Method for position estimation and front-end calibration with in-situ surveying of a multi-antenna system

Abstract

In the method for position estimation and front-end calibration with in-situ surveying of a multi-antenna system, a GNSS antenna with multiple receiving elements is provided in a one- or multi-dimensional arrangement, the relative antenna pattern of which is surveyed. Satellite navigation signals received by the receiving elements are processed together in a signal processing unit. The relative propagation delays and signal strength differences of the receiving channels, which arise in the analog components of the receiving system, are calibrated. The directions of incidence of the incoming satellite signals in the local coordinate system of the antenna system are determined using satellite and position information, as well as an estimated or predefined position of the multi-antenna system in space.The measured steering vectors, which are obtained from the signals after correlation with the pseudorandom sequences of the satellite signals by averaging or eigenvector determination, are used to measure or improve a previously measured relative antenna pattern.

Description

[0001] The invention relates to a method for attitude estimation and front-end calibration with in-situ measurement of a multi-antenna system.

[0002] When using multi-antenna systems, the signals from Global Navigation Satellite Systems (GNSS) can be used not only for global position, speed, and time measurement, but also for global attitude and orientation estimation of the antenna array. In conventional methods, the directions of arrival of the signals are estimated using the signal propagation time differences between the individual receiving elements. The thus estimated directions of arrival – present in a local coordinate system of the antenna array – are compared with the directions of arrival from a globally referenced coordinate system and used to determine the attitude. The determination of a rotation matrix (as a description of the attitude) is known as Wahba's problem [1], and various methods for its mathematical solution exist [2].

[0003] By processing the GNSS signals, the position of the receiver and the position of the received satellites are known in a global coordinate system. Determining the directions of arrival of the signals in a local, globally referenced coordinate system is therefore easily possible. For example, the East-North-Up (ENU) coordinate system can be used here.

[0004] Other methods determine the position of the antenna array directly from the propagation time and signal strength differences at the receiving elements (described by the spatial signature = steering vector), without the intermediate step of using the globally referenced, estimated directions of arrival. In this case, an auxiliary function is defined that calculates the difference between the measured steering vectors and a simulated steering vector. If sufficient satellite signals and thus steering vectors are available, a position can be determined by minimizing the auxiliary function. The method follows the following scheme: 1. Initialize the position estimation (=predefined start value). 2. Rotate the known directions of incidence using the position estimate. 3. Read / calculate the estimated steering vector using the rotated directions of incidence (see step 2). 4. Calculate the auxiliary function. 5. If the auxiliary function exceeds a certain limit: Adjust the estimate and go back to step 2. Otherwise, the current position estimate is accepted.

[0005] Due to varying signal propagation times in the signal processing channels (in the frontend, for example, due to component tolerances, different cable lengths, or different temperatures), the antenna system must also be calibrated. This calibration can be performed before use of the receiving system (inaccurate due to temporal variation of the system) or together with the attitude estimation (e.g., [3]).

[0006] The simulated steering vectors can also be used for beamforming the incoming signals. Such beamforming shifts the phase of the input signals (the antenna elements) so that signals from the desired direction of incidence constructively overlap.

[0007] Both approaches described have in common that the spatial signature of the antenna array (= time of flight and signal strength difference between the receiving elements) must be known for all relevant directions of incidence. The summary of these spatial signatures is referred to below as the antenna pattern. Based on the geometry of the antenna array (= position of the receiving elements relative to each other), the spatial signatures can be mathematically calculated assuming isotropic antenna radiators. However, due to technical imperfections and tolerances, irregularities in and interactions between the receiving elements, as well as environmental influences, the measurable spatial signatures of the antenna array deviate from the theoretical values. Since this deviation significantly impairs the accuracy of the position estimation, the antenna pattern is usually measured before deployment.

[0008] Currently, there are two approaches to measuring the antenna pattern of a GNSS antenna: V1 In an anechoic chamber, a transmitting antenna with precisely known characteristics sends signals to the antenna under test, and the signal strength and signal propagation time are measured using a network analyzer [4]. Depending on the application, the transmitting antenna moves on a sphere around the antenna under test, or the antenna under test is rotated. V2 In the open field using GNSS signals. The antenna to be measured is mounted on a robotic arm whose position is precisely known. The received GNSS signals are analyzed differentially over several days. Due to the daily repeating constellation of GPS satellites, sources of error (e.g., multipath) can be eliminated. The flexible positioning of the robotic arm allows any desired direction of incidence to be measured [5].

[0009] The known methods were designed to measure individual antennas absolutely, i.e., to determine a phase center and its variations. For applications involving an antenna array (e.g., determining the direction of incidence, orientation, beamforming), however, it is not the absolutely measured antenna pattern that is important, but rather the relative propagation time and signal strength differences between the receiving elements, i.e., a relatively measured antenna pattern. If the known methods are used to measure an antenna array, the individual antenna elements of the array are measured one after the other. Interactions between the antennas are therefore not fully represented. In addition, process-related inaccuracies when combining the individual measurements to create a complete antenna pattern of the array lead to deviations in the pattern.

[0010] Both of these antenna measurement methods must be performed at the site of use prior to installation (V1, because the antenna must be placed in a measurement chamber, and V2, because the antenna must be mounted on a robot arm). This carries the risk that the antenna pattern may deviate at the site of use. It also makes re-measurement of the pattern, for example, due to age-related deviations, very difficult.

[0011] In addition, both systems use a mechanical rotation unit, which affects the accuracy of the measurements.

[0012] In addition, there are individual disadvantages for the above-mentioned first approach known in the state of the art for measuring the antenna pattern of a GNSS antenna.

[0013] Depending on the size of the measurement chamber, measuring large antennas and objects in their near field (e.g., tripods, car roofs, etc.) is not possible. Furthermore, the network analyzer uses synthetic signals (usually sine waves) at a specified frequency to measure the pattern. The antenna patterns measured in this way differ from those observed using GNSS signals.

[0014] The object of the invention is to provide an improved method for attitude estimation and front-end calibration with in-situ measurement of a multi-antenna system.

[0015] To achieve this object, the invention provides a first method for attitude estimation and front-end calibration with in-situ measurement of a multi-antenna system, wherein the method - a GNSS antenna with multiple receiving elements in a one- or multi-dimensional array is provided, - Satellite navigation signals received by the receiving elements are processed jointly in a signal processing unit (typically the acquisition, tracking, and decoding of the navigation signals so that the position, speed, and time of the receiver can be determined. The position of the satellite signals (more precisely, the position of the satellites) can also be calculated from the navigation message. Signal processing is performed jointly for each antenna element so that the delays between the antenna elements can be accurately determined. The correlation of all antenna signals is thus performed using the same PRN code, but again separately for each satellite signal. Not every satellite signal is tracked individually for each antenna signal.) - the relative propagation delays and signal strength differences of the receiving channels, which arise in the analogue components of the receiving system, are calibrated, - the directions of incidence of the incoming satellite signals in the local coordinate system of the antenna system are determined by using satellite and position information, as well as an estimated or predetermined position of the multi-antenna system in space, - the measured steering vectors, which are obtained from the signals after correlation with the pseudorandom sequences of the satellite signals by averaging or eigenvector determination, are used to measure or improve a previously measured or otherwise obtained relative antenna pattern (The steering vectors describe the delays of an incoming signal between the individual antenna elements. These are determined as the output signal after the correlation unit. For the received satellite signals, the Doppler signal is removed (Doppler wipeoff) and this signal is correlated with the satellite-typical PRN code. Because we are dealing with a multi-antenna receiver, we not only receive one value for each complete correlation, but also a vector that includes the delays between the antenna elements.The propagation delays mentioned above must be subtracted from this vector, and then this vector can be saved as a steering vector. A brief general description of an isotropic steering vector can be found at Antenna-Theory.com - Steering Vector.

[0016] The method according to the invention is divided into two parts, whereby the first step can be skipped depending on the variant (see below). In the first step, the position of the antenna array and the calibration vector of the receiving system are estimated. For this purpose, a developed method is used that is similar to the method described in [3], but differs in the following points: 1. Instead of an isotropic antenna model, a scanned antenna pattern is used for the method. This applies not only to the pattern itself but also to the mathematical derivation of the pattern, which is calculated numerically in the developed method. The advantage of this adaptation is that even antennas that deviate significantly from the isotropic model can be used. Furthermore, the algorithm can also incorporate changes in the antenna pattern during operation. 2. The absolute phase and amplitude of the satellite signals are no longer iteratively estimated in the developed method. Instead, the absolute phase and amplitude of the estimated steering vector are adjusted to the absolute phase and amplitude of the measured steering vector using the pseudoinverse. This approach reduces the number of required iterations, since the absolute phase of the satellite signals, in particular, changes over time and therefore must be reestimated with each new attitude estimation.

[0017] The second step of the developed method involves using the measured steering vector and the information from the first step (see above) to measure the antenna pattern. The antenna pattern is thus measured in situ during operation. Any incoming satellite signal can be used for the measurement. The exact procedure is as follows: 1. The relative propagation delay and relative gain of the receive channels are calculated from the measured satellite steering vector. These are either known from the first step of the procedure in the form of the calibration vector or are specified and assumed to be static. 2. From the position of the receiver and the positions of the satellites, the directions of arrival are calculated in a globally referenced coordinate system. From these, the directions of arrival in the local coordinate system of the antenna array are then calculated using the known attitude estimate (first step of the procedure or predefined). 3. The calibrated, measured steering vectors (see above under 1.) are saved in the antenna pattern for the corresponding directions of incidence (also calculated as described above under 2.). If measurements already exist for the corresponding direction of incidence, the saved value is calculated from the new and existing measurements, e.g., by averaging. To reduce the number of directions of incidence to be measured, the direction of incidence can be rounded.

[0018] For the in-situ measurement of the antenna pattern, two different variants of the procedure are conceivable: 1. The receiver remains static, and the method is given a fixed position and calibration. The method measures the antenna pattern as described above. 2. The receiver uses an approximate pattern to determine the position and simultaneously measures the antenna pattern. a) The approximate pattern is an (incompletely) measured pattern, which was determined using the method from variant 1. When determining the position, it is advisable to give significantly lower weight to unmeasured directions of incidence than to those for which reliable corrections already exist. b) The approximate pattern is a measured pattern which was determined using the conventional methods (see above under V1 and V2). c) The approximate pattern is a pattern that was calculated / simulated based on the geometry of the antenna array.

[0019] According to the invention, the relative antenna pattern does not necessarily have to have been measured in advance, since the method according to the invention also functions under the assumption of an isotropic antenna radiator (e.g., the isotropic model of the GNSS antenna). Instead of measurement, the geometrically calculable antenna pattern can also be used. Alternatively, it is also possible to completely dispense with a model of the GNSS antenna if the position of the GNSS antenna is estimated in advance (e.g., by determining the directions of incidence of the incoming satellite signals in the local coordinate system of the antenna system using satellite and position information as well as an estimated or predetermined position of the multi-antenna system in space).

[0020] According to an advantageous embodiment of the invention, it is provided that the position of the GNSS antenna in the form of a rotation matrix and the front-end calibration of the receiving system in the form of a complex-valued calibration vector is determined by an iterative method.

[0021] According to a further advantageous embodiment of the invention, the determination is carried out using the globally referenced directions of incidence of the satellite navigation signals and / or the measured steering vectors known by the signal processing unit.

[0022] According to a further advantageous embodiment of the invention, it is provided that the measured relative antenna pattern is used in operational mode for, in particular, improved determination of the directions of incidence.

[0023] According to a further advantageous embodiment of the invention, it is provided that the measured relative antenna pattern is used in operational mode for, in particular, improved determination of the position and front-end calibration of the antenna array.

[0024] According to a further advantageous embodiment of the invention, it is provided that the measured relative antenna pattern is used in operational mode for, in particular, improved beam forming of the input signals.

[0025] In a further expedient embodiment of the invention, it can be provided that the measured relative antenna pattern is used in operational mode for, in particular, improved spatial interference signal suppression in the input signals.

[0026] The following is based on Fig. 1 an embodiment of the invention is described in more detail.

[0027] In Fig.Figure 1 shows the incorporation of the inventive method into a multi-antenna receiver. The lines of varying thickness symbolize multi-channel parallel connections, with the number of individual channels being designated by N (number of antenna elements, e.g., four) and M (number of satellites, e.g., six or fifteen).

[0028] The drawing shows the embedding of the method in a multi-antenna receiver. The signals from the N antenna elements are processed in an RF frontend (e.g., amplified and downconverted to an intermediate frequency) and subsequently digitized using an analog-to-digital converter. Digital preprocessing (e.g., signal filtering, interference suppression) can then take place. Each of the N signals from the antenna elements thus processed contains all M received satellite signals, which are separated by a code division multiplexing method (CDMA). These N signals are then processed separately for each of the M satellite signals. For this purpose, each of the N signals is correlated with the pseudorandom sequence of the satellite signal to be processed. When using the code division multiplexing method, this can also be referred to as despreading. The invention is not limited to the use of the code division multiplexing method, but utilizes it to advantage.Alternatively, a frequency division multiplexing method can be used.

[0029] For the method to be applicable, the signal processing for each of the N input signals must be synchronous for each satellite, i.e., all input channels must be correlated with the same time offset or the delay must be corrected. Only in this way can a relative measurement of the antenna pattern be achieved. The information obtained from the signal processing of the satellite signals is then used to determine position, speed, and time, as in a conventional GNSS receiver. This is done using the satellite positions calculated in this process (the navigation message sent by the satellites contains parameters that allow the satellite positions to be calculated over a certain period of time. This is known to every user, but must also be calculated by each user (for position determination) from the parameters).In addition, a direction of incidence can be calculated in a global coordinate system. The correlation results of the various satellites and the directions of incidence can then be used to perform the described procedure, ie, an attitude estimation and in-situ measurement of the antenna pattern.

[0030] The method according to the invention can be applied to safety-critical and highly automated systems which require robustness against disturbances and at the same time require high-precision positions or high-precision attitude estimation (e.g. time receivers, receivers on UAVs, GNSS-based aviation systems for precise time-of-flight landing, receivers on satellites, etc.). BIBLIOGRAPHY [1] G. Wahba, “A Least Squares Estimate of Satellite Attitude,” SIAM Rev., Vol. 7, No. 3, pp. 409-409, July 1965, doi: 10.1137 / 1007077 [2] J. L. Farrell, J. C. Stuelpnagel, R. H. Wessner, J. R. Velman, und J. E. Brook, „A Least Squares Estimate of Satellite Attitude (Grace Wahba)“, SIAM Rev., Bd. 8, Nr. 3, S. 384-386, Juli 1966, doi: 10.1137 / 1008080 [3] S. Zorn, M. Niestroj, S. Caizzone, M. Brachvogel, und M. Meurer, „Selfcontained Antenna Crosstalk and Phase Offset Calibration by Jointly Solving the Attitude Estimation and Calibration Problem“, in Proceedings of the 30th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2017), Portland, Orgeon, USA, Sep. 2017, S. 3481-3493. doi: 10.33012 / 2017.15382 [4] P. Zeimetz, „On the accuracy of absolute GNSS antenna calibration and the conception of a new Anechoic Chamber“, 2008. Zugegriffen: 27. Februar 2023. [Online]. Verfügbar unter: https: / / www.semanti-cscholar.org / paper / On-the-accuracy-of-absolute-GNSS-antenna-and-theof-Zeimetz / 1454e587c2f85624ec461b8bcf193e3009ed5b6a [5] G. Wübbena and M. Schmitz, “Automated Absolute Field Calibration of GPS Antennas in Real-Time” [6] T. Bamberg, A. Konovaltsev, and M. Meurer, “GNSS Antenna Arrays and RTK: The Effect of Carrier Phase Distortions Induced by Spatial Filtering on RTK Positioning,” in Proceedings of the 2022 International Technical Meeting of The Institute of Navigation, Long Beach, California, Jan. 2022, pp. 799-811. doi: 10.33012 / 2022.18161 QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited non-patent literature

[0000] G. Wahba, „A Least Squares Estimate of Satellite Attitude“, SIAM Rev., Bd. 7, Nr. 3, S. 409-409, Juli 1965, doi: 10.1137 / 1007077

[0030] J. L. Farrell, J. C. Stuelpnagel, R. H. Wessner, J. R. Velman, und J. E. Brook, „A Least Squares Estimate of Satellite Attitude (Grace Wahba)“, SIAM Rev., Bd. 8, Nr. 3, S. 384-386, Juli 1966, doi: 10.1137 / 1008080

[0030] S. Zorn, M. Niestroj, S. Caizzone, M. Brachvogel, und M. Meurer, „Selfcontained Antenna Crosstalk and Phase Offset Calibration by Jointly Solving the Attitude Estimation and Calibration Problem“, in Proceedings of the 30th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2017), Portland, Orgeon, USA, Sep. 2017, S. 3481-3493. doi: 10.33012 / 2017.15382

[0030] P. Zeimetz, „On the accuracy of absolute GNSS antenna calibration and the conception of a new Anechoic Chamber“, 2008. Zugegriffen: 27. Februar 2023. [Online]. Verfügbar unter: https: / / www.semanti-cscholar.org / paper / On-the-accuracy-of-absolute-GNSS-antenna-and-theof-Zeimetz / 1454e587c2f85624ec461b8bcf193e3009ed5b6a

[0030] G. Wübbena und M. Schmitz, „Automated Absolute Field Calibration of GPS Antennas in Real-Time"

[0030] T. Bamberg, A. Konovaltsev, und M. Meurer, „GNSS Antenna Arrays and RTK: The Effect of Carrier Phase Distortions Induced by Spatial Filtering on RTK Positioning“, in Proceedings of the 2022 International Technical Meeting of The Institute of Navigation, Long Beach, California, Jan. 2022, S. 799-811. doi: 10.33012 / 2022.18161

[0030]

Claims

[1] Method for attitude estimation and front-end calibration with in-situ measurement of a multi-antenna system, wherein the method - a GNSS antenna with multiple receiving elements in a one- or multi-dimensional array is provided, - satellite navigation signals received by the receiving elements are processed jointly in a signal processing unit, - the relative propagation delays and signal strength differences of the receiving channels, which arise in the analogue components of the receiving system, are calibrated, - the directions of incidence of the incoming satellite signals in the local coordinate system of the antenna system are determined by using satellite and position information, as well as an estimated or predetermined position of the multi-antenna system in space, - the measured steering vectors, which are obtained from the signals after correlation with the pseudorandom sequences of the satellite signals by averaging or eigenvector determination, are used to measure or improve a previously measured or otherwise obtained relative antenna pattern. [2] Method according to claim 1, characterized by that the position of the GNSS antenna in space is determined in the form of a rotation matrix and the front-end calibration in the form of a complex-valued calibration vector by an iterative procedure. [3] Method according to claim 2, characterized by that the determination is made using the globally referenced directions of incidence of the satellite navigation signals and / or the measured steering vectors known by the signal processing unit. [4] Method according to one of claims 1 to 3, characterized bythat the measured relative antenna pattern is used in operational use, in particular to improve the determination of the directions of incidence. [5] Method according to one of claims 1 to 4, characterized by that the measured relative antenna pattern is used in operational mode, in particular to improve the determination of the position and front-end calibration of the antenna array. [6] Method according to one of claims 1 to 4, characterized by that the measured relative antenna pattern is used in operational mode, in particular to improve the beamforming of the input signals. [7] Method according to one of claims 1 to 4, characterized by that the measured relative antenna pattern is used in operational mode, in particular to improve spatial interference signal suppression in the input signals.

Images