Patent Application
20230273285
This application claims the benefit of Finnish Patent Application No. 20225116, filed Feb. 25, 2022. The entire content of the above-referenced application is hereby incorporated by reference.
Various example embodiments relate to an apparatus comprising at least one processor.
Further embodiments relate to a method of operating related to such apparatus.
Location information, for example associated with a mobile device such as a mobile terminal device, e.g. user equipment (UE), for a communications system, can be used for applications such as e.g. navigation, asset tracking, industrial automation, healthcare, etc.
Some conventional localization approaches utilize time of arrival (ToA) or time difference of arrival (TDoA) measurements, where TDoA methods offer the advantage that an exact transmit time of a target to be tracked does not have to be known. Some TDoA-based localization techniques can for example be used in current 3rd Generation Partnership Project (3GPP) 4G and 5G positioning standards.
Various embodiments of the disclosure are set out by the independent claims. The exemplary embodiments and features, if any, described in this specification, that do not fall under the scope of the independent claims, are to be interpreted as examples useful for understanding various exemplary embodiments of the disclosure.
Some embodiments relate to an apparatus, comprising at least one processor, and at least one memory storing instructions, the at least one memory and the instructions configured to, with the at least one processor, cause the apparatus at least to determine a set of position estimates for a mobile device, wherein each position estimate of the set of position estimates is associated with a a) specific one of a plurality of K many receiver devices configured to receive a reference signal from the mobile device or b) a specific one of a plurality of K many transmitter devices configured to transmit reference signals to the mobile device, and to determine, based at least on the set of position estimates, a set of residual errors associated with at least some position estimates of the set of position estimates. In some embodiments, this enables to determine a position of the mobile device, and/or to assess a quality of the position estimates.
In some embodiments, an uplink TDoA (“UTDoA”) configuration is e.g. characterized by the position estimates of the set of position estimates being associated with a specific one of a plurality of K many receiver devices configured to receive a reference signal from the mobile device.
In some embodiments, a downlink TDoA (“DTDoA”) configuration is e.g. characterized by the position estimates of the set of position estimates being associated with a specific one of a plurality of K many transmitter devices configured to, for example simultaneously, transmit reference signals, e.g. positioning reference signals, to the mobile device.
Note that, in some embodiments, different values may be chosen for the number K of receiver devices or transmitter devices, e.g. depending on a specific implementation or precision requirements or the like.
In some embodiments, the number K of receiver devices is equal to or greater than 2.
In some embodiments, the number K of transmitter devices is equal to or greater than 2.
In some embodiments, e.g. related to a downlink TDoA configuration, K=K1 many transmitters may be used, whereas, in some other embodiments, e.g. related to a uplink TDoA configuration, K=K2 many transmitters may be used, wherein K1 may be different from K2.
In some embodiments, both an uplink TDoA and a downlink TDoA configuration may be used.
While the following description, for the sake of clarity, primarily refers to exemplary UTDoA configurations, the principle according to the embodiments can be applied both to a uplink TDOA configuration and a downlink TDoA configuration.
In some embodiments, the mobile device is a terminal device, for example UE, for a, for example cellular, communications network of e.g. the 5G and/or 6G type.
In some embodiments, e.g. related to an exemplary uplink TDoA configuration, at least some of the K many receiver devices are location measurement units, LMUs, which are configured to receive at least one reference signal of the mobile device, e.g. a sounding reference signal (SRS) of a 5G network. In some embodiments, the receiver devices are not equipped with a transmitter, but rather only a receiver for receiving the reference signal(s), e.g. of a mobile device a position of which is to be determined.
In some embodiments, e.g. related to an exemplary uplink TDOA configuration, at least some of the K many receiver devices can be implemented in the form of, comparatively inexpensive, reduced-functionality 5G sniffers, configured to sense or receive, respectively, the reference signal(s) transmitted by a mobile device a position of which is to be determined.
In some embodiments, e.g. related to an exemplary downlink TDoA configuration, at least some of the K many transmitter devices are reference signal transmission units, which are configured to transmit at least one reference signal, e.g. a positioning reference signal, e.g. of a 5G network, to the mobile device. In some embodiments, the transmitter devices are not equipped with a receiver, but rather only a transmitter for transmitting the reference signals to a mobile device a position of which is to be determined.
In some embodiments, e.g. related to an exemplary downlink TDOA configuration, at least some of the K many transmitter devices can be implemented in the form of, comparatively inexpensive, reduced-functionality 5G transmitters, e.g. configured to transmit the reference signals, e.g. positioning reference signals (PRS), to a mobile device a position of which is to be determined.
In some embodiments, e.g. related to an exemplary UTDoA configuration, at least some of the K many receiver devices can be tightly synchronized, for example to enable TDoA positioning.
In some embodiments, e.g. related to an exemplary DTDoA configuration, at least some of the K many transmitter devices can be tightly synchronized, for example to enable TDoA positioning.
In some embodiments, e.g. related to an exemplary UTDoA configuration, at least some of the K many receiver devices can be configured to transmit received reference signals and/or signals derived from received reference signals to the apparatus according to the embodiments, e.g. for evaluation, e.g. according to the principle of the embodiments.
In some embodiments, e.g. related to an exemplary DTDoA configuration, signals, e.g. positioning signals, received from the plurality of e.g. K many transmitter devices, and/or signals derived from the received (positioning) reference signals may be evaluated, e.g. according to the principle of the embodiments.
In some embodiments, the instructions, when executed by the at least one processor, cause the apparatus to determine updated weights for weighting the set of position estimates based on the set of residual errors. In some embodiments, this enables to attain more precise position estimates.
In some embodiments, the instructions, when executed by the at least one processor, cause the apparatus to determine an updated set of position estimates based on the updated weights.
In some embodiments, the instructions, when executed by the at least one processor, cause the apparatus to determine whether at least one termination criterion is met, and, based on the determination whether the at least one termination criterion is met, to repeat at least one of the preceding steps.
In some embodiments, this enables to provide an iterative procedure which enables to iteratively refine position estimates associated with the mobile device.
In some embodiments, the at least one termination criterion may e.g. comprise a maximum number of iterations.
In some embodiments, the at least one termination criterion may be based on a position mismatch characterizing a difference between a current position estimate and a position estimate of a previous iteration falling below a predetermined threshold.
In some embodiments, e.g. related to an exemplary UTDoA configuration, the instructions, when executed by the at least one processor, cause the apparatus to determine estimated distances associated with the plurality of receiver devices.
In some embodiments, e.g. related to an exemplary DTDoA configuration, the instructions, when executed by the at least one processor, cause the apparatus to determine estimated distances associated with the plurality of transmitter devices.
As mentioned above, the following description, for the sake of clarity, primarily refers to exemplary UTDoA configurations having a plurality of K many receiver devices. However, as also already mentioned above, the principle according to the embodiments can alternatively (or additionally) be applied to exemplary DTDOA configurations having a plurality of K many transmitter devices.
In some embodiments, e.g. related to exemplary UTDoA configurations, the K many receiver devices may be associated with three-dimensional (3D) positions, which, in some embodiments, may e.g. be characterized by a respective 3×1 vector pk=[xk, yk, zk]T.
In some embodiments, e.g. related to exemplary DTDoA configurations, the K many transmitter devices may be associated with 3D positions, which, in some embodiments, may e.g. be characterized by a respective 3×1 vector pl=[xl, yl, zl]T.
In some embodiments, e.g. related to exemplary UTDoA configurations, a distance estimate associated with the k-th receiver device may be denoted as {circumflex over (d)}k.
In some embodiments, e.g. related to exemplary UTDoA configurations, a reference receiver device may be defined. As an example, in some embodiments, choosing e.g. a first receiver device as a reference receiver device, the K−1 distance differences associated with the receiver devices can be characterized by {circumflex over (d)}k,1={circumflex over (d)}k−{circumflex over (d)}1, k=2, . . . , K [equation 1].
In some embodiments, e.g. related to exemplary UTDoA configurations, the instructions, when executed by the at least one processor, cause the apparatus to determine initial weights (e.g., one or more initial weights) for weighting the set of position estimates associated with at least one of the plurality of receiver devices.
In some embodiments, the instructions, when executed by the at least one processor, cause the apparatus to determine the initial weights based on at least one of: i) a reliability of estimated distances associated with a) the plurality of receiver devices or with b) the plurality of transmitter devices, ii) predetermined initial weight values.
In some embodiments, the reliability of the estimated distances associated with the plurality of receiver devices can e.g. be determined based on knowledge about a received signal strength, RSS, since, in some embodiments, multipath reflections tend to be weaker than a direct (line-of-sight) path.
Hence, in some embodiments, measurements, e.g. distance estimates, whose signal strengths are low, e.g. fall below a predetermined RSS threshold, may be downweighted, e.g. weighted with comparatively small weight factors, whereas measurements, e.g. distance estimates, whose signal strengths are high, e.g. exceed the RSS threshold, may receive more emphasis by using larger weight factors.
In some embodiments, if no RSS information or other information characterizing a reliability of the estimated distances associated with the plurality of receiver devices (or transmitter devices, e.g. in case of a downlink TDoA configuration) (or similar) are available, unit initial weights may be used for at least some, for example all, receiver devices (or transmitter devices, e.g. in case of a downlink TDoA configuration).
In some embodiments, e.g. related to exemplary UTDoA or DTDoA configurations, the instructions, when executed by the at least one processor, cause the apparatus to determine K many position estimates {circumflex over (x)}1, {circumflex over (x)}2, . . . {circumflex over (x)}K, and to determine a weighted average position estimate {circumflex over (x)}WA based on the K many position estimates {circumflex over (x)}1, {circumflex over (x)}2, . . . {circumflex over (x)}K and on at least one of initial weights and updated weights for weighting the set of position estimates.
In some embodiments, e.g. related to exemplary UTDoA or DTDoA configurations, for each position estimate of the K many position estimates another receiver device of the K many receiver devices or transmitter device of the K many transmitter devices may be used as reference receiver device or reference transmitter device.
In other words, in some embodiments, e.g. related to exemplary UTDoA or DTDoA configurations, instead of determining a unique reference receiver device or reference transmitter device, which may e.g. be the one that is, for example, a-priori considered the most reliable one (e.g., always the one with the lowest distance estimate) before computing {circumflex over (x)}, it is proposed to use every receiver device or transmitter device of the K many receiver devices or transmitter devices as a reference receiver device or as a reference transmitter device once to obtain the K many position estimates {circumflex over (x)}1, {circumflex over (x)}2, . . . {circumflex over (x)}K, e.g. in the form of a matrix with K distinct position estimates {circumflex over (X)}=[{circumflex over (x)}1, {circumflex over (x)}2, . . . {circumflex over (x)}K].
In some embodiments, an estimate of the mobile device's position x=[x,y,z]T may e.g. be obtained by solving an equation system, characterized e.g. by: Aθ=b+n, [equation 2], with
and with
where d1=∥x−p1∥ (nuisance parameter), and wherein n is a noise vector.
In some embodiments, a closed-form approximate weighted least squares (WLS) solution of equation 2 can be obtained as
{circumflex over (θ)}=(ATWA)−1ATwb, [equation 6],
where W=diag{w2, . . . , wK} is a weighting matrix comprising the weights for weighting the distance difference estimates, i.e., {circumflex over (d)}k,1.
In some embodiments, the position estimates {circumflex over (x)} can be determined by adding e.g. the first three elements {circumflex over (θ)}1:3 of the WLS solution from Eq. (6) to the position p1 of a, for example reference, receiver device:
{circumflex over (x)}=p1+{circumflex over (θ)}1:3, e.g. using the WLS estimator according to equation 6.
In some embodiments, the weighted average position estimate may be obtained as
{circumflex over (x)}
WA=Σk=1Kwnorm,k·{circumflex over (x)}k, [equation 8]
where wnorm,k represents a normalized k-th weight (e.g. associated with the k-th receiver device or transmitter device), wherein, in some embodiments, a normalization such that Σk=1Kwnorm,k=1 is proposed.
In other words, in some embodiments, for determining the weighted average position estimate {circumflex over (x)}WA, multiple individual WLS estimates {circumflex over (x)}k are combined in a weighted fashion.
In some embodiments, e.g. in order to detect receiver devices or transmitter device with erroneous measurements, a measure of a reliability of the position estimates of at least some, for example all, receiver device or transmitter device is provided. In some embodiments, the measure of reliability indicates to which extent the respective estimate can be “trusted”.
Hence, in some embodiments, e.g. related to exemplary UTDoA configurations, the instructions, when executed by the at least one processor, cause the apparatus to determine a residual error ek of a position estimate associated with a k-th receiver device based on
wherein {circumflex over (d)}k characterizes an estimated distance associated with the k-th receiver device, wherein d1=∥x−p1∥, wherein {circumflex over (x)}WA characterizes a weighted average position estimate, wherein pk characterizes a position of the k-the receiver device, wherein p1 characterizes a position of a first receiver device.
Similarly, in some embodiments, e.g. related to exemplary DTDoA configurations, the instructions, when executed by the at least one processor, cause the apparatus to determine a residual error ek of a position estimate associated with a k-th transmitter device based on
wherein {circumflex over (d)}k characterizes an estimated distance associated with the k-th transmitter device, wherein d1=∥x−p1∥, wherein {circumflex over (x)}WA characterizes a weighted average position estimate, wherein pk characterizes a position of the k-the transmitter device, wherein p1 characterizes a position of a first transmitter device.
In some embodiments, e.g. related to exemplary UTDoA configurations or DTDoA configurations, the instructions, when executed by the at least one processor, cause the apparatus to determine whether a residual error of a position estimate associated with a k-th receiver device or transmitter device exceeds a threshold, and, if the residual error of the position estimate associated with a k-th receiver device or transmitter device exceeds the threshold, to discard the k-th receiver device or transmitter device.
In some embodiments, e.g. related to exemplary UTDoA or DTDoA configurations, discarding the k-th receiver device or transmitter device comprises disregarding signals or measurements provided by the k-th receiver device or transmitter device, e.g. for determining position estimates.
In some embodiments, the residual error as exemplarily characterized by equations 9, 9′ represents an averaged deviation between estimated distance differences of K−1 many receiver devices or transmitter devices and a reference receiver device or a reference transmitter device, and actual distance differences (i.e., w.r.t. a current estimate {circumflex over (x)}WA) of the K−1 receiver devices or transmitter devices and the reference receiver device or the reference transmitter device.
In some embodiments, using this formulation of equations 9 or 9′, the residual error may be interpreted as a measure of how much the measured distance differences agree, e.g. correlate, with a current position estimate {circumflex over (x)}WA or—to put it another way—with the distance differences that would result if the mobile device, e.g. UE, were located at a position characterized by {circumflex over (x)}WA.
In some embodiments, e.g. related to exemplary UTDoA configurations, the proposed approach according to the embodiments enables to avoid the choice of a “bad” reference receiver device, wherein e.g. an outlier measurement, which is e.g. associated with a reference receiver device, may severely affect all distance differences e.g. according to equation 1, which, in some embodiments, may e.g. be used to determine a final, e.g. WLS-based, position estimate.
In some embodiments, similar observations also apply to exemplary DTDoA configurations, wherein the proposed approach according to the embodiments enables to avoid the choice of a “bad” reference transmitter device.
Advantageously, in some embodiments, considering a situation with a “bad” reference receiver device or reference transmitter device, an associated residual error will have a comparatively high value, allowing the detection of receiver devices or transmitter devices with e.g. unreliable measurements, and, in some embodiments, downweighting their contribution, e.g. in equation 8, e.g. according to a magnitude of the residual error they produce.
In some embodiments, e.g. after computing a residual error, e.g. based on equation 9 or equation 9′, e.g. for all position estimates, updated weights, which may be used for a next iteration, may be obtained as
w
k
=f
e(ek), [equation 10],
where, in some embodiments, fe( ) is a monotonically increasing function that maps the magnitude of the residual errors ek e.g. to numbers between 0 and 1.
In some embodiments, for the mapping of the magnitude of the residual errors ek to the interval between 0 and 1, a function of the following types may be used: Andrew's sine wave, Huber loss, etc.
As already mentioned above, in some embodiments, e.g. related to exemplary UTDoA configurations, if the residual error ek exceeds a pre-defined maximum value, the corresponding weight may be set to zero, e.g. such that the measurement—and thus also the receiver device, in some embodiments—is discarded.
In some embodiments, this is also applicable to exemplary DTDoA configurations, wherein, if the residual error ek associated with a specific transmitter device exceeds a pre-defined maximum value, the corresponding weight may be set to zero, e.g. such that the measurement—and thus also the transmitter device, in some embodiments—is discarded.
In some embodiments, e.g. related to exemplary UTDoA configurations or exemplary DTDoA configurations, the above-described procedure may be iterated, yielding a robust iteratively reweighted least squares (IRLS) estimator that can e.g. detect and/or reject receiver devices or transmitter device with outlier measurements and/or that can find weights for the remaining receiver devices or transmitter devices or their measurements, respectively, e.g. based on their residual errors.
In some embodiments, after each iteration, a difference of a current position estimate to a position estimate of the previous iteration may be determined, e.g. computed, based on Δ=∥{circumflex over (x)}WA,i−{circumflex over (x)}WA,i-1∥, [equation 11].
In some embodiments, the difference Δ may e.g. be compared to a pre-defined threshold ε, e.g. to check convergence.
In some embodiments, based on the comparison of the difference Δ with the pre-defined threshold ε, a termination criterion may be set, e.g. to stop the process, preventing further iterations to be made.
While the exemplary embodiments explained above primarily refer to exemplary UTDoA configurations, the principle according to the embodiments can similarly be applied to DTDoA configurations.
Further embodiments relate to an apparatus comprising means for determining a set of position estimates for a mobile device, wherein each position estimate of the set of position estimates is associated with a) a specific one of a plurality of K many receiver devices configured to receive a reference signal from the mobile device or b) a specific one of a plurality of K many transmitter devices configured to transmit reference signals to the mobile device, and for determining, based at least on the set of position estimates, a set of residual errors associated with at least some position estimates of the set of position estimates.
Further embodiments relate to a location server for estimating a position of a mobile device which is configured to transmit at least one reference signal, comprising at least one apparatus according to the embodiments.
Further embodiments relate to a system for estimating a position of a mobile device which is configured to transmit at least one reference signal, comprising at least one apparatus according to the embodiments and K, K>=2, many receiver devices and/or K many transmitter devices.
Further embodiments relate to a method comprising: determining a set of position estimates for a mobile device, wherein each position estimate of the set of position estimates is associated with a) a specific one of a plurality of K many receiver devices configured to receive a reference signal from the mobile device or b) a specific one of a plurality of K many transmitter devices configured to transmit reference signals to the mobile device, and determining, based at least on the set of position estimates, a set of residual errors associated with at least some position estimates of the set of position estimates.
Further embodiments relate to a computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method according to the embodiments.
Further embodiments relate to a data carrier signal carrying and/or characterizing the computer program (product).
Some embodiments, see for example
In some embodiments, this enables to determine a position of the mobile device UE, and/or to assess a quality of the position estimates.
In some embodiments,
In some embodiments,
In some embodiments, at least some of the K many receiver devices RD1, RD2, . . . , RDk, . . . , RDK can be implemented in the form of, for example comparatively inexpensive, reduced-functionality 5G sniffers, configured to sense or receive, respectively, the reference signal(s) RS transmitted by a mobile device UE a position of which is to be determined.
In some embodiments, at least some of the K many receiver devices RD1, RD2, . . . , RDk, . . . , RDK are configured to perform TDoA measurements based on the reference signal RS.
In some embodiments, at least some of the K many receiver devices RD1, RD2, . . . , RDk, . . . , RDK can be tightly synchronized, for example to enable TDoA positioning.
In some embodiments, at least some of the K many receiver devices RD1, RD2, . . . , RDk, RDK can be configured to transmit received reference signals RS and/or signals derived from received reference signals to the apparatus 100 according to the embodiments, e.g. via data connections DC, e.g. for evaluation, e.g. according to the principle of the embodiments.
In some embodiments, the data connections DC may enable sub-nanosecond synchronization among the receiver devices RD1, RD2, RDk, RDK, e.g. by using a synchronization protocol such as, but not limited to, the White Rabbit protocol.
In some embodiments, a location server 10 may be provided, e.g. comprising the apparatus 100, for processing incoming signals from the receiver devices RD1, RD2, RDk, . . . , RDK, e.g. to obtain distance estimates. The location server 10, or the apparatus 100, respectively, may be configured to perform the method according to the embodiments exemplarily explained below.
In other words, in some embodiments, the location server 10 may e.g. be configured to receive received reference signals and/or signals derived from received reference signals from at least some of the K many receiver devices RD1, . . . , RDK, e.g. for processing, e.g. according to a method according to the embodiments exemplarily explained below.
In some embodiments,
In some embodiments,
In some embodiments, e.g. related to the exemplary DTDoA configuration of
In some embodiments, the apparatus 10′ may be configured to perform the method according to the embodiments. In some embodiments (not shown), the apparatus 10′ or its functionality, respectively, may also be integrated in the UE′.
In other embodiments, the UE′ of
While the following description of some exemplary embodiments, for the sake of clarity, primarily refers to exemplary UTDoA configurations as exemplarily depicted by
In some embodiments,
In some embodiments,
In some embodiments,
In some embodiments, this enables to provide an iterative procedure which enables to iteratively refine position estimates associated with the mobile device UE.
In some embodiments, the at least one termination criterion may e.g. comprise a maximum number of iterations.
In some embodiments, the at least one termination criterion may be based on a position mismatch characterizing a difference between a current position estimate and a position estimate of a previous iteration falling below a predetermined threshold.
In some embodiments,
In some embodiments, for each active receiver device (
Similarly, in some embodiments, e.g. in case of exemplary DTDoA configurations (
In some embodiments, residual errors are determined, which e.g. rely on distance differences between pairs of locators, e.g. receiver devices (or, in case of exemplary DTDoA configurations (
In some embodiments, the position estimate is refined with the new, i.e. updated, weights W′, see e.g. block 208 of
In some embodiments, if a maximum number of iterations is reached or a stopping or termination criterion is met (e.g., the difference between the current location estimate and the previous one is below a threshold), the procedure can be stopped. Otherwise, one could continue, see arrow 212 of
In some embodiments,
In some embodiments,
In some embodiments, a distance estimate at the k-th receiver device RDk may be denoted as {circumflex over (d)}k.
Similarly, in some embodiments, e.g. in case of exemplary DTDoA configurations (
In some embodiments, a reference receiver device may be defined. As an example, in some embodiments, choosing e.g. a first receiver device RD1 as a reference receiver device, the K−1 distance differences associated with the receiver devices can be characterized by {circumflex over (d)}k,1={circumflex over (d)}k−{circumflex over (d)}1, k=2, . . . , K [equation 1].
In some embodiments,
In some embodiments, the instructions 106, when executed by the at least one processor 102, cause the apparatus 100 to determine 222 the initial weights W-INIT based on at least one of: a) a reliability of estimated distances associated with the plurality of receiver devices, b) predetermined initial weight values.
In some embodiments, the reliability of the estimated distances associated with the plurality of receiver devices can e.g. be determined based on knowledge about a received signal strength, RSS, since, in some embodiments, multipath reflections tend to be weaker than a direct (line-of-sight) path.
Hence, in some embodiments, measurements, e.g. distance estimates, whose signal strengths are low, e.g. fall below a predetermined RSS threshold, may be downweighted, e.g. weighted with comparatively small weight factors, whereas measurements, e.g. distance estimates, whose signal strengths are high, e.g. exceed the RSS threshold, may receive more emphasis by using larger weight factors.
In some embodiments, if no RSS information or other information characterizing a reliability of the estimated distances associated with the plurality of receiver devices (or similar) are available, unit initial weights may be used for at least some, for example all, receiver devices.
In some embodiments,
In some embodiments, for each position estimate of the K many position estimates {circumflex over (x)}1, {circumflex over (x)}2, . . . , {circumflex over (x)}K, another receiver device of the K many receiver devices may be used as reference receiver device.
In other words, in some embodiments, instead of determining a unique reference receiver device, which may e.g. be the one that is, for example, a-priori considered the most reliable one (e.g., always the one with the lowest distance estimate) before computing {circumflex over (x)}, it is proposed to use every receiver device of the K many receiver devices as a reference receiver device once to obtain the K many position estimates {circumflex over (x)}1, {circumflex over (x)}2, . . . , {circumflex over (x)}K, e.g. in the form of a matrix with K distinct position estimates {circumflex over (X)}=[{circumflex over (x)}1, {circumflex over (x)}2, . . . , {circumflex over (x)}K].
In some embodiments, an estimate of the mobile device's position x=[x,y,z]T may e.g. be obtained by solving an equation system, e.g. characterized by: Aθ=b+n, [equation 2], with
and with
where d1=∥x−p1∥ (nuisance parameter), and wherein n is a noise vector.
In some embodiments, a closed-form approximate weighted least squares (WLS) solution of equation 2 can be obtained as
θ=(ATWA)−1ATwb, [equation 6],
where W=diag{w2, . . . , wK} is a weighting matrix comprising the weights for weighting distance difference estimates, i.e., {circumflex over (d)}k,1.
In some embodiments, the position estimates {circumflex over (x)} can be determined by adding e.g. the first three elements {circumflex over (θ)}1:3 of the WLS solution from Eq. (6) to the position p1 of a, for example reference, receiver device:
{circumflex over (x)}=p1+{circumflex over (θ)}1:3, e.g. using the WLS estimator according to equation 6.
In some embodiments, the weighted average position estimate may be obtained as
{circumflex over (x)}
WA=Σk=1Kwnorm,k, [equation 8]
where wnorm,k represents a normalized k-th weight (e.g. associated with the k-the receiver device RDk (
In other words, in some embodiments, for determining 232 (
In some embodiments, e.g. in order to detect receiver devices with erroneous measurements, a measure of a reliability of the position estimates of at least some, for example all, receiver device is provided. In some embodiments, the measure of reliability indicates to which extent the respective estimate can be “trusted”.
Hence, in some embodiments,
([equation 9]), wherein {circumflex over (d)}k characterizes an estimated distance associated with the k-th receiver device RDk, wherein d1=∥x−p1∥, wherein {circumflex over (x)}WA characterizes a weighted average position estimate, wherein pk characterizes a position of the k-the receiver device RDk, wherein p1 characterizes a position of a first receiver device RD1.
In some embodiments, based on the residual error ek, updated weights W′ may be determined, see block 242 of
In some embodiments,
In some embodiments, discarding 252 the k-th receiver device RDk comprises disregarding signals or measurements provided by the k-th receiver device RDk, e.g. for determining position estimates.
In some embodiments, the residual error ek as exemplarily characterized by equation 9 represents an averaged deviation between estimated distance differences of K−1 many receiver devices and a reference receiver device, and actual distance differences (i.e., w.r.t. a current estimate {circumflex over (x)}WA) of the K−1 receiver devices and the reference receiver device.
In some embodiments, using the formulation of equation 9, the residual error ek may be interpreted as a measure of how much the measured distance differences agree, e.g. correlate, with a current position estimate {circumflex over (x)}WA or—to put it another way—with the distance differences that would result if the mobile device UE were located at a position characterized by {circumflex over (x)}WA.
In some embodiments, the proposed approach according to the embodiments enables to avoid the choice of a “bad” reference receiver device, wherein e.g. an outlier measurement, which is e.g. associated with a reference receiver device, may severely affect all distance differences e.g. according to equation 1, which, in some embodiments, may e.g. be used to determine a final, e.g. WLS-based, position estimate.
Advantageously, in some embodiments, considering a situation with a “bad” reference receiver device, an associated residual error ek will have a comparatively high value, allowing the detection of receiver devices with e.g. unreliable measurements, and, in some embodiments, downweighting their contribution, e.g. in equation 8, e.g. according to a magnitude of the residual error they produce, or, in some embodiments, discarding such receiver devices, see block 252 of
In some embodiments, e.g. after determining, e.g. computing, the residual error ek, e.g. based on equation 9, e.g. for all position estimates, updated weights W′, which may be used for a next iteration, see arrow 212 of
w
k
=f
e(ek), [equation 10],
where, in some embodiments, fe( ) is a monotonically increasing function that maps the magnitude of the residual errors ek e.g. to numbers between 0 and 1.
In some embodiments, for the mapping of the magnitude of the residual errors ek to the interval between 0 and 1, a function of the following types may be used: Andrew's sine wave, Huber loss, etc.
As already mentioned above, in some embodiments, if the residual error ek exceeds a pre-defined maximum value, the corresponding weight may be set to zero, e.g. such that the measurement—and thus also the receiver device, in some embodiments—is discarded.
In some embodiments, the above-described procedure may be iterated, see for example arrow 212 of
In some embodiments, e.g. after each m-th, m>=1, iteration, a difference of a current position estimate to position estimate of the previous iteration may be determined, e.g. computed, based on Δ=∥{circumflex over (x)}WA,i−{circumflex over (x)}WA,i-1∥, [equation 11].
In some embodiments, the difference Δ may e.g. be compared to a pre-defined threshold ε, e.g. to check convergence.
In some embodiments, based on the comparison of the difference Δ with the pre-defined threshold ε, a termination criterion may be set, e.g. to stop the process, preventing further iterations to be made.
Further embodiments, see
Further embodiments,
Further embodiments relate to a system 1000, 1000′ for estimating a position of a mobile device UE, UE′ which is configured to transmit at least one reference signal RS, comprising at least one apparatus 100, 100′, 10′ according to the embodiments and A) K, K>=2, many receiver devices RD1, . . . , RDK (
Further embodiments,
Also depicted in the present localization scenario are four receiver devices, e.g. LMUs, individually denoted as LMU1, LMU2, LMU3, LMU4 (2-D case for illustrative purposes), where the top right one (LMU1) measures a reflection a2 (only a reflection a2 is available, as the obstacle O blocks a direct LOS path from the true position of the UE to LMU1), producing wrong angular and ToA or TDoA measures.
In
Further embodiments,
Further embodiments relate to a data carrier signal DCS carrying and/or characterizing the computer program PRG.
The principle according to the embodiments provides a novel, for example TDoA-based, location estimation approach that, for example, has the ability to reject outlier measurements (or to reduce an influence of outlier measurements) and/or to weight e.g. the remaining measurements, e.g. according to, i.e. based on, the residual errors they produce.
In some embodiments, the approach may therefore not only be deployed e.g. as a standalone positioning solution, but can, in some embodiments, also be used in a larger framework, e.g. as a “pre-filtering” method that yields locator weights, e.g. for further, e.g. yet more elaborate, e.g. probabilistic, localization algorithms, and, for example, facilitates a convergence of, e.g. such maximum likelihood (ML) techniques, e.g. by providing a proper initial search point, e.g. for gradient optimization.
In the following, a further example, e.g. based on real-world data according to some embodiments, is presented. In some embodiments, an example using real-world data may help understanding the functionality of some exemplary embodiments. Exemplarily, a case with K=6 LMUs and distance estimates {circumflex over (d)}=[19.53 22.06 21.10 17.16 20.94 26.07]m m is considered, where the last measurement “26.07 m” is an artificially created outlier (10 m more than actually measured). In some examples, e.g., taking the last LMU as the reference, leads to an inaccurate position estimate due to the outlier. Taking each LMU, according to some embodiments, as the reference once, by contrast, yields the residual errors e=[4.47 3.26 3.26 3.69 3.55 11.30] m. The residual error of 11.3 m obtained by taking the outlier LMU as the reference clearly stands out as the highest and can, in some embodiments, be removed, e.g. by properly defining the threshold above which weights are set to 0. After discarding this LMU according to some embodiments, the (remaining) residual errors of the next iteration, e.g. after convergence of the algorithm, are e=[0.33 0.56 0.32 0.93 0.31]m. Notice how the residuals of the remaining LMUs also get smaller, since the outlier measurement was removed and does not impact the TDoA measurements and subsequent iterations anymore. As a consequence, in some embodiments, not only a more accurate position estimate in general can be attained by the principle according to the embodiments, but, in some embodiments, the following additional advantages for probabilistic positioning are enabled: Good initial search point to facilitate the convergence, e.g. of gradient optimization, Outlier measurement does not contribute to a ML function.
| Number | Date | Country | Kind |
|---|---|---|---|
| 20225116 | Feb 2022 | FI | national |
