EP-4067928-B1 - METHOD FOR DETERMINING AN ANGLE OF ARRIVAL, DEVICE, COMPUTER PROGRAM PRODUCT AND NON-VOLATILE STORAGE MEDIUM

Inventors

• MERK, TIMON
• MAHLIG, MATTHIAS
• KARLSSON, PETER
• Rezai, Farshid

Dates

Publication Date

20260506

Application Date

20210329

Claims (12)

1. Method (100) for determining an angle of arrival, AoA, (6) of radio frequency, RF, measurement signals received from a transmitter, the method comprising: - obtaining (101) measurement data based on the received RF measurement signals from an antenna array (5), wherein the RF measurement signals are representative of multiple frequency channels, - determining (102) power spectra, comprising determining at least one power spectrum for each of the multiple frequency channels by using the measurement data, - determining (104) a quality indicator by using the measurement data, wherein the quality indicator is representative of an integrity of the RF measurement signals received from the transmitter, - providing (105) a machine learning algorithm, which is pre-trained to determine an AoA (6) based on power spectra of multiple frequency channels, and - determining (106) the AoA (6) of the RF measurement signals received from the transmitter by using the machine learning algorithm and the determined power spectra and the quality indicator as input features of the machine learning algorithm; wherein - determining (102) the power spectra comprises performing power spectral density, PSD, -based signal processing, which comprises at least one of the following: - a multiple signal classification, MUSIC, algorithm; - a propagator direct data acquisition, PDDA, algorithm; - a self-signal suppression, SSS, algorithm.
2. The method (100) according to claim 1, wherein each power spectrum is a power spectral density, PSD.
3. The method (100) according to any of claims 1 or 2, wherein the quality indicator comprises one or more of a received signal strength indicator, RSSI, and/or a channel state information, CSI, and/or a cyclic redundancy check, CRC, result, and/or a forward correction error, FEC, result and/or a noise analysis result of a constant tone extension.
4. The method (100) according to any of claims 1 to 3, wherein the measurement data comprises in-phase and quadrature, IQ, data.
5. The method (100) according to any of claims 1 to 4, wherein an amount of the determined power spectra is dependent on an amount of the multiple frequency channels and/or a number of antenna rows of the antenna array (5) and/or a polarization of the antenna array (5).
6. The method (100) according to any of claims 1 to 5, wherein the machine learning algorithm comprises a classification and regression tree, CART, algorithm and/or a convolutional neural networks, CNN, and/or a multilayer perceptron, MLP, neural network and/or a random forest algorithm and/or a clustering algorithm and/or a support support-vector machine, SVM, algorithm.
7. The method (100) according to any of claims 1 to 6, wherein the method (100) further comprises: - concatenating (103) the determined power spectra, and - determining (106) the AoA (6) of the received RF measurement signals by using the machine learning algorithm and the concatenated power spectra.
8. The method (100) according to any of claims 1 to 7, wherein the antenna array (5) is positioned in a first environment, in particular when the RF measurement signals are received, and the machine learning algorithm is pre-trained by using pre-train data, wherein the pre-train data is determined based on the antenna array (5) and the first environment.
9. The method (100) according to any of claims 1 to 8, wherein the machine learning algorithm is pre-trained by using pre-train data, wherein the pre-train data is determined based on multiple antenna arrays and multiple second environments.
10. A device (4) comprising a processing unit, wherein - the processing unit is configured to: - obtain measurement data based on RF measurement signals received from a transmitter via an antenna array (5), wherein the RF measurement signals are representative of multiple frequency channels, - determine power spectra, comprising determining at least one power spectrum for each of the multiple frequency channels by using the measurement data, - determine a quality indicator by using the measurement data, wherein the quality indicator is representative of an integrity of the RF measurement signals received from the transmitter, - the processing unit comprises a machine learning algorithm, which is able to determine an AoA (6) based on power spectra of multiple frequency channels, and when the machine learning algorithm is pre-trained, the processing unit is configured to determine the AoA (6) of the received RF measurement signals by using the pre-trained machine learning algorithm and the determined power spectra and the quality indicator as input features of the machine learning algorithm; and - determining (102) the power spectra comprises performing power spectral density, PSD, -based signal processing, which comprises at least one of the following: - a multiple signal classification, MUSIC, algorithm; - a propagator direct data acquisition, PDDA, algorithm; - a self-signal suppression, SSS, algorithm.
11. A computer program product comprising instructions which, when executed by a computing device, cause the computing device to carry out the method (100) according to any of claims 1 to 9.
12. A non-volatile storage medium comprising a computer program product according to claim 11.

Description

TECHNICAL FIELD This disclosure relates to a method for determining an angle of arrival of received radio frequency measurement signals. The disclosure further relates to a device comprising a processing unit. The disclosure also relates to a computer program product and a non-volatile storage medium. BACKGROUND ART In wireless communications radio frequency, RF, signals are often transmitted from a transmitter to a receiver. When receiving such an RF signal, an angle of arrival, AoA, of the received RF signal often needs to be determined. Different signal processing algorithms may be used for determining the AoA. However, different positioning problems may remain, such as a poor AoA estimation, for example in multipath propagation due to reflections and/or diffractions and/or scattering. Furthermore, when determining the AoA, there may exist environment specific disturbances such as so-called Non-Line of Sight, NLoS, measurements and/or so-called Obstructed Line of Sight, OLoS, measurements. Furthermore, the determination of the AoA may be influenced by imperfections relating to an antenna array used for receiving the RF signal and/or relating to RF design. Document US 2020/0191913 A1 discloses a wireless scanning system comprising a transmitter, a receiver, and a processor. The transmitter is configured for transmitting a first wireless signal using a plurality of transmit antennas towards an object in a venue through a wireless multipath channel of the venue. The receiver is configured for: receiving a second wireless signal using a plurality of receive antennas through the wireless multipath channel between the transmitter and the receiver. The processor is configured for obtaining a set of channel information of the wireless multipath channel based on the second wireless signal received by the receiver, and generating an imaging of the object based on the set of channel information. Document WO 2021/016003 A1 discloses a method for estimating the AoA of RF signals that are received by an antenna array. A plurality of RF signal power measurements are received by a plurality of antenna elements at a plurality of RF channels. An estimated RF signal parameter value is computed by applying a machine learning model to the plurality of RF signal power measurements. It is an object of the present disclosure to present a method, a device, a computer program product and a non-volatile storage medium, in which a reliable and robust determination of an angle of arrival of received radio frequency measurement signals is provided. SUMMARY OF INVENTION The above-mentioned object is solved by the subject-matter of the attached independent claims. Further embodiments are disclosed in the attached dependent claims. According to a first aspect of the invention, a method for determining an angle of arrival, AoA, of received radio frequency, RF, measurement signals comprises the steps: obtaining measurement data based on the received RF measurement signals from an antenna array, wherein the RF measurement signals are representative of multiple frequency channels,determining power spectra, comprising determining at least one power spectrum for each of the multiple frequency channels by using the measurement data,providing a machine learning algorithm, which is pre-trained to determine an AoA based on power spectra of multiple frequency channels, anddetermining the AoA of the received RF measurement signals by using the machine learning algorithm and the determined power spectra. The machine learning algorithm according to the first aspect may also be denoted as a machine learning model and/or a data driven model. An advantage of the method according to the first aspect is that a robust and precise estimation of the AoA of the received RF measurement signals may be provided. By using a data driven model, a poor AoA determination, for example in multipath propagation due to reflections and/or diffractions and/or scattering, may be avoided. Furthermore, AoA determination under environment specific disturbances such as NLoS measurements and/or OLoS measurements may be improved. Furthermore, an influence on the AoA determination due to an imperfection of the antenna array and/or imperfect RF design, may be minimized. The use of the data driven model according to the first aspect may be more advantageous for a robust and precise determination of the AoA, compared to using conventional AoA signal processing methods. To overcome the limitation that different antenna array types have a different geometry, such as a different number of antenna elements and/or a different size and/or a different spacing and/or a different configuration, according to the first aspect, the input features of the machine learning algorithm are defined, such that an input feature of same dimension across different antenna arrays is obtained. For example, a non-linear transformation of obtained measurement data such as IQ samples, may be transformed into an