Technical Field

[0001] The present disclosure relates generally to methods and network entities for estimating a position of radio antennas in a wireless communication network. The present disclosure also relates to computer programs and carriers corresponding to the above methods and network entities. The present disclosure further relates to methods and network entities for estimating a direction of a radio antenna in a wireless communication network. The present disclosure also relates to computer programs and carriers corresponding to the above methods and network entities.

Background

[0002] Nowadays different types of wireless communication network technologies are being used or developed, e.g., 3 ^rd Generation (3G) wireless communication network technology, 4 ^th Generation (4G) wireless communication network technology, aka Long-Term Evolution (LTE) technology, 5 ^th Generation (5G) technology, also called New Radio (NR) access, etc.

[0003] In a wireless communication network, multiple user equipment (UE) wirelessly communicates to one or more radio antennas. The one or more radio antennas can be integrated with one or more radio units (RU), or being located separately from the RUs, e.g., the radio antennas and the RUs being located on different floors. The one or more radio antennas are located in one same location but can have different pointing directions. Each radio antenna serves a cell, which covers a logical area, and transmits/receives analog radio signals to/from UEs wirelessly. These UEs are served by the cell and the corresponding radio antenna. The one or more RUs are logical functions which perform transmitting and receiving functions and radio frequency measurements, etc. For example, for uplink communication, the one or more RUs can receive the analog radio signals from the radio antennas and converts the analog radio signals to digital signals. [0004] The one or more RUs are also connected to other nodes of the communication network, for example, a radio access network (RAN) node. The RAN node may be connected to other radio network nodes or core network nodes. The RAN node may have computing capacities and are capable of processing the digital signals received from the RUs. The connection between the RUs and the other nodes can be wired or wireless. The other nodes will send the processed signals to core network for switching.

[0005] The knowledge of geographic positions of the radio antennas is beneficial for many use cases and services, especially in different types of radio access network (RAN). Therefore there is a need to estimate the position of radio antennas which are located in the same position.

[0006] Global Positioning System (GPS) is the most frequently used positioning system for obtaining geographic position. However, the radio antennas usually do not have GPS receiver port. Even if the GPS receiver port is installed on some of the radio antennas, it may face critical drawbacks in various situations, mainly due to blockage of satellite signals in indoor scenarios. In addition, equipping radio antennas with GPS receivers will add extra cost to the system. Hence, an alternative positioning method is needed. A possible alternative method is to extract the location information from the UE measurements, which means that the UEs perform measurements on the signals received from the radio antennas, and the location of the radio antennas are calculated based on the UE measurements. Traditionally, the radio antenna geographic position is calculated by a field technician based on the UE measurements. Automation of the process would save time and cost for the operators. Furthermore, there are other challenges to position the radio antenna, e.g., the relationship between the UE measurements and the location of radio antennas is unknown.

[0007] Therefore, there is a need to automatically and efficiently estimate the position of a radio antennas of an RU. Moreover, when the position of the radio antennas is known, it is also important to estimate the directions of each radio antenna. Therefore, there is also a need to automatically and efficiently estimate the direction of a radio antenna of an RU. Summary

[0008] It is an object of the invention to address at least some of the problems and issues outlined above. It is possible to achieve these objects and others by using methods, and one or more network entities as defined in the independent claims. It is an object of embodiments of the invention to perform radio antennas position estimation automatically and efficiently. It is also an object of embodiments of the invention to perform radio antenna direction estimation automatically and efficiently. It is possible to achieve one or more of these objects and possibly others by using methods and one or more network entities as defined in the attached independent claims.

[0009] According to one aspect, a method is provided. The method is performed by one or more network entities of a wireless communication network. The method is used to estimate a position of N radio antennas of the wireless communication network and N is an integer equal to or larger than 1. The N radio antennas have one same position. Each of the N radio antennas serves one cell which associates with a direction of the antenna, the N radio antennas serve N cells in total. Each of the N radio antennas wirelessly communicates with one or more UEs, each UE is served by one cell, the one cell is the serving cell of the UE and the corresponding radio antenna is the serving radio antenna of the UE. Each UE performs measurements on radio signals received from the serving radio antenna, each measurement comprises signal measurement and the location where the UE performs the measurement. The measurements of UEs being associated with the serving cell and the serving radio antenna of the UE. The method comprises obtaining the measurements of each UE and dividing a geographical area into a grid. The intersections of the grid are vertexes, and the geographical area having a minimum size containing the measurement locations associated with the N cells. The method further comprises: for each vertex, determining N circle sectors, each circle sector having a defined radius and a center line extending away from the vertex. The direction of the center line of each circle sector 506 has the same direction as one of the N radio antennas respectively, and the central angle of each circle sector is the same as the radiation angle of corresponding radio antenna, so that each circle sector is associated with a corresponding radio antenna and a corresponding serving cell. The method further comprises: for each circle sector, determining a number of circle sector measurements, the circle sector measurements are the measurements which are located in the circle sector and associated with the corresponding serving cell, and out of the circle sector measurements determining a number of high quality circle sector measurements, the high quality circle sector measurements are the circle sector measurements having signal measurement higher than a threshold. The method further comprises: for each circle sector, calculating a first ratio of the number of circle sector measurements over the total number of the measurements associated with corresponding serving cell, and a second ratio of the number of high quality circle sector measurements over the total number of the measurements associated with corresponding serving cell, the first ratio and the second ratio being associated with the vertex. The method further comprises estimating the position of the radio antenna based on the one or more of the vertexes that have first ratios and second ratios over corresponding ratio thresholds.

[00010] According to another aspect, one or more network entities are provided. The one or more network entities are configured to operate in a wireless communication network and configured for estimating a position of N radio antennas of the wireless communication network, whereby N being an integer equal to or larger than 1 . The N radio antennas having one same position, each of the N radio antennas serves one cell which associates with a direction of the antenna, the N radio antennas serve N cells in total. Each of the N radio antennas wirelessly communicates with one or more user equipment UEs, each UE being served by one cell, the one cell being the serving cell of the UE and corresponding radio antenna being the serving radio antenna of the UE. Each UE performs measurements on radio signals received from the serving radio antenna, each measurement comprises signal measurement and the location where the UE performs the measurement, each measurement being associated with the serving cell and the serving radio antenna of the UE which performs the measurement. The one or more network entities comprise a processing circuitry and a memory, said memory containing instructions executable by said processing circuitry. The one or more network entities is operative for obtaining the measurements of each UE and dividing a geographical area into a grid, the intersections of the grid being vertexes, the geographical area having a minimum size containing the measurement locations associated with the N cells. The one or more network entities is further operative for: for each vertex, determining N circle sectors, each circle sector having a defined radius and a center line extending away from the vertex, the direction of the center line of each circle sector having the same direction as one of the N radio antennas respectively, and the central angle of each circle sector being the same as a the radiation angle of corresponding radio antenna , so that each circle sector being associated with a corresponding radio antenna and a corresponding serving cell . The one or more network entities is further operative for: for each circle sector, determining a number of circle sector measurements, the circle sector measurements are the measurements which are located in the circle sector and associated with the corresponding serving cell , and out of the circle sector measurements determining a number of high quality circle sector measurements, the high quality circle sector measurements are the circle sector measurements having signal measurement higher than a threshold. The one or more network entities is further operative for: for each circle sector, calculating a first ratio of the number of circle sector measurements over the total number of the measurements associated with corresponding serving cell , and a second ratio of the number of high quality circle sector measurements over the total number of the measurements associated with corresponding serving cell , the first ratio and the second ratio being associated with the vertex. The one or more network entities is further operative for estimating the position of the radio antennas based on the positions of one or more of the vertexes that have first ratios and second ratios over corresponding ratio thresholds.

[00011] According to another aspect, a method is provided. The method is used for estimating the direction of a target radio antenna, when the position of the radio antennas is known. This method is performed by one or more network entities of a wireless communication network. N radio antennas are situated in the wireless communication network, N being an integer equal to or larger than 1. The N radio antennas have one same position. The method is used to estimate a direction of one radio antenna of the N radio antennas, namely a target antenna. Each of the N radio antennas serves one cell which associates with a direction of the radio antenna, the N radio antennas serve N cells in total. Each of the N radio antennas wirelessly communicates with one or more UEs. Each UE is served by one cell, the one cell is the serving cell of the UE and corresponding radio antenna is the serving radio antenna of the UE. Each UE performs measurements on radio signals received from the serving radio antenna, each measurement comprises signal measurement and the location where the UE performs the measurement. Each measurement is associated with the serving cell and the serving radio antenna of the UE which performs the measurement. The method comprises obtaining the measurements of each UE and determining, for the target antenna, one or more half lines, each half line having one end in a position of the target antenna, the other end extends away from the target radio antenna position. The method further comprises: for each half line, determining a segment sector, taking the half line as the center line of the segment sector, taking a distance as the radius of the segment sector, and a central angle of the determined segment sector being the same as the radiation angle of the target radio antenna. The method further comprises: for each determined segment sector, determining N numbers of segment sector measurements, each number comprises the measurements which are located in the segment sector and being associated with one of the cells, and out of the segment sector measurements determining a number of high quality segment sector measurements, the high quality segment sector measurements are the segment sector measurements having signal measurement higher than a threshold. The method further comprises: for each determined segment sector, calculating N first ratios, each first radio being the number of segment sector measurements associated with each cell over the total number of the measurements associated with each same cell, and calculating N second ratios, each second ratio being the number of high quality segment sector measurements associated with each cell over the total number of the measurements associated with each same cell, the N first ratios and N second ratios being associated to the half line. The method further comprises estimating the direction of the target antenna based on the one or more half lines which having N first ratios and N second ratios over corresponding ratio thresholds. [00012] According to another aspect, one or more network entities are provided. The one or more network entities is configured to operate in a wireless communication network, wherein N radio antennas are situated in the wireless communication network, N being an integer equal to or larger than 1 and the N radio antennas having one same position. The one or more network entities is configured for estimating a direction of one radio antenna of the N radio antennas, namely a target antenna, each of the N radio antennas serves one cell which associates with a direction of the radio antenna, the N radio antennas serve N cells in total. Each of the N radio antenna wirelessly communicates with one or more user equipment UE, each UE being served by one cell, the one cell being the serving cell of the UE and corresponding radio antenna being the serving radio antenna of the UE. Each UE performs measurements on radio signals received from the serving radio antenna, each measurement comprises signal measurement and the location where the UE performs the measurement, each measurement being associated with the serving cell and the serving radio antenna of the UE which performs the measurement. The one or more network entities comprises a processing circuitry and a memory, said memory containing instructions executable by said processing circuitry. The one or more network entities is operative for obtaining the measurements of each UE and determining, for the target antenna, one or more half lines, each half line having one end in a position of the target antenna, the other end extends away from the target antenna position. The one or more network entities is further operative for: for each half line, determining a segment sector, taking the half line as the center line of the segment sector, taking a distance as the radius of the segment sector, and a central angle of the determined segment sector being the same as the radiation angle of the target radio antenna. The one or more network entities is further operative for: for each determined segment sector, determining N numbers of segment sector measurements, each number comprises the measurements which are located in the segment sector and being associated with one of the cells, and out of the segment sector measurements determining a number of high quality segment sector measurements, the high quality segment sector measurements are the segment sector measurements having signal measurement higher than a threshold. The one or more network entities is further operative for: for each determined segment sector, calculating N first ratios, each first radio being the number of segment sector measurements associated with each cell over the total number of the measurements associated with each same cell, and calculating N second ratios, each second ratio being the number of high quality segment sector measurements associated with each cell over the total number of the measurements associated with each same cell, the N first ratios and N second ratios being associated to the half line. The one or more network entities is further operative for estimating the direction of the target antenna based on the one or more half lines which having N first ratios and N second ratios over corresponding ratio thresholds.

[00013] According to other aspects, computer programs and carriers are also provided, the details of which will be described in the claims and the detailed description.

[00014] Further possible features and benefits of this solution will become apparent from the detailed description below.

Brief Description of Drawings

[00015] The solution will now be described in more detail by means of exemplary embodiments and with reference to the accompanying drawings, in which:

[00016] Fig. 1 is a schematic block diagram of a wireless communication network in which the embodiments of the present invention may be used.

[00017] Fig. 2 is a schematic block diagram of the radio antennas in which the embodiments of the present invention may be used.

[00018] Fig. 3 is a flow chart illustrating a method performed by one or more network entities, according to possible embodiments.

[00019] Fig. 4 is a schematic block diagram of the UE measurements which are associated with different cells, according to possible embodiments. [00020] Fig. 5 is a schematic block diagram of the UE measurements which are discarded, according to possible embodiments.

[00021 ] Fig. 6 is a schematic block diagram of the grid on a geographic area, according to possible embodiments.

[00022] Fig. 7 is a schematic block diagram of the determined circle sectors of one vertex, according to possible embodiments.

[00023] Fig. 8A and 8B are schematic block diagrams of the probabilities of the vertexes, according to possible embodiments.

[00024] Fig. 9A and 9B are schematic block diagrams of the labelled vertexes in a training phase, according to possible embodiments.

[00025] Fig. 10A is a schematic block diagram of a training phase of a machine learning (ML) method, according to possible embodiments.

[00026] Fig. 10B is a schematic block diagram of an inference phase of the ML method, according to possible embodiments.

[00027] Fig. 11 is a block diagram illustrating the network entity in more detail, according to further possible embodiments, according to possible embodiments.

[00028] Fig. 12 is a flow chart illustrating a method performed by one or more network entities, according to possible embodiments.

[00029] Fig. 13 is a schematic block diagram of the determined half lines, according to possible embodiments.

[00030] Fig. 14 is a schematic block diagram of the determined segment sectors, according to possible embodiments.

[00031 ] Fig. 15A and 15B are schematic block diagrams of the probabilities of the half lines, according to possible embodiments. [00032] Fig. 16A and 16B are schematic block diagrams of the labels of the half lines in a training phase, according to possible embodiments.

[00033] Fig. 17A is a schematic block diagram of a training phase of a ML method, according to possible embodiments.

[00034] Fig. 17B is a schematic block diagram of an inference phase of the ML method, according to possible embodiments.

[00035] Fig. 18 is a block diagram illustrating the network entity in more detail, according to further possible embodiments, according to possible embodiments.

Detailed Description

[00036] Fig. 1 shows a wireless communication network 150 comprising an RU 130 having one or more radio antennas 171 , 172, 173 that are in, or adapted for, wireless communication with a number of wireless devices 140-146. Three radio antennas are shown in fig. 1 , however, N radio antennas can be located in the wireless communication network 150 and having one same location. N is an integer equal to or larger than 1. The radio antennas 171 , 172, 173 provide radio coverage in the wireless communication network 150. The number of wireless devices 140-146 resides in the wireless communication network 150. The radio antennas 171 , 172, 173 are capable of transmitting/receiving radio signals to/from the number of wireless devices 140-146 wirelessly. One or more RUs 130 is connected with the radio antennas 171 , 172, 173. The one or more RUs 130 can be integrated with the radio antennas 171 , 172, 173, as shown in fig. 1 , or being located separately from the radio antennas 171 , 172, 173. One RU can be connected to one antenna, or one RU can be connected to multiple antennas. The RU 130 is capable of transmitting and receiving of radio signals and is capable of converting between analog signals and digital signals. The RU 130 is communicated with to another network node 160.

[00037] Another network node 160 are in, or are adapted for, wireless or wired communication with the RU 130. As fig. 1 shows, the network node 160 has computing capacity and is capable of communicating with the RU 130. [00038] The wireless communication network 150 may be any kind of wireless communication network that can provide radio access to wireless communication devices. Example of such wireless communication networks are Global System for Mobile communication (GSM), Enhanced Data Rates for GSM Evolution (EDGE), Universal Mobile Telecommunications System (UMTS), Code Division Multiple Access 2000 (CDMA 2000), Long Term Evolution (LTE) Frequency Division Duplex (FDD) and Time Division Duplex (TDD), LTE Advanced, Wireless Local Area Networks (WLAN), Worldwide Interoperability for Microwave Access (WiMAX), WiMAX Advanced, as well as 5G wireless communication networks based on technology such as New Radio (NR). However, the embodiments of the following detailed description are described for NR.

[00039] The RU 130 may be any kind of radio unit that provides radio access to the wireless devices 140-146. The radio access is provided via the radio antennas connected to the RU 130.

[00040] The network node 160 can be any kind of network node which has computing capacity and connects to the RU 130. The network node 160 or part of the function of the network node 160 can be implemented in wireless communication network 150, or in a core network, or in a cloud.

[00041 ] The number of wireless devices 140-146 may be any type of device capable of wirelessly communicating with the radio antennas 171 , 172, 173 using radio signals. The number of wireless devices may also be called User Equipment (UE) in this disclosure. For example, the number of wireless devices 140-146 may be a UE, a machine type UE or a UE capable of machine to machine (M2M) communication, a sensor, a tablet, a mobile terminal, a smart phone, a laptop embedded equipped (LEE), a laptop mounted equipment (LME), a USB dongle, a Customer Premises Equipment (CPE) etc.

[00042] The embodiments described may be applicable to single carrier as well as to multicarrier (MC) or carrier aggregation (CA) operation of the number of wireless devices. The term carrier aggregation (CA) may also be called multicarrier system, multi-cell operation, multi-carrier operation, and multi-carrier transmission and/or reception. The embodiments may equally apply for Multi radio bearers (RAB) on some carriers, which means that data and speech are simultaneously scheduled.

[00043] Fig. 2 shows radio antennas 171 , 172, 173. As fig. 2 shows, antennas

171 , 172 and 173 are located in one location, that is the intersection of the antennas. Fig. 2 shows three radio antennas, but the number of the radio antennas can an integer which is equal to or larger than 1 , and the number is configurable. Each antenna serves a cell, which is a logic area served by the antenna, and associated with the antenna direction. For example, the antenna 171 serves the cell 181 , the antenna 172 serves the cell 182 and the antenna 173 serves the cell 183. Although the cells 181 , 182, 183 shown in fig. 2 have sector shapes, the actual logic cells usually do not have regular shapes and may overlap with each other, so the sector shapes are only schematic in fig. 2.

[00044] Fig. 3, in conjunction with figs. 4-7, 8A, 8B, describes a method performed by one or more network entities 160 of a wireless communication network 150. The method is used to estimate a position of N radio antennas 171 ,

172, 173 of the wireless communication network 150 and N is an integer equal to or larger than 1 . The N radio antennas 171 , 172, 173 have one same position. Each of the N radio antennas 171 , 172, 173 serves one cell 181 , 182, 183 which associates with a direction of the antenna, the N radio antennas 171 , 172, 173 serve N cells in total. Each of the N radio antennas 171 , 172, 173 wirelessly communicates with one or more UEs 140-146, each UE 140-146 is served by one cell 181 , 182, 183, the one cell 181 , 182, 183 is the serving cell of the UE 140-146 and the corresponding radio antenna 171 , 172, 173 is the serving radio antenna of the UE 140-146. Each UE 140-146 performs measurements on radio signals received from the serving radio antenna, each measurement comprises signal measurement and the location where the UE performs the measurement. The measurements of UEs 140-146 being associated with the serving cell 181 , 182, 183 and the serving radio antenna 171 , 172, 173 of the UE 140-146. The method comprises obtaining 302 the measurements of each UE 140-146 and dividing 304 a geographical area into a grid. The intersections of the grid are vertexes 502, and the geographical area having a minimum size containing the measurement locations associated with the N cells 181 , 182, 183. The method further comprises: for each vertex 502, determining 306 N circle sectors 506, 508, 510, each circle sector 506, 508, 510 having a defined radius and a center line extending away from the vertex 502. The direction of the center line of each circle sector 506, 508, 510 has the same direction as one of the N radio antennas 171 , 172, 173 respectively , and the central angle of each circle sector 506, 508, 510 is the same as the radiation angle of corresponding radio antenna 171 , 172, 173, so that each circle sector 506, 508, 510 is associated with a corresponding radio antenna 171 , 172, 173 and a corresponding serving cell 181 , 182, 183. The method further comprises: for each circle sector 506, 508, 510, determining 308 a number of circle sector measurements, the circle sector measurements are the measurements which are located in the circle sector 506, 508, 510 and associated with the corresponding serving cell 181 , 182, 183, and out of the circle sector measurements determining a number of high quality circle sector measurements, the high quality circle sector measurements are the circle sector measurements having signal measurement higher than a threshold. The method further comprises: for each circle sector 506, 508, 510, calculating 310 a first ratio of the number of circle sector measurements over the total number of the measurements associated with corresponding serving cell 181 , 182, 183, and a second ratio of the number of high quality circle sector measurements over the total number of the measurements associated with corresponding serving cell 181 , 182, 183, the first ratio and the second ratio being associated with the vertex 502. The method further comprises estimating 312 the position of the radio antenna 171 , 172, 173 based on the one or more of the vertexes that have first ratios and second ratios over corresponding ratio thresholds.

[00045] The one or more network entities 160 correspond to the network node 160 in fig. 1 , and may be realized at or in one of the RAN nodes. Alternatively, the one or more network entities 160 may be arranged at or in any other network node of the wireless communication network 150. Alternatively, the one or more network entities 160 may be realized as a group of network nodes, wherein functionality of the one or more network entities 160 is spread out over the group of network nodes. The group of network nodes may be different physical, or virtual, nodes of the network. This alternative realization may be called a cloudsolution.

[00046] The one or more network entity 160 may be any kind of network entity that provides connection to the RU 130 and computing capacity, alone or in combination with another network entity. For example, the network entity could be a node in the RAN or a node in the core network. Alternatively, the one or more network entity 160 can be a purpose-built hardware with integrated software implemented thereon. The one or more network entity 160 can also be generic hardware, or commercial off the shelf (COTS) hardware platform environment, e.g., Microservice and Cloud native based architecture run on bare metal, Kubernetes environment, x86 hardware, which are deployed in distributed sites or central location. Alternatively, a cloud solution can be used to the one or more network entities 160. Other examples of the one or more network entities 160 are a baseband (BB) system, a base station (BS), a radio BS, a base transceiver station, a BS controller, a central unit (CU) system, a virtualized or cloud based CU, a distributed unit (DU) system, a virtualized or cloud based DU, a rApp, service management and orchestration (SMO) platform, a xApp, a cloud server, a network controller, a Node B (NB), an evolved Node B (eNB), a gNodeB (gNB), a Multi-cell/multicast Coordination Entity, a relay node, an access point (AP), a radio AP, nodes in a distributed antenna system (DAS), a multistandard radio BS (MSR BS), a Radio Access Network (RAN) node, an Open Radio Access Network (O-RAN) node, etc.

[00047] N cells per frequency are served by the N radio antennas, wherein each radio antenna serves one cell per frequency. Each UE 140-146 performs measurements on received signals from corresponding radio antenna. Each measurement includes signal measurement and the location where the UE performs the measurement. The signal measurement includes the measurements of signal strength, signal quality, and/or signal to noise ratio, e.g., received signal strength indication (RSSI), reference signal received power (RSRP), reference signal received quality (RSRQ), signal-to-noise ratio (SNR), signal-to-interference-plus-noise ratio (SINR), Timing Advance (TA), etc. Other relevant signal measurements can also be included.

[00048] Each measurement is associated with a serving cell and a serving radio antenna. Referring to fig. 1 , fig. 2 and fig. 4, the dots in fig. 4 refer to the measurements performed by the UEs located different locations and associate with the cells 181 -183 and the radio antennas 171-173. The white dots refer to the measurements associated with the serving cell 181 and the serving radio antenna 171 , that is, the UEs which perform these measurements are served by the cell 181 and the radio antenna 171. Similarly, the black dots refer to the measurements associated with the serving cell 182 and the serving radio antenna 172. The dots with diamond mesh refer to the measurements associated with the serving cell 183 and the serving ratio antenna 173.

[00049] According to another embodiment, the method further comprises grouping the obtained 302 measurements which are associated with the N cells 181 , 182, 183, based on the locations of the measurements and a group distance threshold. The method further comprises discarding the measurements of one or more groups before determining 308 the number of circle sector measurements and the number of high quality circle sector measurements, the measurements in the discarded one or more groups having an average distance from an average position of the measurements which belong to the group having the largest number of measurements, the average distance is higher than a discarding distance threshold.

[00050] As shown in fig. 5, the measurements are grouped according to the distance between each other. For example, when the distance between two measurement locations is less than a group distance threshold, the two measurements are grouped into one group. Among all the groups, there is one group which has the largest number of measurements. The measurements in this largest group have an average position. Measurements of one or more of other groups are discarded, if an average distance from the measurements in the one or more discarded groups to the average position is higher than a discarding distance threshold. These measurements in the one or more groups are deemed as outliers and discarded. The discarding should be performed before the determining 308 of the circle sector measurements number and the high quality circle sector measurement number, so that the determination of the numbers are more accurate. As the fig. 5 shows, the measurements which are grouped in the circles are discarded as outliers.

[00051 ] In the step of dividing 304, referring to fig. 6, a geographical area is defined. The geographical area has a minimum size which contains all the measurement locations associated with the N cells 181 , 182, 183. The shape of the geographical area can be e.g., rectangle. The geographical area is divided into a grid, and the size of the grid is configurable, e.g., 20 meters * 20 meters. Each of the intersections of the grid is a vertex, e.g., the vertex 502.

[00052] In the step of determining 306, referring to fig. 7, for each vertex 502, N circle sectors 506, 508, 510 are determined. N is three in this embodiment. For each circle sector, the circle sector has a defined radius, which is configurable. A center line of the circle sector is extending away from the vertex 502 and has the same direction as a radio antenna. The central angle of the circle sector is the same as the radiation angle of corresponding radio antenna, e.g., the same as the horizontal radiation angle of the radio antenna. For example, for the circle sector 506, the center line 512 has a same extending direction as the radio antenna 171 , and the central angle is the same as the horizontal radiation angle of the corresponding radio antenna 171. The central angle refers to the angle formed by the two arms of the sector, and the radiation angle refers to the horizonal angle of the main beam of the antenna, and the radiation angle is an attribute of the antenna. Obviously, the center line 512 divides the central angle of the circle sector 506, so that inside the circle sector 506, the angle to the left of the center line 512 is the same as the angle to the right of the center line 512, and both of them are equal to one half of the antenna horizonal radiation angle. The circle sector 506 is associated with the radio antenna 171 and the serving cell 181.

[00053] Similarly, for the circle sector 508, the center line 514 extends in the same direction of the antenna 172, and the central angle of the circle sector 508 is equal to the horizontal radiation angle of the antenna 172. The circle sector 508 is associated with the radio antenna 172 and the serving cell 182. For the circle sector 510, the center line 516 extends in the same direction of the antenna 173, and the central angle of the circle sector 510 is equal to the horizontal radiation angle of the antenna 173. The circle sector 510 is associated with the radio antenna 173 and the serving cell 183. Thus N circle sectors are determined in this way.

[00054] In the step of determining 308, for each circle sector, a number of circle sector measurements is determined. The circle sector measurements refer to those measurements which are located in the circle sector, and being associated with corresponding serving cell. For example, for the circle sector 506, the circle sector measurements are those measurements which are located therein and associated with the serving cell 181. Since the white dots in the fig. 7 all associated with the serving cell 181 , the circle sector measurements of the circle sector 506 are those white dots which are located in the circle sector 506. In fig. 7, the number of the circle sector measurements of the circle sector 506 is 7.

[00055] Another number, the number of high quality circle sector measurements are also determined. The high quality circle sector measurements refer to those measurements which are located in the circle sector, being associated with corresponding serving cell and having signal measurement higher than a threshold, that is, those circle sector measurements which have signal measurement higher than a signal measurement threshold. For example, for the circle sector 506, among the circle sector measurements, 4 of them have a signal measurement higher than the threshold. Therefore, the high quality circle sector measurement of the circle sector 506 is 4.

[00056] Similarly, the number of circle sector measurement of the circle sector 508 is 12, by counting the black dots located in the circle sector 508, because the black dots are the measurements which are associated with corresponding serving cell 182 of the circle sector 508. The number of high quality circle sector measurement of the circle sector 508 is e.g., 10, which is the number of the circle sector measurements which have a signal measurement higher than a threshold. [00057] Similarly, the number of circle sector measurement of the circle sector 510 is 19, by counting the diamond mesh dots located in the circle sector 510, because the diamond mesh dots are the measurements which are associated with corresponding serving cell 183 of the circle sector 510. The number of high quality circle sector measurement of the circle sector 510 is e.g., 12, which is the number of the circle sector measurements which have a signal measurement higher than a threshold.

[00058] In the step of calculating 310, for each circle sector, a first ratio is calculated via dividing the number of circle sector measurements by the total number of the measurements associated with corresponding serving cell. Referring to fig. 7, for the circle sector 506, the first ratio is to divide the number of circle sector measurements of the circle sector 506 by the total number of the measurements associated with the serving cell 181. The total number of the measurement associated with the serving cell 181 is the total number of the white dots, which is 18. The first ratio of the circle sector 506 is dividing 7 by 18, which is about 0.39.

[00059] A second ratio of each circle sector is also calculated. The second ratio is calculated via dividing the number of high quality circle sector measurements by the total number of the measurements associated with corresponding serving cell. Referring to fig. 7, for the circle sector 506, the second ratio is to divide the number of high quality circle sector measurement of the circle sector 506 by the total number of the measurements associated with the serving cell 181 . The second ratio of the circle sector 506 is dividing 4 by 18, which is about 0.22.

[00060] Similarly, for the circle sector 508, the total number of the measurements associated with the corresponding serving cell 182 is the total number of black dots in fig. 7, that is 20. Thus the first ratio is 12/20, 0.6, the second ratio is 10/20, 0.5.

[00061 ] Similarly, for the circle sector 510, the total number of the measurements associated with the corresponding serving cell 130 is the total number of diamond mesh dots in fig. 7, that is 19. Thus the first ratio is 19/19, 1 , the second ratio is 12/19, 0.63.

[00062] Since there are N circle sectors, for each vertex, there are N first ratios and N second ratios associated with the vertex. For example, 3 first ratios: 0.39, 0.6 and 1 , and 3 second ratios: 0.22, 0.5 and 0.63 are associated with the vertex 502.

[00063] In the estimating step 312, since each vertex has N first ratios and N second ratios, the position of the radio antennas is estimated based on the position of the one or more vertexes which have the N first ratios and N second ratios over corresponding thresholds. For example, when each of the N first ratios and the N second ratios of one vertex is higher than 0.7, the position of the radio antennas is considered as the position of this vertex.

[00064] In this method, each vertex is deemed as a candidate of the position of the radio antennas. When the first radios and the second ratios are near to 1 , the position of corresponding vertex is near to the real position of the radio antennas. It is obvious that if the real position of the radio antennas is taken as a vertex, the first radios and second ratios of this vertex should be very close to 1 . By such a method, the position of the radio antennas can be calculated efficiently, precisely and automatically.

[00065] According to an embodiment, referring to in fig. 8A and 8B, the estimating 312 step further comprises: for each vertex 502, determining a probability ranging from 0 to 1 , based on the first and second ratios of each circle sector 506, 508, 510, the probability indicating the probability of the N radio antennas 171 , 172, 173 being positioned at the vertex 502. The method further comprises estimating the position of the N radio antennas 171 , 172, 173 based on the positions of the one or more vertexes which have the probability higher than a probability threshold.

[00066] For example, referring to fig. 8A, a probability is determined for each vertex, e.g., 0.1 , 0.3, etc. The threshold of the probability is e.g., 0.9, so that the position of the vertex with the probability 0.9 is estimated as the position of the N radio antennas.

[00067] For another example, referring to fig. 8B, the threshold of the probability is e.g., 0.7, so that the position of the radio antennas 171 , 172, 173 are estimated based on the positions of the four vertexes which having the probabilities 0.7, 0.7, 0.8, 0.9.

[00068] According to this embodiment, when the number of the vertexes which have the probability higher than the threshold is equal to or larger than 1 , there are several alternatives to estimate the radio antenna position based on the vertex position. The first alternative is to estimate the radio antenna position based on the position of the vertex which has the highest probability. The second alternative is to estimate the radio antenna position based on the average position of the vertexes which have probability higher than the threshold. The third alternative is to estimate the radio antenna position based on the average position of one subset of the vertexes which have probability higher than the threshold.

[00069] If there is no vertex which has a probability higher than the threshold, the position of the radio antennas cannot be estimated.

[00070] By this method, the position of the radio antennas is estimated based on the positions of the vertexes which having high probabilities. The estimation of the radio antenna position becomes more accurate.

[00071] According to another embodiment, referring to fig. 9A, 9B, 10A, 10B, the method is performed by machine learning (ML) method using a ML model, the ML method comprising a training phase followed by an inference phase. In the inference phase, the method in the embodiment above is performed, wherein the probability of each vertex 502 is determined by the ML model, taking the first ratios and the second ratios of each vertex 502 as the inputted features of the ML model, and the determined probability of each vertex 502 being a prediction result of the ML model. [00072] The training phase is performed before the inference phase and comprises obtaining the measurements of each UE 140-146 and dividing a geographical area into a grid 190, the intersections of the grid 190 being vertexes 502. The geographical area has a minimum size containing the measurement locations associated with the N cells 181 , 182, 183. The training phase further comprises for each vertex 502, determining N circle sectors 506, 508, 510, each circle sector 506, 508, 510 having a defined radius and a center line extending away from the vertex 502, the direction of the center line of each circle sector 506, 508, 510 having the same direction as one of the N radio antennas 171 , 172, 173 respectively, and the central angle of each circle sector 506, 508, 510 being the same as the horizontal radiation angle of corresponding radio antenna 171 , 172, 173, so that each circle sector 506, 508, 510 being associated with a corresponding radio antenna 171 , 172, 173 and a corresponding serving cell 181 , 182, 183. The training phase further comprises: for each circle sector 506, 508, 510, determining a number of circle sector measurements, the circle sector measurements are the measurements which are located in the circle sector 506, 508, 510 and associated with the corresponding serving cell 181 , 182, 183, and out of the circle sector measurements determining a number of high quality circle sector measurements, the high quality circle sector measurements are the circle sector measurements having signal measurement higher than a threshold. The method further comprises: for each circle sector 506, 508, 510, calculating a first ratio of the number of circle sector measurements over the total number of the measurements associated with corresponding serving cell 181 , 182, 183, and a second ratio of the number of high quality circle sector measurements over the total number of the measurements associated with corresponding serving cell 181 , 182, 183, the first ratio and the second ratio being associated with the vertex 502. The training phase further comprises receiving a label for each vertex 502, the label being 0 or 1 , 0 indicating that the radio antennas 171 , 172, 173 are not located in the position of the vertex 502, 1 indicating that the radio antennas 171 , 172, 173 are located in the position of the vertex 502. The training phase further comprises training the ML model based on the first ratios and second ratios of each vertex 502 as inputted features and based on the received label for each vertex 502 as the inputted target of the ML model.

[00073] The training phase actually have same steps as the inference phase, except for the receiving step and the training step. As shown in fig. 9A and 9B, labels 0 or 1 for each vertex are received from technician and being used as inputted target for training the ML model. The real position of the radio antennas is known in the training phase. In one alternative, for the vertex which is the closest to the real position of the radio antennas, a label 1 is received. Labels 0 are received for other vertexes. In another alternative, for the four vertex which are closest to the real position of the radio antennas, four labels 1 are received. Labels 0 are received for other vertexes. The ML model are trained based on the vertex features and the received labels.

[00074] As shown in fig. 10A and 10B, the machine learning method comprises a training phase and an inference phase. Referring to fig. 10A, the training phase comprises data preparation, grid generation, feature extraction, training or retraining ML model and an optional post processing. The data preparation is corresponding to the step of obtaining the UE measurements and the optional step of discarding outliers. The grid generation is corresponding to the step of dividing the grid. The feature extraction is corresponding to the step of determining N circle sectors, the step of determining the number of circle sector measurements/the number of high quality circle sector measurements and the step of determining the N first ratios and the N second ratios. The training or retraining ML model is corresponding to the step of receiving label and the step of training the ML model, taking the received 0 or 1 labels as ground truth. An optional post processing can be added.

[00075] Similarly, referring to fig. 10B, the inference phase comprises data preparation, grid generation, feature extraction, prediction using trained model and an optional post processing. The data preparation is corresponding to the step of obtaining 302 the UE measurements and the optional step of discarding outliers. The grid generation is corresponding to the step of dividing 304 the grid. The feature extraction is corresponding to the step of determining 306 N circle sectors, the step of determining 308 the number of circle sector measurements/the number of high quality circle sector measurements and the step of determining 310 the N first ratios and the N second ratios. The prediction using trained model is corresponding to the step of estimating 312 the position of the radio antenna. The optional post processing can be added.

[00076] By such an embodiment, a ML model can be used to estimate the position of the radio antennas. When the amount of the measurement becomes quite large, the ML model has significant advantages.

[00077] According to another embodiment, one or more network entities 160 are provided. The one or more network entities 160 are configured to operate in a wireless communication network 150 and configured for estimating a position of N radio antennas 171 , 172, 173 of the wireless communication network 150, whereby N being an integer equal to or larger than 1. The N radio antennas 171 , 172, 173 having one same position, each of the N radio antennas 171 , 172, 173 serves one cell 181 , 182, 183 which associates with a direction of the antenna, the N radio antennas 171 , 172, 173 serve N cells in total. Each of the N radio antennas 171 , 172, 173 wirelessly communicates with one or more user equipment UEs 140-146, each UE 140-146 being served by one cell 181 , 182, 183, the one cell 181 , 182, 183 being the serving cell of the UE 140-146 and corresponding radio antenna 171 , 172, 173 being the serving radio antenna of the UE 140-146. Each UE 140-146 performs measurements on radio signals received from the serving radio antenna, each measurement comprises signal measurement and the location where the UE performs the measurement, each measurement being associated with the serving cell 181 , 182, 183 and the serving radio antenna 171 , 172, 173 of the UE 140-146 which performs the measurement. The one or more network entities 160 comprise a processing circuitry 603 and a memory 604, said memory 604 containing instructions executable by said processing circuitry 603. The one or more network entities 160 is operative for obtaining the measurements of each UE 140-146 and dividing a geographical area into a grid 190, the intersections of the grid 190 being vertexes 502, the geographical area having a minimum size containing the measurement locations associated with the N cells 181 , 182, 183. The one or more network entities 160 is further operative for: for each vertex 502, determining N circle sectors 506, 508, 510, each circle sector 506, 508, 510 having a defined radius and a center line extending away from the vertex 502, the direction of the center line of each circle sector 506, 508, 510 having the same direction as one of the N radio antennas

171 , 172, 173 respectively, and the central angle of each circle sector 506, 508, 510 being the same as a the radiation angle of corresponding radio antenna 171 ,

172, 173, so that each circle sector 506, 508, 510 being associated with a corresponding radio antenna 171 , 172, 173 and a corresponding serving cell 181 , 182, 183. The one or more network entities 160 is further operative for: for each circle sector 506, 508, 510, determining a number of circle sector measurements, the circle sector measurements are the measurements which are located in the circle sector 506, 508, 510 and associated with the corresponding serving cell 181 , 182, 183, and out of the circle sector measurements determining a number of high quality circle sector measurements, the high quality circle sector measurements are the circle sector measurements having signal measurement higher than a threshold. The one or more network entities 160 is further operative for: for each circle sector 506, 508, 510, calculating a first ratio of the number of circle sector measurements over the total number of the measurements associated with corresponding serving cell 181 , 182, 183, and a second ratio of the number of high quality circle sector measurements over the total number of the measurements associated with corresponding serving cell 181 , 182, 183, the first ratio and the second ratio being associated with the vertex 502. The one or more network entities 160 is further operative for estimating the position of the radio antennas 171 , 172, 173 based on the positions of one or more of the vertexes that have first ratios and second ratios over corresponding ratio thresholds.

[00078] According to another embodiment, estimating of the position of the radio antennas 171 , 172, 173 further comprises: for each vertex 502, determining a probability ranging from 0 to 1 , based on the first and second ratios of each circle sector 506, 508, 510, the probability indicating the probability of the N radio antennas 171 , 172, 173 being positioned at the vertex 502, and estimating the position of the N radio antennas 171 , 172, 173 based on the positions of the one or more vertexes which having the probability higher than a probability threshold.

[00079] According to another embodiment, the one or more network entities 160 further being configured to performed a machine learning ,ML, method using a ML model, the ML method comprising a training phase followed by an inference phase, in the inference phase, the one or more network entities 160 is configured to perform steps above, wherein the probability of each vertex 502 is determined by the ML model, taking the first ratios and the second ratios of each vertex 502 as the inputted features of the ML model, the determined probability of each vertex 502 being a prediction result of the ML model. The training phase is performed before the inference phase and comprises obtaining the measurements of each UE 140-146 and dividing a geographical area into a grid 190, the intersections of the grid 190 being vertexes 502, the geographical area having a minimum size containing the measurement locations associated with the N cells. The training phase further comprises for each vertex 502, determining N circle sectors 506, 508, 510, each circle sector 506, 508, 510 having a defined radius and a center line extending away from the vertex 502, the direction of the center line of each circle sector 506, 508, 510 having the same direction as one of the N radio antennas 171 , 172, 173 respectively, and the central angle of each circle sector 506, 508, 510 being the same as the horizontal radiation angle of corresponding radio antenna 171 , 172, 173, so that each circle sector 506, 508, 510 being associated with a corresponding radio antenna 171 , 172, 173 and a corresponding serving cell 181 , 182, 183. The training phase further comprises: for each circle sector 506, 508, 510, determining a number of circle sector measurements, the circle sector measurements are the measurements which are located in the circle sector 506, 508, 510 and associated with the corresponding serving cell 181 , 182, 183, and out of the circle sector measurements determining a number of high quality circle sector measurements, the high quality circle sector measurements are the circle sector measurements having signal measurement higher than a threshold. The training phase further comprises: for each circle sector 506, 508, 510, calculating a first ratio of the number of circle sector measurements over the total number of the measurements associated with corresponding serving cell 181 , 182, 183, and a second ratio of the number of high quality circle sector measurements over the total number of the measurements associated with corresponding serving cell 181 , 182, 183, the first ratio and the second ratio being associated with the vertex 502. The training phase further comprises receiving a label for each vertex 502, the label being 0 or 1 , 0 indicating that the radio antennas 171 , 172, 173 are not located in the position of the vertex 502, 1 indicating that the radio antennas 171 , 172, 173 are located in the position of the vertex 502, and training the ML model based on the first ratios and second ratios of each vertex 502 as inputted features and based on the received label for each vertex 502 as the inputted target of the ML model.

[00080] According to another embodiment, the one or more network entities 160 is further operative for grouping the obtained measurements which are associated with the N cells 181 , 182, 183, based on the locations of the measurements and a group distance threshold, and discarding the measurements of one or more groups before determining the number of circle sector measurements and the number of high quality circle sector measurements , the measurements in the discarded one or more groups having an average distance from an average position of the measurements which belong to the group having the largest number of measurements, the average distance is higher than a discarding distance threshold.

[00081 ] According to other embodiments, referring to fig. 11 , the network entity 160 may further comprise a communication unit 602, which may be considered to comprise conventional means for wireless communication with other devices, such as a transceiver for wireless transmission and reception of signals. The instructions executable by said processing circuitry 603 may be arranged as a computer program 605 stored e.g. in said memory 604. The processing circuitry 603 and the memory 604 may be arranged in a sub-arrangement 601 . The subarrangement 601 may be a micro-processor and adequate software and storage therefore, a Programmable Logic Device, PLD, or other electronic component(s)/processing circuit(s) configured to perform the methods mentioned above. The processing circuitry 603 may comprise one or more programmable processor, application-specific integrated circuits, field programmable gate arrays or combinations of these adapted to execute instructions.

[00082] The computer program 605 may be arranged such that when its instructions are run in the processing circuitry, they cause network entity 160 to perform the steps described in any of the described embodiments of the network entity 160 and its method. The computer program 605 may be carried by a computer program product connectable to the processing circuitry 603. The computer program product may be the memory 604, or at least arranged in the memory. The memory 604 may be realized as for example a RAM (Randomaccess memory), ROM (Read-Only Memory) or an EEPROM (Electrical Erasable Programmable ROM). In some embodiments, a carrier may contain the computer program 605. The carrier may be one of an electronic signal, an optical signal, an electromagnetic signal, a magnetic signal, an electric signal, a radio signal, a microwave signal, or computer readable storage medium. The computer-readable storage medium may be e.g. a CD, DVD or flash memory, from which the program could be downloaded into the memory 604. Alternatively, the computer program may be stored on a server or any other entity to which the network entity 160 has access via the communication unit 602. The computer program 605 may then be downloaded from the server into the memory 604.

[00083] According to another embodiment, referring to figs. 1 , 2, 4, 5, 12-14, 15A and 15B, another method is disclosed, the method is used for estimating the direction of a target radio antenna, when the position of the radio antennas is known. This method is performed by one or more network entities 160 of a wireless communication network 150. N radio antennas 171 , 172, 173 are situated in the wireless communication network 150, N being an integer equal to or larger than 1. The N radio antennas 171 , 172, 173 have one same position. The method is used to estimate a direction of one radio antenna of the N radio antennas 171 , 172, 173, namely a target antenna 171. Each of the N radio antennas 171 , 172, 173 serves one cell 181 , 182, 183 which associates with a direction of the radio antenna, the N radio antennas 171 , 172, 173 serve N cells in total. Each of the N radio antennas 171 , 172, 173 wirelessly communicates with one or more UEs 140- 146. Each UE 140-146 is served by one cell 181 , 182, 183, the one cell 181 , 182, 183 is the serving cell of the UE 140-146 and corresponding radio antenna 171 , 172, 173 is the serving radio antenna of the UE 140-146. Each UE 140-146 performs measurements on radio signals received from the serving radio antenna, each measurement comprises signal measurement and the location where the UE performs the measurement. Each measurement is associated with the serving cell

181 , 182, 183 and the serving radio antenna 171 , 172, 173 of the UE 140-146 which performs the measurement. The method comprises obtaining 1202 the measurements of each UE 140-146 and determining 1204, for the target antenna 171 , one or more half lines 702, 710, each half line 702, 710 having one end in a position of the target antenna 171 , the other end extends away from the target radio antenna 171 position. The method further comprises: for each half line 702, 710, determining 1206 a segment sector 704, taking the half line 702 as the center line of the segment sector 704, taking a distance as the radius of the segment sector 704, and a central angle of the determined segment sector 704 being the same as the radiation angle of the target radio antenna 171 . The method further comprises: for each determined segment sector 704, determining 1208 N numbers of segment sector measurements, each number comprises the measurements which are located in the segment sector 704 and being associated with one of the cells 181 , 182, 183, and out of the segment sector measurements determining a number of high quality segment sector measurements, the high quality segment sector measurements are the segment sector measurements having signal measurement higher than a threshold. The method further comprises: for each determined segment sector 704, calculating 1210 N first ratios, each first radio being the number of segment sector measurements associated with each cell 181 ,

182, 183 over the total number of the measurements associated with each same cell 181 , 182, 183, and calculating N second ratios, each second ratio being the number of high quality segment sector measurements associated with each cell 181 , 182, 183 over the total number of the measurements associated with each same cell 181 , 182, 183, the N first ratios and N second ratios being associated to the half line 702. The method further comprises estimating 1212 the direction of the target antenna 171 based on the one or more half lines 702 which having N first ratios and N second ratios over corresponding ratio thresholds.

[00084] The direction of the target antenna is also known as the antenna azimuth. Similar as above embodiments, N cells per frequency are served by the N radio antennas, wherein each radio antenna serves one cell per frequency.

Each UE 140-146 performs measurements on received signals from corresponding radio antenna. Each measurement includes signal measurement and the location where the UE performs the measurement. The signal measurement includes the measurements of signal strength, signal quality, and/or signal to noise ratio, e.g., received signal strength indication (RSSI), reference signal received power (RSRP), reference signal received quality (RSRQ), signal- to-noise ratio (SNR), signal-to-interference-plus-noise ratio (SINR), Timing Advance (TA), etc. Other relevant signal measurements can also be included.

[00085] Similar as above embodiments, each measurement is associated with a serving cell and a serving radio antenna. Referring to fig. 1 , fig. 2 and fig. 4, the dots in fig. 4 refer to the measurements performed by the UEs located different locations and associate with the cells 181-183 and the radio antennas 171-173.

The white dots refer to the measurements associated with the serving cell 181 and the serving radio antenna 171 , which is the target antenna. Similarly, the black dots refer to the measurements associated with the serving cell 182 and the serving radio antenna 172. The dots with diamond mesh refer to the measurements associated with the serving cell 183 and the serving ratio antenna 173.

[00086] According to one embodiment, the method further comprises grouping the obtained 1202 measurements which are associated with the N cells 181 , 182, 183, based on the locations of the measurements and a group distance threshold, and discarding the measurements of one or more groups before determining 1208 the number of segment sector measurements and the number of high quality segment sector measurements, the measurements in the discarded one or more groups having an average distance from an average position of the measurements which belong to the group having the largest number of measurements, the average distance is higher than a discarding distance threshold.

[00087] Similarly, as shown in fig. 5, the measurements are grouped according to the distance between each other. For example, when the distance between two measurement locations is less than a group distance threshold, the two measurements are grouped into one group. Among all the groups, there is one group which has the largest number of measurements. The measurements in this largest group have an average position. Measurements of one or more of other groups are discarded, if an average distance from the measurements in the one or more discarded groups to the average position is higher than a discarding distance threshold. These measurements in the one or more groups are deemed as outliers and discarded. The discarding should be performed before the determining 1208 of the circle sector measurements number and the high quality circle sector measurement number, so that the determination of the numbers are more accurate. As the fig. 5 shows, the measurements which are grouped in the circles are discarded as outliers.

[00088] In the step of dividing 1204, referring to fig. 12, one or more half lines 702, 710 etc. are determined. The half lines extend respectively away from the position of the target antenna 171 . The number of the half lines and the angle between the half lines can be configurable, wherein the angle between each two half lines can be the same or different.

[00089] In the step of determining 1206, for each half line, taking the half line 702 in fig. 13 as an example, a segment sector 704 is determined based on the half line 702. The half line 702 is the center line of the segment sector 704, which means the angle between one radius 706 and the half line 702 is the same with the angle between the other radius 708 and the half line 702. The segment sector 704 has a central angle which is the same as the radiation angle of the target antenna 171 , e.g., horizontal radiation angle of the target antenna 171 , which means that the angle between the radius 706 and the radius 708 is the same as the horizontal radiation angle of the target antenna 171 . The length of the radius 706 and 708 is configurable. For example, a circle drawn with this radius is a minimum circle containing all the measurement locations associated with the N cells 181 , 182, 183. Thus each half line is associated with a segment sector.

[00090] In the determining 1208 step, for each segment sector, N numbers of segment sector measurements and N numbers of high quality segment sector measurements are determined. In fig. 13, N is three. For the segment sector 704, 3 numbers of segment sector measurements and 3 numbers of high quality segment sector measurements are determined. For cell 181 , the number of segment sector measurements of segment sector 704 is 14, because the number of white dots which are located in the segment sector 704 is 14. For cell 182, the number of segment sector measurements of segment sector 704 is 2, because the number of black dots which are located in the segment sector 704 is 2. Similarly, for cell 183, the number of segment sector measurements of segment sector 704 is 0, because the number of diamond mesh dots which are located in the segment sector 704 is 0. Thus the N numbers of segment sector measurements of the segment sector 704 are determined, they are 14, 2, 0.

[00091] N numbers of high quality segment sector measurements are also determined. Referring to fig. 13, for the segment sector 704, for the cell 181 , among those 14 white dots which are located in the segment sector 704, for example, 10 of them have a signal measurement higher than a threshold. Therefore, for the cell 181 , the number of high quality segment sector measurements of the segment sector 704 is 10. Similarly, for the cell 182, the number of high quality segment sector measurements of the segment sector 704 is e.g., 1 . For the cell 183, the number of high quality segment sector measurements of the segment sector 704 is 0. Thus the N numbers of high quality segment sector measurements of the segment sector 704 are determined, they are 10, 1 , 0.

[00092] In the step of calculating 1210, for each segment sector, N first ratios are calculated via dividing the number of segment sector measurements associated with each cell 181 , 182, 183 by the total number of the measurements associated with each same cell 181 , 182, 183. Referring to fig. 13, for the segment sector 704, the first ratio of the cell 181 is to divide the number of segment sector measurements associated with the cell 181 by the total number of the measurements associated with the cell 181. The total number of the measurement associated with the cell 181 is the total number of the white dots, which is 18. Thus for the cell 181 , the first ratio of the segment sector 704 is dividing 14 by 18, which is about 0.78.

[00093] Similarly, for the cell 182, the first ratio of the segment sector 704 is dividing the number of segment sector measurements associated with the cell 182 by the total number of the measurements associated with the cell 182. The total number of the measurements associated with the cell 182 is the total number of the black dots, which is 20. Thus for the cell 182, the first ratio of the segment sector 704 is dividing 2 by 20, which is 0.1 .

[00094] Similarly, for the cell 183, the first ratio of the segment sector 704 is dividing the number of segment sector measurements associated with the cell 183 by the total number of the measurements associated with the cell 183. The total number of the measurements associated with the cell 183 is the total number of the diamond mesh dots, which is 19. Thus for the cell 183, the first ratio of the segment sector 704 is dividing 0 by 19, which is 0. In summary, the N first ratios of the segment sector 704 are 0.78, 0.1 and 0.

[00095] Also, for each segment sector, N second ratios are calculated via dividing the number of high quality segment sector measurements associated with each cell 181 , 182, 183 by the total number of the measurements associated with each same cell 181 , 182, 183. Referring to fig. 13, for the segment sector 704, the second ratio of the cell 181 is to divide the number of high quality segment sector measurements associated with the cell 181 by the total number of the measurements associated with the cell 181 , which is dividing 10 by 18, which is about 0.56.

[00096] Similarly, for the cell 182, the second ratio of the segment sector 704 is dividing the number of high quality segment sector measurements associated with the cell 182 by the total number of the measurements associated with the cell 182, which is dividing 1 by 20, which is 0.05. [00097] Similarly, for the cell 183, the second ratio of the segment sector 704 is dividing the number of high quality segment sector measurements associated with the cell 183 by the total number of the measurements associated with the cell 183, which is dividing 0 by 19, which is 0. In summary, the N second ratios of the segment sector 704 are 0.56, 0.05 and 0.

[00098] Thus, for the segment sector 704, N first ratios 0.78, 0.1 and 0 and N second ratios 0.56, 0.05 and 0 are calculated in the step of calculating 1210, the N first ratios and the N second rations are associated with the half line 702 of the segment sector 704.

[00099] In the estimating step 1212, since each half line has N first ratios and N second ratios, the direction of the target antenna is estimated based on the direction of the one or more half lines which have the N first ratios and N second ratios over corresponding thresholds. For example, when each of the N first ratios and the N second ratios of one half line is higher than 0.8, the direction of the target antenna is estimated based on the direction of this half line.

[000100] In this method, each half line is deemed as a candidate of the direction of the target antenna. When the first radios and the second ratios are near to 1 , the direction of corresponding half line is near to the real direction of the target antenna. It is obvious that if the real direction of the target antenna is taken as a half line, the first radios and second ratios of this half line should be very close to 1 . By such a method, the direction of the target antenna can be calculated efficiently, precisely and automatically.

[000101] According to another embodiment, the estimating 1212 step further comprises: for each half line 702, determining a probability ranging from 0 to 1 , based on the N first ratios and the N second ratios of each half line 702, the probability indicating the probability of the direction of the target antenna 171 being the same as the half line 702. The estimating step 1212 further comprises estimating the direction of the target antenna 171 based on one or more half lines which having the probability higher than a threshold. [000102] For example, referring to fig. 15A, a probability is determined for each half line, e.g., 0.1 , 0.3, etc. The threshold of the probability is e.g., 0.9, so that the direction of the half line with the probability 0.9 is estimated as the direction of the target antenna.

[000103] For another example, referring to fig. 15B, the threshold of the probability is e.g., 0.8, so that the direction of the target antenna 171 is estimated based on the directions of the two half lines which having the probabilities 0.8 and 0.9.

[000104] According to this embodiment, when the number of the half lines which have the probability higher than the threshold is equal to or larger than 1 , there are several alternatives to estimate the target antenna direction based on the direction of the half lines. The first alternative is to estimate the target antenna direction based on the direction of the half line which has the highest probability. The second alternative is to estimate the target antenna direction based on the average direction of the half lines which have probability higher than the threshold. The third alternative is to estimate the target antenna direction based on the average direction of one subset of the half lines which have probability higher than the threshold.

[000105] If there is no half line which has a probability higher than the threshold, the direction of the target antenna cannot be estimated.

[000106] By this method, the direction of the target antenna is estimated based on the directions of the half lines which having high probabilities. The estimation of the target antenna direction becomes more accurate.

[000107] According to another embodiment, the method being performed by machine learning (ML) method using a ML model, the ML method comprising a training phase followed by an inference phase, in the inference phase. The method in above embodiment is performed, wherein the probability of each half line 702 is determined by the ML model, taking the N first ratios and the N second ratios of each half line 702 as the inputted features of the ML model, and the determined label of each half line 702 being a prediction result of the ML model. [000108] The training phase is performed before the inference phase and comprises obtaining the measurements of each UE 140-146 and determining, for the target antenna 171 , one or more half lines 702, 710, each half line 702, 710 having one end in a position of target antenna 171 , the other end extends away from the target antenna 171 position. The method further comprises: for each half line 702, determining a segment sector 704, taking the half line 702 as the center line of the segment sector 704, taking a distance as the radius of the segment sector 704, and a central angle of the determined segment sector 704 being the same as the radiation angle of the target antenna 171. The method further comprises: for each determined segment sector 704, determining N numbers of segment sector measurements, each number comprises the measurements which are located in the segment sector 704 and being associated with one of the cells

181 , 182, 183, and out of the segment sector measurements determining a number of high quality segment sector measurements, the high quality segment sector measurements are the segment sector measurements having signal measurement higher than a threshold. The method further comprises: for each determined segment sector 704, calculating N first ratios, each first radio being the number of segment sector measurements associated with each cell 181 , 182, 183 over the total number of the measurements associated with each same cell 181 ,

182, 183, and calculating N second ratios, each second ratio being the number of high quality segment sector measurements associated with each cell 181 , 182, 183 over the total number of the measurements associated with each same cell 181 , 182, 183, the N first ratios and N second ratios being associated to the half line 702. The method further comprises receiving a label for each half line 702, the label being 0 or 1 , 0 indicating that the direction of the target antenna 171 is not the same as the direction of the half line 702, 1 indicating that the direction of the target antenna 171 is the same as the direction of the half line 702. The method further comprises training the ML model based on the N first ratios and N second ratios of each half line 702 as inputted features and based on the received label for each half line 702 as the inputted target of the ML model.

[000109] The training phase actually have same steps as the inference phase, except for the receiving step and the training step. As shown in fig. 16A and 16B, labels 0 or 1 for each vertex are received from technician and being used as inputted target for training the ML model. The real direction of the target antenna is known in the training phase. In one alternative, for the half line which is the closest to the real direction of the target antenna, a label 1 is received. Labels 0 are received for other half lines. In another alternative, for the two half lines which are closest to the real direction of the target antenna, two labels 1 are received. Labels 0 are received for other half lines. The ML model are trained based on the half line features and the received labels.

[000110] As shown in fig. 17A and 17B, the machine learning method comprises a training phase and an inference phase. Referring to fig. 17A, the training phase comprises data preparation, half lines generation, feature extraction, training or retraining ML model and an optional post processing. The data preparation is corresponding to the step of obtaining the UE measurements and the optional step of discarding outliers. The half lines generation is corresponding to the step of determining the half lines. The feature extraction is corresponding to the step of determining segment sectors, the step of determining the number of segment sector measurements/the number of high quality segment sector measurements and the step of determining the N first ratios and the N second ratios. The training or retraining ML model is corresponding to the step of receiving label and the step of training the ML model, taking the received 0 or 1 labels as ground truth. An optional post processing can be added.

[000111 ] Similarly, referring to fig. 17B, the inference phase comprises data preparation, half lines generation, feature extraction, prediction using trained model and an optional post processing. The data preparation is corresponding to the step of obtaining 1202 the UE measurements and the optional step of discarding outliers. The half lines generation is corresponding to the step of determining 1204 the half lines. The feature extraction is corresponding to the step of determining 1206 segment sectors, the step of determining 1208 the number of segment sector measurements/the number of high quality segment sector measurements and the step of determining 1210 the N first ratios and the N second ratios. The prediction using trained model is corresponding to the step of estimating 1212 the direction of the target antenna. The optional post processing can be added.

[000112] By such an embodiment, a ML model can be used to estimate the direction of the target antenna. When the amount of the measurement becomes quite large, the ML model has significant advantages.

[000113] According to another embodiment, the target antenna can be any of the antennas of the N radio antennas 171 , 172, 173 which having the same location.

[000114] By this method, the direction of all the radio antennas 171 , 172, 173 can be estimated independently. Alternatively, when the angle differences between each two of the N radio antennas are known, the directions of other antennas can be calculated based on the angle differences, when the direction of the target antenna is estimated. The angle difference between each two of the N radio antennas are configurable.

[000115] According to another embodiment, the method can be performed multiple times for one target antenna 171 , each time determining different half lines 702, 710, the direction of the target antenna 171 being estimated based on the estimated direction of each time.

[000116] For example, for the first time, six half lines A, B, C, D, E, F are determined and the angles between every two half lines are the same, that is 60 degrees. After performing the method for the first time, the half line A is determined to have the N first ratios and N second ratios above thresholds. It means that the direction of the target antenna is near to the half line A. When performing the method for the second time, multiple half lines can be determined near the half line A, with smaller angle intervals, e.g., two half lines G, H are -20 degrees and -10 degrees away from the half line A, and two half lines I, J are +20 degrees and +10 degrees away from the half line A. No other half lines are determined. The second time of the method is performed based on the five half lines A, G, H, I, J. Therefore, a more accurate direction of the target antenna can be estimated. [000117] According to another embodiment, the direction of the target antenna is already known from network operator. However, the direction known from the operator may not be accurate. In this circumstance, when performing the method, taking the known direction as the first half line A. Based on the first half line A, determining other half lines in two directions away from the half line A, and only a small range of angles need to be covered, e.g., -15 degrees and +15 degrees. Therefore, the direction of the target antenna can be estimated accurately in a small range.

[000118] According to another embodiment, one or more network entities 160 is configured to operate in a wireless communication network 150, wherein N radio antennas 171 , 172, 173 are situated in the wireless communication network 150, N being an integer equal to or larger than 1 and the N radio antennas 171 , 172, 173 having one same position. The one or more network entities 160 is configured for estimating a direction of one radio antenna of the N radio antennas 171 , 172, 173, namely a target antenna 171 , each of the N radio antennas 171 , 172, 173 serves one cell 181 , 182, 183 which associates with a direction of the radio antenna, the N radio antennas 171 , 172, 173 serve N cells in total. Each of the N radio antenna 171 , 172, 173 wirelessly communicates with one or more user equipment UE 140-146, each UE 140-146 being served by one cell 181 , 182, 183, the one cell 181 , 182, 183 being the serving cell of the UE 140-146 and corresponding radio antenna 171 , 172, 173 being the serving radio antenna of the UE 140-146. Each UE 140-146 performs measurements on radio signals received from the serving radio antenna, each measurement comprises signal measurement and the location where the UE performs the measurement, each measurement being associated with the serving cell 181 , 182, 183 and the serving radio antenna 171 , 172, 173 of the UE 140-146 which performs the measurement. The one or more network entities 160 comprises a processing circuitry 803 and a memory 804, said memory 804 containing instructions executable by said processing circuitry 803. The one or more network entities 160 is operative for obtaining the measurements of each UE 140-146 and determining, for the target antenna 171 , one or more half lines 702, 710, each half line 702, 710 having one end in a position of the target antenna 171 , the other end extends away from the target antenna 171 position. The one or more network entities 160 is further operative for: for each half line 702, determining a segment sector 704, taking the half line 702 as the center line of the segment sector 704, taking a distance as the radius of the segment sector 704, and a central angle of the determined segment sector 704 being the same as the radiation angle of the target radio antenna 171 . The one or more network entities 160 is further operative for: for each determined segment sector 704, determining N numbers of segment sector measurements, each number comprises the measurements which are located in the segment sector 704 and being associated with one of the cells 181 , 182, 183, and out of the segment sector measurements determining a number of high quality segment sector measurements, the high quality segment sector measurements are the segment sector measurements having signal measurement higher than a threshold. The one or more network entities 160 is further operative for: for each determined segment sector 704, calculating N first ratios, each first radio being the number of segment sector measurements associated with each cell 181 , 182, 183 over the total number of the measurements associated with each same cell 181 , 182, 183, and calculating N second ratios, each second ratio being the number of high quality segment sector measurements associated with each cell 181 , 182, 183 over the total number of the measurements associated with each same cell 181 , 182, 183, the N first ratios and N second ratios being associated to the half line 702. The one or more network entities 160 is further operative for estimating the direction of the target antenna 171 based on the one or more half lines 702 which having N first ratios and N second ratios over corresponding ratio thresholds.

[000119] According to another embodiment, the estimating of the direction of the target antennas 171 further comprises: for each half line 702, determining a probability ranging from 0 to 1 , based on the N first ratios and the N second ratios of each half line 702, the probability indicating the probability of the direction of the target antenna 171 being the same as the half line 702, and estimating the direction of the target antenna 171 based on one or more half lines which having the probability higher than a probability threshold. [000120] According to another embodiment, one or more network entities 160 is further being configured to perform a machine learning, ML, method using a ML model, the ML method comprising a training phase followed by an inference phase. In the inference phase, the one or more network entities 160 is configured to perform the steps above, wherein the probability of each half line 702 is determined by the ML model, taking the N first ratios and the N second ratios of each half line 702 as the inputted features of the ML model, the determined probability of each half line 702 being a prediction result of the ML model. The training phase is performed before the inference phase and comprises obtaining the measurements of each UE 140-146 and determining, for the target antenna 171 , one or more half lines 702, 710, each half line 702, 710 having one end in a position of target antenna 171 , the other end extends away from the target antenna 171 position. The training phase further comprises: for each half line 702, determining a segment sector 704, taking the half line 702 as the center line of the segment sector 704, taking a distance as the radius of the segment sector 704, and a central angle of the determined segment sector 704 being the same as the radiation angle of the target antenna 171 . The training phase further comprises: for each determined segment sector 704, determining N numbers of segment sector measurements, each number comprises the measurements which are located in the segment sector 704 and being associated with one of the cells 181 , 182, 183, and out of the segment sector measurements determining a number of high quality segment sector measurements, the high quality segment sector measurements are the segment sector measurements having signal measurement higher than a threshold. The training phase further comprises: for each determined segment sector 704, calculating N first ratios, each first radio being the number of segment sector measurements associated with each cell 181 , 182, 183 over the total number of the measurements associated with each same cell 181 , 182, 183, and calculating N second ratios, each second ratio being the number of high quality segment sector measurements associated with each cell 181 , 182, 183 over the total number of the measurements associated with each same cell 181 , 182, 183, the N first ratios and N second ratios being associated to the half line 702. The training phase further comprises receiving a label for each half line 702, the label being 0 or 1 , 0 indicating that the direction of the target antenna 171 is not the same as the direction of the half line 702, 1 indicating that the direction of the target antenna 171 is the same as the direction of the half line 702, and training the ML model based on the N first ratios and N second ratios of each half line 702 as inputted features and based on the received label for each half line 702 as the inputted target of the ML model.

[000121] According to another embodiment, the one or more network entities 160 are further operative for grouping the obtained measurements which are associated with the N cells 181 , 182, 183, based on the locations of the measurements and a group distance threshold, and discarding the measurements of one or more groups before determining the number of segment sector measurements and the number of high quality segment sector measurements, the measurements in the discarded one or more groups having an average distance from an average position of the measurements which belong to the group having the largest number of measurements, the average distance is higher than a discarding distance threshold.

[000122] According to another embodiment, the target antenna can be any of the antennas of the N radio antennas 171 , 172, 173 which having the same position.

[000123] According to another embodiment, the one or more network entities 160 being operative for performing the steps multiple times for one target antenna 171 , each time determining different half lines 702, 710, the direction of the target antenna 171 being estimated based on the estimated direction of each time.

[000124] According to other embodiments, referring to fig. 18, the network entity 160 may further comprise a communication unit 802, which may be considered to comprise conventional means for wireless communication with other devices, such as a transceiver for wireless transmission and reception of signals. The instructions executable by said processing circuitry 803 may be arranged as a computer program 805 stored e.g. in said memory 804. The processing circuitry 803 and the memory 804 may be arranged in a sub-arrangement 801 . The subarrangement 801 may be a micro-processor and adequate software and storage therefore, a Programmable Logic Device, PLD, or other electronic component(s)/processing circuit(s) configured to perform the methods mentioned above. The processing circuitry 803 may comprise one or more programmable processor, application-specific integrated circuits, field programmable gate arrays or combinations of these adapted to execute instructions.

[000125] The computer program 805 may be arranged such that when its instructions are run in the processing circuitry, they cause network entity 160 to perform the steps described in any of the described embodiments of the network entity 160 and its method. The computer program 805 may be carried by a computer program product connectable to the processing circuitry 803. The computer program product may be the memory 804, or at least arranged in the memory. The memory 804 may be realized as for example a RAM (Randomaccess memory), ROM (Read-Only Memory) or an EEPROM (Electrical Erasable Programmable ROM). In some embodiments, a carrier may contain the computer program 805. The carrier may be one of an electronic signal, an optical signal, an electromagnetic signal, a magnetic signal, an electric signal, a radio signal, a microwave signal, or computer readable storage medium. The computer-readable storage medium may be e.g. a CD, DVD or flash memory, from which the program could be downloaded into the memory 804. Alternatively, the computer program may be stored on a server or any other entity to which the network entity 160 has access via the communication unit 802. The computer program 805 may then be downloaded from the server into the memory 804.

[000126] Although the description above contains a plurality of specificities, these should not be construed as limiting the scope of the concept described herein but as merely providing illustrations of some exemplifying embodiments of the described concept. It will be appreciated that the scope of the presently described concept fully encompasses other embodiments which may become obvious to those skilled in the art, and that the scope of the presently described concept is accordingly not to be limited. Reference to an element in the singular is not intended to mean "one and only one" unless explicitly so stated, but rather "one or more." Further, the term “a number of”, such as in “a number of wireless devices” signifies one or more devices. All structural and functional equivalents to the elements of the above-described embodiments that are known to those of ordinary skill in the art are expressly incorporated herein by reference and are intended to be encompassed hereby. Moreover, it is not necessary for an apparatus or method to address each and every problem sought to be solved by the presently described concept, for it to be encompassed hereby. In the exemplary figures, a broken line generally signifies that the feature within the broken line is optional.