Technical Field

The present disclosure relates to a technique for estimating a Time of Arrival (TOA) for signals provided by one or more cells of a cellular telecommunications network. More specifically, and without limitation, a method of estimating the TOA based on Positioning Reference Signals (PRSs) sent by a cell of the telecommunications network and a device therefor are provided.

Background

There are various reasons for determining the location of a mobile device wirelessly connected to a telecommunications network. For example, a network operator may provide location-based services to subscribers, such as indicating restaurants in a vicinity of the mobile device, upcoming gas stations or any other information depending on the current location of the mobile device. Furthermore, some jurisdictions require the network operator to determine the location of the mobile device when an emergency call is placed from the mobile device. Moreover, some location tracking systems include mobile devices accessing the telecommunications network to track the location of vehicles, such as freight trains, cars and trucks for fleet operations.

Many location techniques, e.g., multilateration, are at least partly based on a difference of the TOA, as observed at the mobile device. Observed Time Difference of Arrival (OTDOA) for positioning has been introduced as a Long Term Evolution service by 3GPP in Release 9 and has been further specified in Release 10 (cf. 3GPP TS 36.211, V10.7.0, Sections 6.10.4, 7.2; 3GPP TS 36.355, VIO.11.0, Section 6.5; and 3GPP TS 36.133, V10.12.0, Sections 8.1, 9.1, A.8.12 to A.8.14, A.9.8). The mobile device, which is also referred to as User Equipment (UE) in LTE

implementations, is required to measure the TOA of radio signals received from cells, for which the telecommunications network has requested a TOA measurement. The requested cells send Positioning Reference Signals (PRSs) to assist the TOA measurements, from which Reference Signal Time Difference (RSTD) values are derived. The mobile device is further required to report the RSTD values measured for the requested cells to the telecommunications network. The range of timing misalignment allowed for synchronous cells is ±500 ps

(cf. 3GPP TS 36.355, VIO.11.0, Section 6.5), i.e., the TOA estimation range exceeds the length of a Cyclic Prefix (CP), which is in the range of 5.2 ps to 33.3 \ S for LTE implementations. The expected time of arrival of a signal from a cell to the UE is signaled by the telecommunications network using a dedicated Information Element (IE) parameter "expectedRSTD". The uncertainty range of aforementioned parameter is signaled by the telecommunications network using another IE parameter

"expectedRSTD-Uncertainty" (cf. 3GPP TS 36.355, V10.110, Section 6.5.1.2).

A technique for estimating the TOA, when the TOA estimation range is not bounded to the CP length, was suggested in document WO 2011/141403 A. One of the embodiments described therein suggests determining two or more timing hypotheses and combining a Power Delay Profile (PDP) associated with each of the hypotheses in order to capture a maximum in the combined PDP for deriving the TOA. However, cell timing information can be lost when two or more PDP hypotheses are combined.

Summary

Accordingly, there is a need for a technique that allows accurately estimating a TOA with low complexity also when a timing misalignment between cells exceeds the CP length.

According to one aspect, a method of estimating a Time of Arrival (TOA) based on Positioning Reference Signals (PRSs) sent by a cell of a cellular telecommunications network is provided. The method comprises a step of selecting a Power Delay Profile (PDP) out of a plurality of PDPs determined based on the PRSs received from the cell; and a step of estimating the TOA based on the selected PDP.

By selecting the PDP, a range (also referred to as search window) that exceeds the CP length can be covered in at least some implementations of the technique. The search window may be covered by the plurality of PDPs.

A mobile device, e.g., in wireless communication with the telecommunications network, may perform the method. Determining the plurality of PDPs may include deriving bounds for a search window in a number of OFDM symbols based on the "expected RSTD" and "expected RSTD- Uncertainty" parameters received from the telecommunications network.

The PDP may be selected based on a selection criterion. The selection criterion may be computed for each of the plurality of PDPs. The PDP may be selected based on a Peak-to-Average Ratio (PAR), e.g., a peak-to-average power ratio.

The plurality of PDPs may be determined for different first time offsets. The first time offset may be defined relative to a reference cell. The reference cell may be a cell other than a serving cell. Alternatively or in addition, the reference cell may be a cell other than the cell from which the PRSs (used for determining the plurality of PDPs) are received. E.g., a neighboring cell other than a present candidate cell (for which the plurality of PDPs is determined) may function as the reference cell. The communication may include Orthogonal Frequency-Division Multiplexing (OFDM) symbols. The mobile device may receive OFDM symbols from the

telecommunications network. The first time offset may be integer multiples of a length of the OFDM symbols. E.g., the first time offset may assume the values -2, -1, 0, +1 and +2 in units of the OFDM symbol length.

An OFDM frame or subframe structure may be defined and/or synchronized with the reference cell. The reference cell may be a cell serving the mobile device. The OFDM frame or subframe structure may be used for receiving the PRSs from two or more cells other than the reference cell.

Each of the plurality of PDPs may be determined based on PRSs received and/or extracted from different OFDM symbols. For example, the PRSs may be received in PRS occasions. Each PRS occasion may include a plurality of Resource Elements (REs), e.g., in one or more OFDM symbols. The PDP determination may include extracting the PRSs at predetermined REs, e.g. according to a subset of a time- frequency-grid. The PRSs extraction may be shifted according to the first time offset.

Radio signals received at the mobile device may be stored. At least some of the received OFDM symbols may be subjected to a Fourier transformation. The received radio signals may be stored in a buffer, e.g., in a frequency domain representation, optionally following the Fourier transformation in an OFDM receiver. The received radio signals may be stored in the form of In-Phase and Quadrature components (also referred to as an IQ-signal), e.g., for each Resource Element (RE) or subcarrier.

Determining the plurality of PDPs may include generating a PRS sequence, e.g. for each OFDM symbol, as sent by the cell.

Each of the plurality of PDPs may be determined based on a Channel Impulse

Response (CIR). For example, the PDP may be a sequence of power or absolute values of a sequence of complex values representing the CIR. A Channel Frequency Response (CFR) may be computed based on the generated PRS sequence and the received OFDM symbols (e.g. the PRS extracted from the received OFDM symbols). The CIR may be computed from the CFR. The CFR may be computed in the

frequency domain. The computed CFR may be transformed to the time domain, resulting in the CIR.

Determining the plurality of PDPs may include a first averaging of the CFR or the CIR taking phase information into account. The first averaging may include averaging the CFR or the CIR over one or more PRS symbols of a PRS occasions. A second averaging may include averaging the power or absolute values of the averaged CIR.

A spectral resolution of the extracted PRS and/or the CFR may be less than a spectral width of the RE including one of the PRSs in the received OFDM symbol.

A second time offset may be determined by comparing the highest PAR and the second highest PAR. The closer the PAR value of the second highest PAR is to the PAR value of the highest PAR, the greater the second time offset. The second time offset may be zero, if the second highest PAR is significantly less than the highest PAR. When the second highest PAR is close to the highest PAR, the second time offset may assume a maximum for the second time offset. The maximum for the second time offset may be a down-sampling factor of the PRS set compared to the system sampling rate.

A third time offset may be determined based on a peak position indicated by the selected PDP.

The TOA may be estimated by adding up at least two of the first time offset, the second time offset and the third time offset. According to another aspect, a computer program product is provided. The computer program product comprises program code portions for performing one or more of the steps of the method aspect described herein, e.g., when the computer program is executed on one or more computing device. The computer program product may be stored on a computer-readable recording medium, such as a permanent or rewritable memory. The computer program product may also be provided for download via a data network, e.g., the telecommunications network and/or the Internet.

According to a hardware aspect, a device for estimating a Time of Arrival (TOA) based on Positioning Reference Signals (PRSs) sent by a cell of a cellular

telecommunications network is provided. The device comprises a selecting unit adapted to select a Power Delay Profile (PDP) out of a plurality of PDPs determined based on the PRSs received from the cell; and an estimating unit adapted to estimate the TOA based on the selected PDP.

The device may be implemented at a mobile device in wireless communication with the telecommunications network.

The device may further comprise any feature disclosed in the context of the method aspect. E.g., any one of the units, or a dedicated unit, may be adapted to perform any one of the steps described in the context of the method aspect.

Brief Description of the Drawings

In the following, the present disclosure is described in more detail with reference to exemplary embodiments illustrated in the drawings, in which:

Fig. 1 schematically illustrates a cellular telecommunications network including a plurality of cells and a mobile device in wireless communication with the telecommunications network;

Fig. 2 schematically illustrates a block diagram of a device for estimating a

Time Of Arrival (TOA) based on Positioning Reference Signals (PRSs) sent by at least some of the cells of Fig. 1; shows a flowchart for a method of estimating a TOA based on PRSs sent by at least some of the cells of Fig. 1; schematically illustrates a block diagram of an embodiment for the device of Fig. 2; shows a flowchart for a method of controlling and configuring the device of Fig. 2 or 4; shows a flowchart for a method of deriving size and bounds for a search window implementable in as a substep in the method of Fig. 3 or 5; shows a flowchart for a method of determining a Power Delay Profile (PDP) implementable as a substep in the method of Fig. 3 or in a unit of the device of Fig. 4; schematically illustrates a time-frequency grid including Resource

Elements allocated to PRSs; shows a flowchart for an implementation of the method of Fig. 3 that can be performed by units of the device of Fig. 4; and schematically illustrates time offsets between a reference cell and a neighboring cell for determining a second time offset in the method of Fig. 9.

Detailed Description

In the following description, for purposes of explanation and not limitation, specific details are set forth, such as specific device configurations and specific Resource Element (RE) structures in order to provide a thorough understanding of the technique disclosed herein. It will be apparent to one skilled in the art that the technique may be practiced in other embodiments that depart from these specific details. Moreover, while the following embodiments are primarily described for a positioning service according to Long Term Evolution (LTE) and LTE-Advanced, it will readily apparent that the technique may also be practiced in the context of other cellular telecommunications systems including the Global System for Mobile

Communications (GSM) and the Universal Mobile Telecommunications System

(UMTS).

Those skilled in the art will further appreciate that methods, steps, functions and units explained herein may be implemented using individual hardware circuitry, using software functioning in conjunction a programmed microprocessor or a general purpose computer, using an Application Specific Integrated Circuit (ASIC), using a Field Programmable Gate Array (FPGA) and/or using one or more Digital Signal Processor (DSPs). It will also be appreciated that, while the following embodiments are primarily described in the form of methods and devices, the technique disclosed herein may also be embodied in a computer processor and memory coupled to the processor, wherein the memory stores one or more programs that perform the steps discussed herein when executed by the processor.

Fig. 1 schematically illustrates an exemplary telecommunications network 100 including a plurality of cells 102, 104, 106, 108 and 110 with associated base stations 112, 114, 116, 118 and 120, respectively. A mobile device 130 is wirelessly connected to the cell 110. The cell 110 serving the mobile device 130 is an example for a reference cell. The mobile device 130 receives radio signals including

Positioning Reference Signals (PRSs) from three or more of the cells 102 to 110.

In an exemplary LTE implementation, the PRSs are transmitted by each of the base stations 112 to 120 during designated PRS occasions. PRS occasions typically occur once every 160 to 1280 subframes (i.e., every 160 to 1280 ms). Each PRS occasion comprises 1, 2, 4 or 6 consecutive subframes. The mobile device 130 measures a Time of Arrival (TOA) based on the PRSs from different cells. The mobile device 130 reports the TOA measured for each cell to the telecommunications network 100. A location server in the telecommunications network 100 uses the TOA measurements to determine the location of the mobile device 130.

Fig. 2 shows a block diagram of a device 200 for estimating a TOA based on PRSs sent by one or more cells of a cellular telecommunications network. The device 200 includes a selecting unit 202 for selecting a Power Delay Profile (PDP). An estimating unit 204 is configured for estimating the TOA based on the selected PDP.

The device 200 can be implemented in the mobile device 130. The PRSs can be sent by some or all of the cells of the telecommunications network 100.

Fig. 3 shows a flowchart of a method 300 of estimating a TOA based on PRSs sent by one or more cells of a cellular telecommunications network. The method 300 comprises a step 302 of selecting a PDP out of a plurality of PDPs determined based on PRSs received from the one or more cells of the telecommunications network. In a step 304 of the method 300 the TOA is estimated based on the selected PDP.

The mobile device 130 in wireless communication with the telecommunications network 100 can perform the method 300. E.g., the steps 302 and 304 can be performed by the units 202 and 204, respectively.

An exemplary implementation of the method 300 further includes determining a search window for a cell of interest. Size and bounds of the search window are determined. E.g., the size and the bounds are determined in terms of samples and as a number of Orthogonal Frequency-Division Multiplexing (OFDM) symbols,

respectively, e.g., relative to a frame or subframe structure of the reference cell 110.

The selecting step 302 may include de-noising the OFDM symbols, e.g. by means of coherent accumulation. For the cell of interest, a PDP is determined for each of the de-noised OFDM symbols.

The REs allocated to PRSs are grouped in PRS sets. The number of PRS sets is also referred to as PRS cardinality. In the frequency domain, each PRS set spans over a number of subcarriers, e.g., 6 subcarriers or half the size of a Resource Block (RB). Each PRS set is used for determining a frequency component of a Channel Impulse Response (CIR). The PDP is determined based on an inverse Fourier transformation (e.g., an Inverse Fast Fourier Transformation or IFFT) of the CIR. For an exemplarily bandwidth of 50 RBs, 100 frequency components for the CIR are determined. Zero- Padding results in 128 frequency components suitable for the IFFT.

Since the PRS cardinality is less than the number of subcarriers, e.g., 1200

subcarriers compared to an IFFT size of 128, the CIR is down-sampled compared to a system sampling rate. Herein, "down-sampled" may refer to a number of distinct frequency components that is less than the number of subcarriers used for the communication, e.g. for the downlink, of the mobile device 130.

Binning in the frequency domain, e.g., by using x=6 subcarriers for determining one frequency component of the CIR, corresponds to a periodic PDP or shortened PDP compared to the OFDM symbol length in the time domain. E.g., a PDP extending over a sixth of the OFDM symbol length may include the PDP information for the corresponding OFDM symbol. The PDP may be periodic, e.g., including x=6 periods, over the OFDM symbol length. The periodic structure can be used for determining at least one of a fine sample timing (as a second time offset) and a coarse sample timing (as a third time offset). The factor x is also referred to as a down-sampling factor.

All PDPs are analyzed for a best peak, e.g., the highest peak in the PDP in units of an average value of the PDP. The PDP having the best peak is selected. The selection provides a first time offset, e.g., with a temporal resolution corresponding to the OFDM symbol length or a temporal granularity corresponding to the OFDM symbol structure. The first time offset is also referred to as OFDM symbol timing.

For example, the PDP having the best peak gives explicit timing information at the OFDM symbol boundaries within a subframe in relation to the serving cell 110. Since the timing information is relative to the serving cell 110, the timing information is also referred to as a derived timing difference. The first time offset is recorded in association with the corresponding cell of interest, e.g., as a number of the corresponding OFDM symbol within the subframe.

Optionally, a second time offset (also referred to as coarse sample timing) and/or a third time offset (also referred to as fine sample timing) are estimated for the OFDM symbol with the best peak in the corresponding PDP. The second time offset is computed, only if a sample timing residue is greater than the cardinality of the PRS set in terms of the system sampling rate. The sample timing residue may be defined as the offset (e.g., in terms of the system sampling rate) of the PDP peak within the selected OFDM symbol.

For example, the second time offset is computed, only if the cardinality is reduced by a factor x, and the sample timing residue is greater than the OFDM symbol length divided by the factor x. The second time offset can be represented by an integer value, e.g., as the sample timing residue divided by cardinality (e.g., the OFDM symbol length divided by the factor x) and rounded off:

(second time offset) = (sample timing residue) div (cardinality),

The third time offset corresponds to the sample timing residue that fits into the PRS cardinality in terms of the system sampling rate. In other words, the third time offset is based on an estimation of a position of the peak within a time interval having the length of the OFDM symbol length divided by the factor x:

(second time offset) = (sample timing residue) mod (cardinality),

The TOA is computed as the sum of the first time offset (i.e., the OFDM symbol timing), the second time offset (i.e., the coarse sample timing, e.g., multiplied by the cardinality when represented by above integer value) and the third time offset (i.e., the fine sample timing).

Fig. 4 schematically illustrates a top-level block diagram of an embodiment of the device 200. The device 200 can be implemented as an Observed Time Difference Of Arrival (OTDOA) positioning subsystem for the physical layer (also referred to as layer 1 or LI) at the side of the mobile device 130. The OTDOA positioning subsystem is initiated by the telecommunications network 100 that transmits LTE Positioning Protocol (LPP) Information Elements (IEs) at certain times (cf. 3GPP TS 36.355, VIO.11.0, Section 6.5). The LPP IEs enable a Control and Configuration unit 402 to initialize the OTDOA capability of the mobile device 130. The Control and Configuration unit 402 configures a scheduler unit 404, a PDP estimation unit 406 and a Reference Signal Time Difference (RSTD) estimation unit 408. The RSTD estimation unit 408 performs the RSTD measurement of candidate cells configured by the Control and Configuration unit 402. The scheduler unit 404 calls, e.g. in response to the configuration, the PDP estimation unit 406 for processing of the PRS occasions that are transmitted within PRS subframes in the RSTD reporting interval (cf. 3GPP TS 36.133, V10.12.0, Section 8.1, 9.1, A.8.12 to A.8.14, A.9.8).

For each PRS occasion, IQ samples are received and stored after a Fourier transformation to the frequency domain, e.g. in a post-FFT IQ-sample buffer unit 410 (cf. 3GPP TS 36.211, V10.7.0, Sections 6.10.4, 7.2). The IQ sample buffer unit 410 can be implemented as part of an OFDM receiver at the mobile device 130.

The OFDM receiver remains synchronized with the serving cell 110, e.g., also during receiving the PRSs (e.g., from the cell 102) and/or estimating the TOA. Here, the serving cell 110 is assumed to be the reference cell. If the reference cell is not the serving cell, the device 200 assigns the serving cell to become the reference cell, optionally subject to the condition that the serving cell is on a list of candidate cells provided (e.g., via the Control and Configuration unit 402) by the

telecommunications network 100.

The PDP estimation unit 406 determines and accumulates the PDP results based on the IQ-sample buffer unit 410 for each of the candidate cells.

At a later stage, e.g. in the estimating step 304, the PDP estimation unit 406 performs a post-processing of the accumulated PDPs in order to produce the final PDP report for each cell. Moreover, e.g. in the estimating step 304, the RSTD estimation unit 408 is triggered for processing the PDP resulting in the TOA estimation and/or corresponding RSTD values. The TOA and/or RSTD results are passed to higher layers and/or conveyed back to the telecommunications network 100 by the mobile device 130, e.g. in response to a positioning request.

Fig. 5 illustrates an exemplary method 500 for controlling and configuring the positioning subsystem in the mobile device 130. Only LPP parameters that are most relevant for the embodiment 200 of Fig. 4 and corresponding steps for the method 300 are shown in Fig. 5. A more comprehensive list of LPP parameters specified in document 3GPP TS 36.355, VIO.11.0, Section 6.5. The method 500 may be performed in advance of the step 302 and 304. The method 500 may be

implemented in the unit 402.

The LI LPP parameters are extracted in a step 502 from the OTDOA positioning IEs sent by the telecommunications network 100, e.g., via the serving cell 110. The extracted parameters include, e.g., a PRS configuration index, I _PRS . A PRS occasions periodicity, T _PRS , and a number of PRS occasions, N _PRS , are derived from I _PRS in a step 504. The resulting T _PRS is in the range from 160 ms to 1280 ms (cf. 3GPP TS 36.211, V10.7.0, Sections 6.10.4, 7.2). N _PRS further dependents on other network

configurations. The resulting N _PRS is in the range from 8 to 32 (cf. 3GPP TS 36.133, V10.12.0, Section 8.1, 9.1, A.8.12 to A.8.14, A.9.8).

The configuration parameters are passed to the scheduler unit 404 that triggers in return the PDP estimation unit 406 every T _PRS for processing of the post-FFT IQ- samples buffered in the unit 410. In a step 506, the number of candidate cells, NCELLS, requested by the telecommunications network 100 in the IEs is extracted. The resulting NCELLS is in the range from 3 to 72 (cf. 3GPP TS 36.355, VIO.11.0,

Section 6.5). The unit 402 passes the parameter to the PDP estimation unit 406 and the RSTD estimation unit 408. The units 406 and 408 process the requested number of candidate cells.

In a step 508, the size of the search window in samples of the system sampling rate, T _s , and the bounds in a number of OFDM symbols of the search window are derived for neighbor cells based on cell-specific LPP parameters. In an exemplary LTE implementation, the cell-specific LPP parameters include an expected RSTD

("expectedRSTD") and an expected uncertainty of the RSTD ("expectedRSTD- Uncertainty").

The search range (e.g., the search window) includes the value expectedRSTD and extends to earlier times and to later times up to ±0.1 ms, respectively (3GPP TS 36.355, VIO.11.0, Section 6.5). The span of the search window is extended to include complete OFDM symbols. E.g., the length of the search window includes (almost) 3 or up to 5 consecutive OFDM symbols.

The parameters derived for the search window in the step 508 and/or by the unit 402 are communicated to the PDP estimation unit 406 and/or the RSTD estimation unit 408.

Fig. 6 illustrates an exemplary implementation of the step 508 of deriving the search window size and bounds. For each RSTD reporting interval, the size and bounds are calculated and remain unchanged during this interval. After deriving the expected RSTD values in a step 602 (e.g., using the commands "GET expectedRSTD" and "GET expectedRSTD-Uncertainty"), the search window size is calculated in a step 604 according to a formula given in the clause

"expetdedRSTD-Uncertainty" of document 3GPP TS 36.355, VIO.11.0, Section 6.5. Based on a search window size derived in the step 604, the search window bounds are derived in a number of OFDM symbol in step 606. The step 606 includes extending (also referred to as "bounding") the search window size on its left-hand side and right-hand side to the end and to the start of the neighboring OFDM symbols, respectively.

The extended search window is shifted (e.g., translated in time) to different OFDM symbols according to the first time offset. Based on the derived bounds, indices representing the first offset relative to the serving cell 110 are derived for the delayed OFDM symbols of each cell in the list of neighboring cells in a step 608. It can be shown, at least for an exemplary LTE implementation, that a maximum number of OFDM symbols to be searched is 5 for a given (valid) range of

expectedRSTD-Uncertainty (cf. 3GPP TS 36.355, VIO.11.0, Section 6.5). Moreover, valid relative OFDM symbol indices are in the range of [-5, 4] at all times for the exemplary LTE implementation. In a preferred implementation, the shifted search window is contiguous and/or relative OFDM symbol indices that bound the search window always stay in ascending order.

The search process is repeated for all cells in the list of neighboring cells according to branching point 610. Finally, the search window bounds and respective relative OFDM symbol indices are passed to the rest of the subsystem in a step 612.

Fig. 7 illustrates an exemplary method 700 for determining a plurality of PDPs for one candidate cell. For each cell out of a list of candidate cells, the method 700 can be repeated. The method 700 can be implemented as a substep of the step 302.

In a step 702 of the method 700, a PRS sequence is generated for each OFDM symbol in a PRS subframe including a RE allocated to a PRS according to 3GPP TS 36.211, VIO.7.0, Section 7.2 (also referred to as a PRS symbol). The PRS sequence is optionally stored for future re-use, subject to the number of PRS symbols in the search window.

The received PRSs are extracted in a step 704 for each PRS symbol in the PRS subframe or PRS occasion. A PRS channel estimation, e.g., a Channel Frequency Response (CFR) is computed in a step 706. The PRS channel estimation is computed from the PRS sequence and the extracted PRS. For example, the computation includes a point-wise multiplication of the extracted PRSs and the (complex-conjugated) PRS sequence.

The PRS channel estimation results are coherently accumulated in a step 708. E.g., all PRS channel estimation results are coherently accumulated for PRSs extracted from a coherent sub-block. Fig. 8 shows a time-frequency grid of an exemplary coherent sub-block 800. Time is shown on a horizontal axis in units of OFDM symbols. Frequency is shown on a vertical axis in units of subcarriers. REs allocated to a PRS are marked by "R6 . Dimensions of the coherent sub-block depend on the system coherence time and coherence bandwidth. In the example illustrated in Fig. 8, the time dimension equals one subframe and the frequency dimension equals half of one Resource Block (RB). In an advanced embodiment, the dimensions are configured based on the Channel State Information (CSI).

By means of coherent accumulation, a de-noising effect is achieved of the PRS channel estimation results. A coherent sub-block configured to the half size of one RB is beneficial for an Extended Typical Urban (ETU) channel, e.g., with low Doppler frequencies.

In a variant of the method 700, the extracted PRSs are coherent accumulation prior to computing the PRS channel estimation. In other words, steps 706 and 708 may be interchanged.

The steps 702 to 708 are repeated for all PRS OFDM symbols from the coherent sub- block 800 according to branching point 710.

The de-noised channel estimation results are transformed to the time domain in a step 712. For example, the CFR is subjected to an Inverse Fourier Transformation (IFFT) resulting in a Channel Impulse Response (CIR).

The CIR is non-coherently accumulated in a step 714 in order to re-construct the average propagation channel CIR that is also referred to as PDP. This process is repeated for all configured PRS occasions and respective coherent sub-blocks 800 according to branching point 716. From each PDP, a PDP maximum, a PDP index at which the PDP maximum is assumed (or interpolated) and a PDP average are calculated and stored in steps 718, 720 and 722, respectively. The steps 702 to 722 are cell-specific. The steps 702 to 722 are repeated multiple times depending on the number of OFDM symbols in the search window, e.g., for the different hypotheses for the first time offset, according to branching point 724.

The PDP, the PDP maximum, the PDP maximum index and the PDP average are passed to the rest of the subsystem for further processing in a step 726.

Fig. 9 illustrates an exemplary method 900 for TOA and RSTD estimations for one candidate cell. For each cell in a list of candidate cells, the method 900 can be repeated.

A PDP quality metric is evaluated in a step 902 for a given PDP, e.g., by calculating a Peak-to-Average Ratio (PAR). The calculation step 902 is repeated for each of the plurality of PDPs according to branching point 904. The calculation step 902 may be based on the steps 718 to 722.

The PDP quality metric is indicative of how distinct the peak power is for the given PDP. E.g., The PAR is calculated for all PDPs belonging to the search window. A PDP with the best PAR quality metric is selected in a step 906.

At least the steps 902 to 906 can be implemented as substeps of the step 302.

The selected PDP is further used for the TOA sample timing estimation, e.g., estimating the second and third time offsets. In an additional processing step 908, the PDP quality post-estimated. The step 908 includes determining whether or not to continue with the given cell, e.g., subject to the PDP quality. The PAR may serve as the quality metric also used as the quality criterion in the step 908. Alternatively, a different quality metric is calculated in step 908.

Cells that do not pass the quality criterion or criteria are discarded and/or marked to the higher layers as not present. Otherwise, processing is continued in a step 910 of deriving the first time offset, e.g., the best PDP OFDM Symbol Timing that returns a TOA estimation at OFDM symbol boundaries including the CP. If the search window is restricted to the CP length, only one PDP is required for the given cell. Therefore, further processing is skipped by branching point 912 to a step 918 of derive the third time offset, e.g., a fine sample timing. The fine symbol timing is a function of the PRS symbol mapping to the OFDM symbol, the cardinality of the PDP set, the system bandwidth and the size of the coherent sub-block.

If the search window is beyond the CP length, a second-best PDP is selected in a step 914. The best and second-best PDPs serve as an aid for deriving the second time offset, e.g., the coarse TOA sample timing estimation, which is carried out in a step 916.

Due to the cardinality of the PDP set being less than the system FFT size of, e.g., N = 2048, the second time offset indicates how many times the PDP repeats itself in terms of the system sampling rate N, i.e., within one OFDM symbol length. It can be shown that in an exemplary system with 50 PRS RBs, with the coherent sub-block size being half of one RB and with the PDP cardinality set being 128, the PDP maximum size in terms of the system sampling rate is Q = 341. A corresponding recalculation coefficient from the PDP cardinality is 2.(6). For the exemplary configuration, the number of repetitions k of Q can range from 0 to 2, to the left or to the right of the reference sampling point, depending on a time misalignment direction.

The number of repetitions themselves can be derived based on the power levels in the second-best PDP in relation to the power levels in the best PDP. If the second- best PDP peak power (e.g., represented by PAR ₂ -1 wherein PAR ₂ is the PAR of the second-best PDP) is not present at all (e.g., close to zero) or significantly smaller compared with the best PDP peak power (e.g., represented by PARi-1 wherein PARi is the PAR of the best PDP), it is assumed that only an insignificant part of the PRS OFDM symbol overlaps with the 2nd repetition (i.e., only an insignificant part of the the PRS OFDM symbol exceeds Q), as is schematically illustrated in Fig. 10(a). In Fig. 10, a reference cell (RC) timing is shown at reference sign 1002 and a neighbor cell (NC) relative timing is shown at reference sign 1004.

For the small time delay of Fig. 10(a) or similar relatively small time delays, the second offset time, k, is set to 0. If the power difference is in the range of small to medium, the second time offset is set to k = 1, correspond to a symbol timing schematically illustrated in Fig. 10(b).

If the power levels in the second-best PDP are significant (but still less than the power levels in the best PDP), the second time offset is set to k = 2, corresponding to a symbol timing as schematically illustrated in Fig. 10(c).

The second time offset (which is also referred to as a coarse TOA sample timing) is optionally returned in units of the system sampling rate, e.g., as (k * Q).

The overall TOA sample timing result, i.e., the second and third time offsets together, is always in a range of [-1024, 1023] samples for the discussed exemplary config ration.

Following the fine sample timing calculation in the step 918, the TOA is calculated in a step 920 as the sum of three summands:

TOA = (OFDM Symbol Timing) + (Coarse Sample Timing) + (Fine Sample Timing).

The TOA result can be positive or negative, subject to the time misalignment direction of the reference cell 110 and the neighbor cell 102.

The RSTD is obtained from the TOA in a step 922 by computing the RSTD according to document 3GPP TS 36.214, VIO.1.0, Section 5.1.12.

The OTDOA measurement quality is calculated in a step 924, e.g., according to clause "OTDOA-MeasQuality" in 3GPP TS 36.355, VIO.11.0, Section 6.5. The RSTD is reported for all candidate cells to the higher layers in a step 926.

At least some of the steps 910 to 926 can be implemented as substeps of the step 304.

As has become apparent from above description of exemplary embodiments, the technique disclosed herein allows determining a search window size for one or more cells of interest. Furthermore, the technique allows defining appropriate search window bounds. A cell timing information, which is conventionally lost, can be used for accurately estimating the TOA. For example, a Signal-to-Noise Ratio of a PDP may be greater compared with a conventionally combined PDP, resulting in a more accurate TOA estimation. At least some of the embodiments allow accurately estimating the TOA, even if a cell timing misalignment exceeds the CP length.

Same or other embodiments dynamically and/or explicitly determine a required search window size in terms of samples and search window bounds as a number of OFDM symbols, e.g. for each cell of interest and/or derived from control signaling of the telecommunications network. The control signaling may be compatible with, e.g. an expected RSTD and/or an expected RSTD uncertainty.

The technique can be implemented with low computational complexity, e.g., based on a common and/or not-repeated Fast Fourier Transform (FFT) of a given OFDM symbol. The OFDM symbol can be received according to a synchronization with a serving cell, so that the FFT output for the received OFDM symbol can be reused for estimating TOAs for all other cells of interest.

At least some of the embodiments ensure a wide operational range, e.g., in fading environments. The technique can be implemented so as to cover both channel propagation components (e.g., a line-of-site or LOS component) and non-line-of-site (NLOS) components.

An accurate TOA estimation can be obtained from a multistep implementation of the technique, which includes a sum of, at most, three summands even in a worst-case scenario. In at least some embodiments, an OFDM sample timing summand determines the TOA within the subframe at the OFDM symbol boundaries. A coarse sample timing summand is present, if a sample timing residue is greater than a cardinality of a set of positioning reference symbols compared to a system sampling rate. The coarse sample timing summand determines the coarse TOA within the OFDM symbol having the best PDP peak. A fine sample timing summand determines a fine TOA within the OFDM symbol having the best PDP peak.

In the foregoing, principles, preferred embodiments and various modes of

implementing the technique disclosed herein have been exemplarily described.

However, the present invention should not be construed as being limited to particular principles, embodiments and modes discussed above. Rather, it will be appreciated that variations and modifications may be made by a person skilled in the art without departing from the scope of the invention as defined in the following claims.