FIELD OF THE INVENTION

The present invention relates to locating a radio unit relative to a plurality of radio anchors, in environments that cause distortions to free-field models for localization.

In a free-field situation, one can use a radio propagation model to compute the distance between two radio units from the received signal strength, provided that the transmit power of the transmitting unit is known. When several radio anchors, being radio units with known locations, are present, one can use the received signal strengths to estimate the location of a radio unit by means of triangulation.

However, the accuracy of such triangulation approaches is limited due to large ambiguities in the radio propagation model in an environment that cannot be considered as a free field, i.e. any environment containing structures that interact with radio signals and that can cause distortions to the radio signals. A typical non- free- field situation is an interior of a building, wherein walls, floors, furniture and various physical structures interact with radio signals in a complex manner that differs substantially from a free-field situation.

It is however possible to locate a single radio unit relative to a plurality of radio anchors, operating in the radio frequency (RF) domain, in a non- free- field environment, using the signal strengths received between the radio unit and the radio anchors, but it requires a method than can deal with model ambiguities in non- free- field situations.

BACKGROUND OF THE INVENTION

US 2008/0004042 describes a method for locating radio units within a non- free- field environment. In this method, signal strength values between a radio unit and multiple radio anchors are measured, forming a sequence of signal strength values, which are subsequently compared against a database that contains multiple signal strength sequences and their corresponding locations. The sequence in the database that yields the closest match with the measured sequence is selected, and the corresponding location is selected as the estimated location of the radio unit.

A major drawback of the prior art is that it requires a database, containing many locations and corresponding signal strength sequences, making the preparation of the database elaborate. In effect, all locations that are a possible outcome of this localization method need to be present in the database, along with the corresponding signal strength sequences which are determined by either field measurement or by computation. Filling the database thus implies, for many locations throughout the localization area, performing either physical signal strength measurements or intensive computations using a radio propagation model that requires detailed input about the topology of the entire localization area.

SUMMARY OF THE INVENTION

It is an object of the invention to provide a method locating a radio unit in a non- free- field environment without the need for extensive calibration. The present invention discloses a method of locating a radio unit relative to a plurality of radio anchors, the radio anchors being radio units with known locations, the radio unit and the radio anchors being arranged to send or receive signals, the method comprising (a) grouping neighboring radio anchors into anchor groups, the total number of anchors exceeding the total number of anchor groups; and (b) associating a region to each anchor group, the region situated near the radio anchors of the anchor group and within a localization area that includes the regions; and (c) measuring the signals strengths between the radio unit and the radio anchors; and (d) combining the signal strengths of radio anchors of each anchor group into a respective group strength; and (e) determining at least one anchor group based on the group strength of the anchor group relative to other group strengths; and (f) selecting the region associated to the determined anchor group , the selected region being the estimated location of the radio unit.

The context of this invention concerns a radio unit with an unknown location in a non- free- field area containing multiple radio anchors with known locations. Within this area the radio unit is typically within range of several anchors such that signals can be sent or received between the radio unit and the radio anchors. The measured strength of these signals is the basis for the further localization procedure.

The method includes grouping of neighboring radio anchors into anchor groups, and associating a region with each anchor group. The region refers to locations situated near the radio anchors of the anchor group and lying within the localization area. In order to compare anchor groups on the basis of signal strength, the method assigns a group strength to each anchor group. The group strength is computed by combining the signals strengths of the individual radio anchors in the group. Determining the estimated location of the radio unit is done as follows. Group strengths of anchor groups are compared, and at least one anchor group is determined based on the group strength, for example the anchor group with the highest group strength or the anchor groups with group strengths near the highest group strength. In case only one anchor group is determined, the associated region is selected as the estimated location of the radio unit. In case multiple anchor groups are determined, one of the multiple regions associated to the multiple determined anchor groups is selected as the estimated location of the radio unit, wherein the selection is based, for example, on the spatial locations of the regions.

Optionally, the radio unit is measuring the strength of the radio signals transmitted by the radio anchors. The radio unit then acts as a receiver, while the radio anchors act as senders.

Optionally, the radio anchors are measuring the strength of the radio signals transmitted by the radio unit. The radio anchors then act as receivers, while the radio unit acts as a sender.

Optionally, the determined anchor group is the anchor group having the highest group strength. The estimated location of the radio unit is then region associated to the anchor group having the highest group strength.

Optionally, the at least one anchor group comprises determining a set of anchor groups having a group strength larger than the highest group strength minus a preset margin. The set of anchor groups then have group strengths that are near the highest group strength. The selected region is then among the regions associated to the anchor groups in the set of anchor groups.

Optionally, a plurality of anchor groups is determined, and wherein selection of the region comprises selection of one of the plurality of regions associated to the respective determined anchor groups, based on the spatial locations of the plurality of regions. For example, in case three anchor groups are determined, then, from the three regions associated to the three anchor groups, the most central region of the three regions may be selected as the estimated location of the radio unit.

Optionally, neighboring regions share an overlap area, the overlap area belonging to each of the neighboring regions. The overlap area may enhance the spatial resolution of the localization method to a sub-region size. An example case is when the anchor groups of two neighboring regions have the two highest group strengths values and these two values are similar. The overlap area then becomes the estimated location rather than one of the two neighboring regions Optionally, neighboring anchor groups share some radio anchors, the shared radio anchors belonging to each of the neighboring anchor groups. Sharing radio anchors provides additional flexibility in the grouping of radio anchors.

Optionally, the combining of the signal strengths into a group strength comprises averaging the signal strengths of the radio anchors of the radio group. The group strength being the average signal strength then becomes the single value representing the signal strength of the anchor group.

Optionally, the combining of the signal strengths into a group strength comprises averaging the N highest signal strengths of the radio anchors of the radio group, N being one or more. As the highest signals strengths tend to have a higher signal-to-noise ratio, the average from the highest signal strengths will also have a higher signal-to-noise ratio, making the average, being the group strength, a more stable value.

Optionally, the combining of the signal strengths of the radio anchors of the anchor group into a group strength comprises computing a weighted average of the signal strengths of the radio anchors, the weighted average comprising (a) associating a weight to each radio anchor in the anchor group, (b) multiplying the signal strength of the each radio anchor with the associated weight, and (c) summing the multiplied signal strengths. Using the weighted average, signal strengths from some radio anchors in the anchor group can contribute more to the weighted average than signal strengths from other radio anchors in the anchor be group. The weights may depend on any property of the signals strength or on any property of the radio anchors that is considered to be important. A weighted average has the benefit of being a flexible expression as the dependency can be described by any function.

Optionally, a weight depends on the signal strength of the radio anchor to which the weight is associated. Weights increasing with signal strength cause high signal strengths to contribute more to the weighted average than low signal strengths. Consequently, similar to what is described above, this may benefit the signal-to-noise ratio of the group strength.

Optionally, the weight depends on a relative position of the respective radio anchor within the anchor group, the relative position being the position relative to the region associated to the anchor group. More confidence can be expressed in signal strengths from radio anchors at certain relative positions by increasing the weights for the certain relative positions. Such assignment of weights may be part of a calibration procedure.

The invention discloses a system for locating a radio unit relative to a plurality of radio anchors, the radio anchors being radio units with known locations, the radio unit and the radio anchors being arranged to send or receive radio signals, the system comprising (a) the plurality of radio anchors; and (b) first data containing identifiers of radio anchors and identifiers of the associated anchor groups, the total number of anchors exceeding the total number of anchor groups; and (c) second data containing identifiers of anchor groups and the regions associated to the anchor groups, the regions situated near the radio anchors of the anchor groups and within a localization area that includes the regions; and (d) a measurement unit arranged to measure the signal strengths of radio signals between the radio unit and the radio anchors; and (e) a localization module arranged to combine the signal strengths of radio anchors of each anchor group into a group strength, and determine at least one anchor group based on the group strength of the anchor group relative to other group strengths, and select the region associated to the determined anchor, the selected region being the estimated location of the radio unit.

The system comprises a plurality of radio anchors, relative to which a radio unit is localized. The radio unit may receive signals transmitted by the radio anchors, putting the radio unit in the role of the sender and the radio anchor in the role of receivers.

Alternatively, the radio anchors may receive signals transmitted by the radio unit, putting the radio anchors in the role of senders and the radio unit in the role of receiver. The measuring unit provides the capability to measure the signal strength between the radio unit and the radio anchors, and can be comprised by either the radio unit or the radio anchors.

The system comprises a localization module arranged to receive the measured signal strengths from a measurement unit. The localization module is also arranged to perform a number of programming steps and to retrieve information, required by the programming, from the data. The programming steps finally yield the estimated location of the radio unit.

In order to compute group strengths, the system retrieves from, first data, the associations between radio anchors and anchor groups. Furthermore, in order to select a region after determining at least one anchor group the system retrieves, from second data, the associations between anchor groups and regions.

Optionally, a radio unit comprises a measurement unit and/or a localization unit. Comprising the measurement unit, the radio unit acts in the role of receiver, measuring the strength of radio signals sent by the radio anchors. Comprising the localization module, the radio unit itself derives the estimated location from the measured signals strengths.

Optionally, a radio anchor comprises a measurement unit and/or a localization module. Comprising a measurement unit, the radio anchor acts in the role of receivers, measuring the strength of radio signals sent by the radio unit. Comprising the localization module, the radio anchor itself derives the estimated location from the measured signals strengths.

Optionally, a computer program product comprises instructions for causing a processor system to perform the combining, the determining and the selecting according to the method for locating a radio unit relative to a plurality of radio anchors, as described above. The computer program product could be a so-called 'app', installed on a mobile phone for navigating within a building.

Optionally, a luminaire comprises a radio anchor. By adding radio anchors to luminaires, multiple luminaires are enabled to, for example, locate mobile phones inside a building with the purpose to present advertisements or messages to the persons carrying the mobile phones. As the spatial density of luminaires in a building is typically dense, each room typically containing at least one luminaire, luminaires having radio anchors become particularly suitable to implement the invention, as a the high spatial density of radio anchors enables the grouping while maintaining a high spatial resolution for localization.

An advantage of the invention is that the preparation of the localization method requires minor effort compared to US 2008/0004042 which discloses a localization method in a non- free- field environment. Defining anchor groups and assigning regions is a small effort compared to (1) exhaustive field measurements throughout the localization area or compared to (2) collecting detailed information on the topology and subsequently performing complex computations of a radio propagation model using the detailed information.

BRIEF DESCRIPTION OF THE DRAWINGS

In the drawings:

Fig. 1 schematically shows a radio unit receiving signals from two groups of radio anchors, Fig. 2 schematically shows a two groups of radio anchors receiving signals from a radio unit, Fig. 3 schematically shows two adjacent groups of radio anchors and their associated regions, Fig. 4a illustrates exclusive grouping, and

Fig. 4b illustrates overlapping grouping, and

Fig. 5a schematically shows two adjacent regions,

Fig. 5b schematically shows two regions that overlap, Fig. 5c schematically shows two regions that are separated by a space,

Fig. 6 schematically shows two adjacent regions, within each region an anchor group, Fig. 7 shows the localization method schematically,

Fig. 8 schematically shows the localization,

Fig. 9a schematically shows a radio unit receiving radio signals sent by radio anchors, Fig. 9b schematically shows radio anchors receiving radio signals sent by the radio unit.

DETAILED DESCRIPTION OF EMBODIMENTS The localization area concerns a non- free- field situation with a plurality of radio anchors, being radio units with an known locations, and a radio unit with an unknown location.

FIG. 1 schematically shows a radio unit receiving signals from two groups of radio anchors. FIG. 1 illustrates an environment with a radio unit 101, anchor group A 104 with radio anchors A1-A3 102, anchor group B with radio anchors B1-B3, and anchor group C with anchors C1-C3. The radio unit 101 receives a signal 103 from each of the radio anchors A1-A3, B1-B3 with signal strengths S _AI , S _A2 , S _A 3 107, S _BI , S _B2 , S _B 3- The radio unit 101 is out of receptive range of radio anchors C1-C3, such that the radio unit 101 cannot detect their signals. The direction of the arrows indicate the direction of the radio signals, being sent from the radio unit 101 to the radio anchors A1-A3, B1-B3 . The thickness of the arrows illustrate the differences between the signal strengths S _AI , S _A2 , S _A 3, S _BI , S _B2 , S _B 3- Region RA 105 comprises anchor group A as well as the radio unit, region RB comprises anchor group B, and region RC comprises anchor group C. Localization area 106 comprises the regions RA, RB, and RC. For example, the radio unit 101 may be a mobile phone, receiving and measuring signals sent by radio anchors inside a building, the mobile phone using the measured signals to estimate its own location inside the building.

FIG. 2 schematically shows two groups of radio anchors receiving signals from a radio unit. FIG. 2 is identical to FIG. 1 except for the difference that the roles of sender and receiver are interchanged. In FIG. 2 radio unit 201 sends signals 203 to the six anchors A1-A3 202, B1-B3 , and the signals are received with signal strengths S _AI , S _A2 , S _A3 207, S _BI , S _B2 , S _B3 - The anchors C1-C3 are out of receptive range of radio unit 201, such that the radio unit 201 cannot detect their signals. The direction of the arrows indicate the direction of the radio signals, being sent by the radio anchors A1-A3, B1-B3 to the radio unit 201. The thickness of the arrows illustrate the differences between the signal strengths S _AI , S _A2 , S _A 3, S _BI , S _B2 , S _B 3. Region RA 205 comprises anchor group A 204 as well as the radio unit, region RB comprises anchor group B, and region RC comprises anchor group C.

Localization area 206 comprises the regions RA, RB, and RC. For example, the radio anchors may reside inside a building, the radio anchors receiving and measuring the signals sent by a mobile phone inside the building, using the measured signals to estimate the location of the person carrying the mobile phone.

Differences between signal strengths S _AI , S _A2 , S _A 3, S _BI , S _B2 , S _B 3 in FIG. 1 and Fig. 2 can be caused by various physical parameters in the environment. Apart from the obvious factors being transmit power and the radio frequency used, variations in signal strengths in a free field environment are caused by the distance between sender and receiver, thus the distances between the radio unit 101 and the radio anchors A1-A3, B1-B3. Yet, in a non- free- fie Id environment the interactions of the radio signals with various obstacles cause additional variations. In a building, for example, this may concern obstacles such as furniture, walls, floors and other physical structures.

In the invention, measurement of the signals strength of a radio anchor entails temporally averaging several short term signal strength measurements in order to minimize temporal fluctuations in the signal strength measurement, such as caused by fast channel fading. For example, for high data rates such as offered by standard radio communicatons such as WiFi, it is in practice possible to average several hundreds of short term signal strength values within a few seconds, so that the temporal averaged signal strength itself still has a temporal resolution of a few seconds. In general, the pace of the movement of the radio unit determines the temporary resolution needed for the localization. In this the temporal resolution of the localization method equals the temporal resolution of the signal strength measurement, a radio unit could be localized once every few seconds, which would be adequate for locating a walking person inside a building, for example.

In one embodiment, the radio unit is a mobile phone transmitting radio signals, the interior of a building is the localization area, and the building comprises multiple radio anchors that receive radio signals, the radio anchors being distributed throughout the building. The mobile phone is typically carried by a person walking through the building, while several radio anchors measure the signal strength of the radio signal transmitted by the mobile phone. A localization module receives, from the radio anchors and through a communication link, the measured signal strength values, and subsequently estimates the location of the radio unit by processing the measured signal strengths, according to the method of the invention. As an example application, the location of the mobile phone can be used by an information system in the building that shows messages to a person carrying the mobile phone. The information may contain, for example, commercial ads or local news. In addition, if the central unit or radio anchors recognize the identity of the mobile phone from previous visits to the building, the advertisements or messages may be even personalized for the person carrying the mobile phone.

FIG. 3 schematically shows two adjacent anchor groups and their associated regions. Anchor group A with radio anchors A1-A3, anchor group B with anchors B1-B3, region A and region B. Region A is an area near neighboring radio anchors A1-A3 and is associated to anchor group A. Likewise, region B is an area near the neighboring radio anchors B1-B3 and is associated to anchor group B. In general, each region is associated to a nearby group of anchors and each region is part of a localization area that includes all regions. For example, regions A and B could each be a room inside a building being the localization area. Or, regions A and B may be sub-regions of a room, region A being one half of the room and region B being the other half of the room.

Grouping of radio anchors into anchor groups and the assignment of regions to anchor groups enables a straightforward localization method with little preparation compared to established methods. Computing a group strength provides a single value representing the signal strength of the anchor group. Using the group strengths, groups can be compared on the basis of their signal strength, so that localization of a radio unit effectively becomes selecting the region associated to the anchor group with the highest group strength.

Computation of the group strength is described in what follows.

FIG. 4a illustrates exclusive grouping and FIG. 4b illustrates overlapping grouping. FIG. 4a illustrates exclusive grouping of two anchor groups, A and B, with radio anchors A1-A3 and B1-B3, respectively. In this grouping, radio anchors A1-A3 belong exclusively to anchor group A, and, likewise, radio anchors B1-B3 belong exclusively to anchor group B. Optionally, as an alternative to exclusive grouping, FIG. 4b illustrates overlapping grouping of two anchor groups, A and B: radio anchors Al and A2 belong to anchor group A, anchors Bl and B2 belong to anchor group B, and anchor AB belongs to both anchor group A and anchor group B. Allowing sharing of radio anchors between anchor groups provides additional flexibility in the grouping of radio anchors.

As the present invention comprises grouping of radio anchors, the number of anchor groups should exceed at least the number of radio anchors. As each anchor group has a region assigned to it, the number of regions therefore also should exceed the number of radio anchors.

FIG. 5a schematically shows two adjacent regions, whereas FIG. 5b schematically shows two regions that overlap. Regions may be adjacent neighbors, thus having borders that partially coincide, as shown in Fig. 5a. Yet, neighbors may also be separated by some space that is not a region, as shown in Fig. 5b. The separation between regions may be caused by a structure such as a wall, for example.

FIG. 5c schematically shows two adjacent regions, RA and RB, that share an overlap area OAB. The overlap area OAB may be used in case a radio unit is located near the border of the regions RA and RB, such that the localization system is inconclusive whether the radio unit is located in region RA or in region RB. In such a case, the localization system may locate the radio unit to the overlap region OAB. Using overlap areas in this manner, in addition to regions, effectively enhances the spatial resolution of the localization system. A case wherein the localization system is inconclusive may occur when group strengths of the anchor groups associated to region RA and RB are similar.

Fig. 6 schematically shows two adjacent regions, within each region a group of six radio anchors. FIG. 6 illustrates a radio unit 601 and two regions, each region comprising an anchor group. Region RA comprises anchor groups A with radio anchors Al- A6, and region RB comprises anchor group B with radio anchors B1-B6. The radio unit 601 is within range of all radio anchors A1-A6,B1-B6 such that the signal strength between the radio unit 601 and radio anchors A1-A6,B1-B6 can be measured. The signal strengths between radio unit 601 and radio anchors A1-A6 are denoted in this description as

S _AI -S _A6 , whereas the signal strengths between radio unit 601 and radio anchors B1-B6 are denoted as S _BI -S _B6 - In what follows, Fig. 6 is used to illustrate the localization method, being computing group strengths, comparing group strengths and selecting a region as the estimated location of the radio unit.

The radio unit 601 can be localized as follows. The group strength for anchor group A is computed from the group strengths S _AI -S _A6 of the individual anchors, the group strength being a single value representing the signal strength of the anchor group. A straightforward implementation is to compute the group strength G _A as the average of the signals strengths of the radio anchors. If the group strength G _B is computed in an analogous manner, then: GA— (SA1+SA2+SA3+SA4+SA5+SA6) 6

GB = (SB1+SB2+SB3+SB4+SB5+SB6) 6 Anchor groups A and B can now be compared on the basis of their group strengths, the highest group strength directly indicating whether the radio unit 601 resides within region A or within region B: if G _A > G _B then the radio unit 601 is located to region A,

if G _A ≤ G _B then the radio unit 601 is located to region B.

This comparison is done under the assumption that radio unit 601 resides in either region A or region B.

Additionally, an overlap area OAB could be defined, such as depicted in FIG. 5c, when the group strengths G _A and G _B are similar. When |G _A - G _B | < Θ the radio unit 601 could be assigned to overlap area OAB, wherein similarity threshold Θ is a small number relative to G _A and G _B . Choosing the size of overlap area, such as R _AB , and the size of similarity threshold Θ would be part of setting up the localization system.

Optionally, the computation of the group strength G _A can be arranged such that, for a group of anchors, the weakest signals are discarded, so that the group strength is computed from the strongest signals in the anchor group only. For example, for the anchor group A in FIG. 6, the group strength can be computed by averaging the four strongest signals only, thus discarding the two weakest signals . The group strength can be expressed for anchor group A as follows. If the signal strengths S _AI ,S _A2 ,S _A 3S _A4 ,¾5,S _A 6 are sorted according to decreasing value, the sorted signal being named T _I ,T ₂ ,T3,T ₄ ,T ₅ ,T _A 6 (thus T ₅ and T ₆ being the two lowest signal strength values), then the group strength G _A is computed by

G _A = (T ₁ +T ₂ +T ₃ +T ₄ )/4 Alternatively, a preset margin D can be used to determine a plurality of anchor groups with group strengths G larger than the highest group strength Gmax minus the preset margin D. In mathematical terms, anchor groups are determined if the anchor groups have group strengths G that test positively on G > Gmax-D. In case several group strengths G test positively, the several respective anchor groups are determined, and, consequently, the several regions associated to the respective anchor groups are candidates for being the estimated location of the radio unit. Selection of the region being the estimated location of the radio unit can be based on the spatial locations of the candidate regions. For example, if three anchor groups have group strengths G wherein G > Gmax-D, then the three regions associated to the three anchor groups are the candidate regions. One may select the candidate region nearest to the center location of the three candidate regions as the estimated location of the radio unit. Alternatively, the estimated location of the radio unit may be selected by averaging spatial locations of the candidate regions. Alternatively, the estimated location of the radio unit may be selected based on criteria other than spatial locations of the candidate regions. For example, the estimated location may be selected as the candidate region that is most frequently visited by radio units.

Alternatively, the group strength can be computed as a weighted average of the signal strengths of the individual anchors. Computation of the group strength GA for group A in FIG. 5 is described by a weighted average according to:

GA= (WI SA1+ W ₂ SA2+ W3SA3+ W ₄ SA4+ WSSA 5+ W ₆ SA6) ( Wl+ W ₂ + W3+ W ₄ + W5+ w ₆ )

Weights wi -w ₆ can may depend, for example, on the signal strength, but the weights may also depend on other factors. Weights wi -w6 may depend on signal strengths SAI -SA ₆ in various ways. A possible implementation is to assign a first constant value a to a weight w; , wherein i=l,2,3,4,5, or 6, if the weight w; exceeds a threshold t, and to assign second constant value b to the weight w; if the weight w; does not exceed the threshold t: if Wi >t then w; =a ,

if Wi <t then w; =b .

In order to favor high signal strengths in the computation of the group strength GA, constant a should be larger than constant b, thus a > b.

Optionally, the weights can be made dependent on other factors than the signal strength. For example, the weights could be preset and depend on the radio anchor identifiers, so that the relative contribution of the signal strength from a radio anchor to the weighted average can be higher or lower than from other anchors. In this sense, the contribution of some radio anchors can be favored relative to other radio anchors in an anchor group. For example, in FIG. 6, radio anchors A2 and A5 that reside in the center of the region RA could be favored with a factor 2 over the radio anchors A1,A3,A4,A6 that reside at the left and right of regions RA: w2=w5=2, wl=w3=w4=w6=l . The group strength G _A would then be computed as: GA = (SAI+ 2 SA2+ SA3+ SA4+ 2 SAS+ SA ₆ )/8

In case the weights are specific for each anchor group then the weights should be stored in a database (not shown in the figures), so that the weights for a certain anchor group be retrieved at any moment. The weight specific for each radio anchor (identifier) could be determined by a calibration procedure, thus presetting the relative contribution of each radio anchor to the group strength of its anchor group.

The motivation to assign a higher weight to an anchor in the center of the region could be that anchors in the center are considered more representative for the position of the region than anchors that reside nearer to the edge of the region. Another motivation for weights depending on radio anchor identifier could that the signal strength from the respective anchors are considered less or more reliable. For example, a radio anchor that has a physical structure in its direct vicinity, causing much interference to radio signals, may be considered less reliable.

Computations, comparisons and other operations involving signal strength as described thus far in this document are valid under the assumption that the transmit power of the senders are similar. Although, it is not necessary in the invention that the transmit power of all senders are all similar, it is preferred that the relative strengths of the transmit powers are at least known. In the case that the transmit powers are not the similar for the senders, a distinction is made between a corrected signal strength and a raw signal strength, the raw signal strength being the signal strength as directly measured by a receiver. A corrected signal strength S is then computed from the raw signal strength Sraw by accounting for the difference between a transmit power Τοΐ the sender and a reference transmit power Trej ^' by: S = Sraw x Tref / T

In what is described above, all equations and comparisons comprising 'signal strength' are valid either using the corrected signal strength or assuming that transmit powers of all senders are similar.

The desired localization accuracy determines the necessary spatial density of the radio anchors. For example, if a radio unit is to be localized to a room, then a single region is typically assigned to the room containing multiple radio anchors. As another example, if a radio unit is to be localized to one half of a room, then the room typically contains two regions, each region containing multiple radio anchors. In the latter example, the localization accuracy may be extended by creating an additional overlap area and by refining the comparison of the group strengths of the two anchor groups: if the two group strengths are sufficiently similar then the radio unit is located to the overlap area, otherwise the radio unit is located to one of the other two regions (as explained above and in FIG. 5b).

FIG. 7 shows the localization method schematically. A first unit 701 selects radio anchors that are within range of a radio unit 706 with unknown location, such that the signal strengths between the radio anchors and the radio unit can be measured. Unit 701 passes on identifiers of the selected radio anchors to a second unit 702 that measures the signal strengths between the radio unit 706 and the radio anchors. A third unit 703 receives from the second unit 702 anchor identifiers and corresponding signal strengths, then groups the radio anchors into anchor groups and computes a group strength for each anchor group. A unit 704 receives from the third unit 703 anchor group identifiers and the corresponding group strength, and the fourth unit 704 selects the anchor group with the highest group strength. A fifth unit 705 receives from the fourth unit 704 the identifier of the selected anchor group and selects the region associated to the selected anchor group, the selected region being the estimated location of radio unit 706. The arrow starting from unit 705 back to unit 701 indicates that the localization method is repeated at the next moment in time.

FIG. 8 shows the localization system schematically. FIG. 8 shows the localization system 800 with four units SELAMEAS 810, GRPCOMP 820, GRPSEL 840, and REGNLUT 850, and two databases, DB1 830 and DB2 860. The first unit SELAMEAS selects radio anchors ANIDs 811 and measures the signal strengths STRs 812 between the radio unit and the selected radio anchors. The first unit SELAMEAS passes on the radio anchor identifiers ANIDs and corresponding signal strengths STRs to the second unit

GRPCOMP. The second unit GRPCOMP fetches from the first database DB1 the anchor group identifiers GRPIDs 831 corresponding to the radio anchor identifiers ANIDs , and then computes the group strengths GRPSTRs 821 for the fetched anchor groups. The third unit GRPSEL then receives from the second unit GRPCOMP the anchor group identifiers GRPIDs and corresponding computed group strengths GRPSTRs, and selects the anchor group with the highest group strength, the selected anchor group having identifier GRPID 811. The fourth unit REGNLUT receives, from the third unit GRPSEL, the anchor group identifier GRPID and fetches the region coordinates REGN 861 corresponding to the group identifier GRPID from the second database DB2, region coordinates REGN representing the estimated location of the radio unit. The fourth unit REGNLUT then presents the coordinates REGN as final output of the localization system 800.

In FIG.8, the combined units GRPCOMP, GRPSEL and REGNLUT are considered as a localization unit LOC 870 that acts as a programming unit for performing the operations and computations to derive the estimated location REGN from the signal strengths STRs and corresponding radio anchor identifiers ANIDs.

FIGs. 9a, 9b show two different embodiments of the radio anchors and the radio unit, the radio anchors and the radio unit acting in different roles of sender and receiver

FIG. 9a schematically shows a radio unit receiving radio signals sent by radio anchors. FIG. 9a shows the radio unit 915 being a receiver and radio anchors 911-914 being senders, depicted in a cloud 940. The radio unit 915 receives signals 921-924 from the radio achors 911-914. The radio unit 915 measures the strengths of the signals 921-924 and passes the radio anchor identifiers ANIDs 811 and corresponding signal strengths STRs 812 to unit GRPCMP 820. Effectively, the depicted cloud, comprising the radio station 915 and the plurality of radio anchors 91 1-914, is an embodiment of unit SELAMEAS in FIG. 8.

FIG. 9b schematically shows radio anchors receiving radio signals sent by the radio unit. FIG. 9b shows the radio unit 955 being a receiver and radio anchors 951-954 being senders, depicted in a cloud 990. The radio anchors 951-954 receive signals from the radio unit 955 with signal strengths 961-964, respectively. A unit COLL 975 collects the signal strength values STRs 812 from the radio anchors 961-964 and passes the signal strength values STRs 812 and corresponding radio anchor identifiers ANIDs 811 to unit GRPCMP 820. Effectively, the combination of the unit COLL 975 and the depicted cloud, the cloud comprising the radio station 955 and the plurality of radio anchors 951-954, is an embodiment of unit SELAMEAS in FIG.8.

In an embodiment, the radio unit is a mobile phone receiving radio signals transmitted by the radio anchors. The mobile phone measures the signal strengths of the radio signals and computes the estimated location of the phone using the signal strengths. This requires that the mobile phone comprises a localization unit to compute the estimated location and that the mobile phone also comprises the databases that associate radio anchors with anchor groups and anchor groups with regions.

In a variant of the previous embodiment, the mobile phone measures the radio signals transmitted by the radio anchors, but, different from the previous embodiment, the mobile phone does not comprise the localization unit. The mobile phone sends the measured signal strengths, through a communication link, to a central programming unit comprising the localization unit that computes the estimated location from the received measured signal strengths. The central unit and the communication link are not shown in the figures.

The invention exploits the trend of the increasing spatial density of radio networks, as a high spatial density enables a straightforward localization of a radio unit based on a comparison of group strengths. The group strength value is, in principle, a more stable value than the signal strength value of the individual radio anchors that can suffer from local distortions from obstacles and physical structures in their vicinity. By combining the signal strengths of several neighboring radio anchors into a group strength such local distortions tend to cancel out in the combined value being the group strength. Consequently, group strengths from different anchor groups can be compared directly, and the anchor group having the highest group strength directly leads to the region being the estimated location of the radio unit.

It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be able to design many alternative embodiments without departing from the scope of the appended claims.

In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. Use of the verb "comprise" and its conjugations does not exclude the presence of elements or steps other than those stated in a claim. The article "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention may be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the device claim enumerating several means, several of these means may be embodied by one and the same item of hardware. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.