Method for including bistation paths in angle estimation with cooperative radar sensor

Abstract

A method for a collaborative radar sensor network having a plurality of individual radar sensors for estimating a radar target angle or for preparing a radar sensor network for this purpose. An assignment relationship is determined that assigns a control vector with a bistatic component for each of the plurality of angles. A measurement vector having a bistatic component associated with the detected radar target is calculated. According to some aspects, a bistatic l-th component of a control vector corresponds to a product resulting from a respective single-station first component of a control vector of a respective first radar sensor and a respective single-station second component of a control vector of a respective second radar sensor. The bistatic lth component of the measurement vector corresponds to a product resulting from the measurement values of the redundant bistatic channel associated with the radar target. According to other aspects, a bistatic lth component of a control vector corresponds to a product resulting from a respective first component of a transmit control vector of a respective first radar sensor and a respective second component of a receive control vector of a respective second radar sensor.

Description

Field of the Invention

[0001] The present invention relates to angle estimation by means of a collaborative radar sensor network. In particular, the present invention relates to a method for a collaborative radar sensor network having a plurality of individual radar sensors, the method being for preparing the collaborative radar sensor network for bistatic angle estimation. Further, the present invention relates to a method for a collaborative radar sensor network having a plurality of individual radar sensors, the method being for estimating the angle of a radar target. Background Art

[0002] In motor vehicles, radar sensors are used to implement comfort functions such as adaptive cruise control and safety functions such as emergency braking assistance. An important advantage of such sensors is that they directly measure physical quantities rather than, for example, interpreting images from cameras. The radar sensor emits high-frequency radar beams through its antenna array and receives the beams reflected from objects. Here, the detected objects can be stationary or moving. The distance to the object and the direction (angle) relative to the object can be calculated by means of the received radar beams. In addition, the speed of the object relative to the radar sensor can be calculated.

[0003] When performing angle estimation, the received signal is compared with the antenna pattern associated with the previously measured angle. For the determination of only a single target (or multiple targets, where the multiple targets can be clearly distinguished from each other based on distance and relative speed), the estimated angle results as the position of the best agreement between the received signal and the antenna pattern. For the case of general multi-target estimation, some special estimation algorithms are known, which provide estimated values for the positioning angles of all the participating targets.

[0004] The effective power of environmental detection can be significantly increased by using a collaborative radar sensor network Here, the data of multiple radar sensors are analyzed and utilized jointly and phase-coherently, which on the one hand achieves increased sensitivity of environmental detection and on the other hand achieves increased accuracy of environmental detection, because overall there is a larger antenna array and more HF channels available than in a single sensor.

[0005] Current non-phase-coherent radar sensor networks for motor vehicles usually operate in a single-station manner. In other words, radar targets are detected separately by the respective radar sensors, which emit signals and receive the signals reflected by the radar targets. In angle-resolved measurements, the receiving antenna elements on the respective radar sensors are arranged at different positions in such a direction: in this direction, the radar sensor performs angle resolution. For an idealized, near-point-like radar target at a respective angular position, there is a characteristic phase and amplitude relationship between the signals received in different receiving antenna elements for the transmitted signal. Here, the amplitude ratio between the received signals depends on the direction angle and the sensitivity curve of the receiving antenna elements. By analyzing and utilizing the phase relationship and / or amplitude relationship, the angular position of the radar target to be determined can be determined. For the angle estimation of the detected radar target, the signal processing steps for each combination of the transmitting and receiving antenna elements of the respective individual sensors are usually carried out separately. Such a combination corresponds to the so-called single-station path of the radar sensor network. Then, the angle estimation is performed based on the values jointly associated with the detected target for the spectra of the individual single-station paths of the radar sensors. The calibration of the respective radar sensors is carried out by measuring the antenna pattern of the respective radar sensors, in particular by creating a calibration matrix for the single-station path of the radar sensor.

[0006] A sensor system and a method for calibrating a radar sensor node of a sensor system are known from DE 102019201138 A1, in which two radar sensor nodes are arranged spaced apart from each other on a vehicle. The transmitting and receiving elements of the first radar sensor node form the node-internal field of the aperture elements of a virtual receiving aperture, and the transmitting element of the first radar sensor node and the receiving element of the second radar sensor node form the cross-node field of the aperture elements of a virtual receiving aperture. The spacing between the two radar sensor nodes and the spacing between the transmitting and receiving elements within the first radar sensor node are selected such that the position of the first aperture element of the node-internal field corresponds to the position of the second aperture element of the cross-node field. Thereby, the calibration of the first radar sensor node can be transferred to the second radar sensor node, thus avoiding the laborious calibration of multiple radar sensor nodes.

[0007] In a radar sensor network operating in a phase-coherent and collaborative manner, the radar sensors operate phase-coherently or phase-synchronously in the system composite structure, so that the transmitted signals of the individual radar sensors have a fixed phase relationship relative to each other. Therefore, in addition to the radar signals related to the individual sensors of the single-hop paths, the two-hop paths can also be analyzed and utilized, in which the radar signal is transmitted from the transmitting antenna element of one radar sensor and received by the receiving antenna element of another radar sensor. However, compared with individual radar sensors, collaborative radar sensor networks are mostly very large and their final geometric arrangement can only be achieved through installation (for example, installed in a vehicle). Therefore, it is impossible or difficult to perform calibration measurements with a collaborative radar sensor network, especially in an antenna measurement chamber, in order to create a calibration matrix for the two-hop paths or the entire radar sensor network. Summary of the Invention

[0008] The object of the present invention is to provide a method for a collaborative radar sensor network that achieves an improvement in the effective power during angle estimation and / or greatly reduces the calibration cost of a system operating in a collaborative manner.

[0009] In particular, the object of the present invention is to achieve an angular analysis and utilization of the two-hop paths of a collaborative radar sensor network without measuring the calibration matrix of the two-hop paths or the entire radar sensor network.

[0010] Another object of the present invention is to achieve an angular analysis and utilization of the full aperture of a collaborative radar sensor network without measuring the calibration matrix of the two-hop paths or the entire radar sensor network.

[0011] According to a first aspect of the present disclosure, at least one of the above-mentioned objects is solved by a method for a collaborative radar sensor network having a plurality of individual radar sensors, the method being for preparing the collaborative radar sensor network for two-hop angle estimation, wherein the method includes: determining an assignment relationship that assigns a corresponding control vector to each of a plurality of angles, wherein the corresponding control vector has at least a two-hop component, and wherein the l-th two-hop component of the control vector corresponds to the product resulting from: the corresponding single-hop first component of the control vector of the corresponding first radar sensor among the radar sensors, wherein the single-hop first component corresponds to the single-hop path from the m-th transmitting antenna element to the n-th receiving antenna element of the first radar sensor; and the corresponding single-hop second component of the control vector of the corresponding second radar sensor among the radar sensors, wherein the single-hop second component corresponds to the single-hop path from the m-th transmitting antenna element to the n-th receiving antenna element of the second radar sensor.

[0012] Here, l is a natural number. Further, m and n are natural numbers. For example, for multiple l, the l-th component of the bistatic of the control vector can respectively correspond to the product as given, where, for each l, the numbers m and n can correspond to different combinations of the transmitting and receiving antenna elements of the associated radar sensor.

[0013] In particular, the cooperative radar sensor network can be a phase-coherent cooperative radar sensor network. In a phase-coherent cooperative radar sensor network, the individual radar sensors have a relationship in which the phases of their transmitted signals are fixed relative to each other. Here, the individual radar sensors can have the same fundamental frequency (radar fundamental frequency) or different fundamental frequencies.

[0014] Associated with the cooperative radar sensor network, the term "bistatic" refers to the radar signal path from the transmitting antenna element through the radar target to the receiving antenna element, in which the transmitting antenna element and the receiving antenna element belong to different radar sensors of the radar sensor network. Such a path can also be referred to as a bistatic path or as a path that bridges the participating radar sensors. Here, the corresponding channel (analysis and utilization channel or transmit-receive channel) is called a bistatic channel. Different from this, the term "monostatic" refers to a radar signal path in which the transmitting antenna element and the receiving antenna element belong to the same radar sensor of the radar sensor network. Such a path can also be referred to as a monostatic path or as a path limited to a single radar sensor. The corresponding channel (analysis and utilization channel or transmit-receive channel) is called a monostatic channel here. If not otherwise stated, then here the control vector, channel, and signal path are understood as the transmit-receive control vector, transmit-receive channel, and transmit-receive signal path.

[0015] Bistatic angle estimation is understood to include angle estimation (estimation of the angle of the radar target) of signals including at least one bistatic path of the radar sensor network. In particular, the method can be a method for preparing a cooperative radar sensor network for bistatic angle estimation of signals including signals of (multiple) bistatic paths of the radar sensor network. The angle to be estimated can be, for example, the azimuth angle. The angle to be estimated can also be the elevation angle or a combination formed by the elevation angle and the azimuth angle.

[0016] The radar sensors each have a plurality of transmit antenna elements and / or a plurality of receive antenna elements. In particular, the radar sensors each have an antenna array, which includes a plurality of transmit antenna elements and / or a plurality of receive antenna elements. The plurality of transmit antenna elements and / or the plurality of receive antenna elements are arranged at different positions on the respective radar sensors in such a direction that the radar sensor or the cooperative radar sensor network is angle-resolving. For example, the radar sensors each have at least N tx transmit antenna elements and at least N rx receive antenna elements. Preferably, the N rx receive antenna elements on the respective radar sensors are arranged at different positions in such a direction that the radar sensor or the cooperative radar sensor network is angle-resolving. In a preferred embodiment, each individual radar sensor in a MIMO (multiple-input multiple-output) radar sensor has at least N tx transmit antenna elements and at least N rx receive antenna elements. In order to be able to separate the signals of the transmit antenna elements on the receive antenna elements, the transmit signals must be uncorrelated (orthogonal). This can be achieved by time-division multiplexing, frequency-division multiplexing, or code-division multiplexing.

[0017] Hereinafter, the bistatic component of the product of the respective monostatic components of the control vector corresponding to the relevant radar sensor is also referred to as the bistatic product component.

[0018] The control vector gives the geometric characteristics and wave propagation characteristics of the respective antenna array under consideration. Knowledge of the control vector enables the establishment of a (uni-valued under suitable conditions) relationship between the angle of the object (or "radar target") and the received signal and the inference of the angle of the object from the amplitude and phase relationships of the received signals. In particular, the control vector gives the expected phase relationship between the components of the respective measurement vector. The individual components of the control vector can in particular give the expected amplitude and the expected phase position of the signal of the channel (transmit-receive channel) or the channel product respectively assigned to the radar sensor for the angle assigned to the control vector. In the case of the bistatic product component of the control vector, this relates to the signals of the bistatic channels and the respective bistatic paths respectively assigned to the radar sensor network. The control vector including the bistatic product component can also be referred to as the product control vector. In addition to the bistatic component, the respective control vector can also have a monostatic component, which can correspond to the respective monostatic component of the control vector of a single radar sensor. The respective control vector having a bistatic component can be referred to as the bistatic control vector.

[0019] The assignment relationship that assigns a corresponding control vector to each of multiple angles can be, for example, a control matrix and can also be referred to as a calibration matrix or an antenna pattern. Hereinafter, the determination of the assignment relationship can also be referred to as the creation of the calibration matrix. The corresponding rows of the control matrix can correspond to the control vectors assigned to the corresponding angles.

[0020] The l-th bistatic component of the control vector is understood as the l-th component of the control vector, i.e., one of the components of the control vector, where this component is a bistatic component. It corresponds to the (complex) product of the corresponding monostatic first component of the control vector of the corresponding first radar sensor and the corresponding monostatic second component of the control vector of the corresponding second radar sensor. In particular, it can be a product. This product includes at least one addition of the relevant phases of the corresponding monostatic components of the considered control vector. For example, the mentioned monostatic components of the control vector of the radar sensor can be complex and give an amplitude and a phase. Then, the product of the two monostatic components corresponds to the product of their amplitudes and the addition of their phases. Ignoring the corresponding amplitudes, the mentioned component of the control vector can also only give a single phase. In this case, the product of the two monostatic components corresponds to the addition of their phases.

[0021] The method can include: storing or preparing, for all virtual channels and angle racks (e.g., in a memory), the calibration matrix of a single sensor and / or the element-by-element product of the corresponding matrix.

[0022] The method has the advantage that, in order to prepare a cooperative radar sensor network for bistatic angle estimation, an assignment relationship is determined that assigns a corresponding control vector to each of multiple angles, where the control vector can be calculated based on the monostatic components of the control vectors of individual radar sensors. Thus, a control vector with bistatic components can be calculated and the assignment relationship can be determined without the need for laborious measurements of the bistatic components of the installed radar sensor network. In particular, the bistatic components of the corresponding control vector can be calculated based on the (monostatic) control vectors of individual radar sensors. For example, the control vectors or calibration matrices or antenna patterns assigned to the corresponding angles of individual radar sensors can be determined in a common way by calibration measurements of individual radar sensors. They can be measured, for example, at the factory for each individual sensor before installing the radar sensor network. For example, the determination of the assignment relationship can be carried out before actually using the radar sensor network, e.g., after installing the cooperative radar sensor network.

[0023] Thus, the determination of the assignment relationship allows for the realization of angle analysis of the bistatic paths of a cooperative radar sensor network or the preparation of a cooperative radar sensor network for bistatic angle estimation without measuring the bistatic paths or the calibration matrix of the entire radar sensor network. This realizes an improvement in the effective power of the cooperative radar sensor network during angle estimation. Since the bistatic paths can also be used during angle analysis, a much larger virtual antenna aperture can be spanned. The calibration of the entire cooperative system can be obtained based on the monostatic calibration of the individual sensors used.

[0024] Here, it is fully utilized that if the spacing or the spatial offset between the m-th transmitting antenna element and the n-th receiving antenna element of the first radar sensor corresponds to the spacing or the spatial offset between the m-th transmitting antenna element and the n-th receiving antenna element of the second radar sensor (as is usually the case, for example, in two structurally identical radar sensors or radar sensors with identically constructed antenna arrays), then the thus determined assignment relationship and the control vector are particularly well-suited for use in angle estimation. Because then it is assumed that since the two bistatic virtual channels correspond to the same virtual spatial antenna positions, the two bistatic virtual channels (the two bistatic virtual channels corresponding to the bistatic path from the m-th transmitting element of the first radar sensor to the n-th receiving element of the second radar sensor or from the m-th transmitting element of the second radar sensor to the n-th receiving element of the first radar sensor) are redundant. Thus, the complex spectral values of the bistatic virtual channels on the virtual sensor exist redundantly in terms of their spatial positions. Then, for an ideal radar target detected at an angular position, it holds that the l-th bistatic component of the control vector assigned to the angle corresponds to the product resulting from: the measured value assigned to the radar target of the bistatic channel (the bistatic channel corresponding to the m-th transmitting antenna element of the first radar sensor and the n-th receiving antenna element of the second radar sensor); and the measured value assigned to the radar target of the bistatic channel (the bistatic channel corresponding to the m-th transmitting antenna element of the second radar sensor and the n-th receiving antenna element of the first radar sensor). This equality holds for the virtual spatial antenna positions. Thus, the product of the spectral values measured for the radar target in the two redundant bistatic channels in the d,v spectrum (spacing - velocity spectrum) can be analyzed using the corresponding bistatic components of the control vector assigned to the respective angle for angle estimation. In the present disclosure, for the sake of simplicity of illustration, a radar target in the far field is taken as the starting point.

[0025] Hereinafter, a single radar sensor including channels of a single station is also referred to as an actual radar sensor. Different from this, a bistatic virtual sensor (or radar sensor) includes bistatic channels corresponding to a bistatic path, and the bistatic path is generated by a combination of a transmitting antenna element and a receiving antenna element of two different actual radar sensors. Hereinafter, a single radar sensor and a bistatic virtual sensor are generally also referred to as an actual or virtual sensor (or radar sensor) and can also be referred to as an actual or virtual sensor node. Here, different from the overall aperture of a radar sensor network, the aperture of a single actual and bistatic virtual sensor is also referred to as a sub-aperture.

[0026] According to a second aspect of the present disclosure, at least one of the above-mentioned tasks is solved by a method for a cooperative radar sensor network having a plurality of single radar sensors, the method for estimating an angle of a radar target, wherein the method includes: calculating a measurement vector assigned to the detected radar target, wherein the measurement vector has at least a bistatic component, and wherein the l-th bistatic component of the measurement vector corresponds to a product generated by: a measurement value assigned to the radar target of a bistatic channel (the bistatic channel corresponding to the m-th transmitting antenna element of the corresponding first radar sensor of the radar sensor and the n-th receiving antenna element of the corresponding second radar sensor of the radar sensor); and a measurement value assigned to the radar target of a bistatic channel (the bistatic channel corresponding to the m-th transmitting antenna element of the second radar sensor and the n-th receiving antenna element of the first radar sensor), wherein the method further includes: estimating an angle of the detected radar target, wherein the estimated angle is determined based on a result of an association between the measurement vector and a control vector assigned to different angles, or at least based on a product of the measurement vector and the control vector.

[0027] The method according to the second aspect is complementary to the method according to the first aspect because the method according to the first aspect is used to prepare for the method according to the second aspect. The estimation of the angle is particularly performed in an actual use of the radar sensor network. The following explanation supplements the above explanation of the first aspect, and the above explanation of the first aspect is not repeated here.

[0028] In order to detect a radar target, a baseband signal is generated by mixing the received signal received by the receiving antenna element with the transmitted signal transmitted from the transmitting antenna element, and the baseband signal is scanned and analyzed for utilization. The frequency and phase of the baseband signal correspond to the frequency difference and phase difference between the signal transmitted at a given point in time and the signal received at the same point in time. Due to the frequency modulation of the transmitted signal, this frequency difference depends on the round-trip time of the signal from the radar sensor to the object and back and thus on the distance of the object. However, due to the Doppler effect, the frequency difference also contains a component determined by the relative velocity of the object.

[0029] The method described and the determination of the measurement vector for the detected radar target can advantageously be used especially in a cooperative radar sensor network with FMCW radar sensors, which operate with so-called fast chirp sequences. Here, a large number of frequency ramps (chirps) are traversed quickly and continuously, which have a large slope and only a relatively short duration. Then, for one channel, the d,v spectrum can be calculated by Fourier transform, in which the detected radar target detected as a peak is assigned spectral bin coordinates, which respectively correspond to the distance d or the relative velocity v.

[0030] Thus, for the detected radar target (object), a measurement vector can be obtained, the components of which correspond to different configurations of the transmitting antenna elements and / or receiving antenna elements of the radar sensor network. The measurement vector has at least bistatic components. The bistatic components of the measurement vector correspond to the products of the respective channels (transmit-receive channels) of the associated bistatic virtual sensors, more precisely, the (complex) products generated from the measured values (especially two-dimensional spectra or spectral values of the d,v spectra of the bistatic channels) assigned to the detected radar target of the respective bistatic channels. In particular, it can be a product. The l-th bistatic component of the mentioned measurement vector is understood as the l-th component of the measurement vector, i.e., one of the components of the measurement vector, where this component is a bistatic component. In particular, i.e., the measurement vector can be composed and / or calculated from measured values existing as vectors of, for example, individual transmit-receive channels, where the product components are calculated for the bistatic components. The product includes at least one addition of the relevant phases of the measured values assigned to the radar target of the considered bistatic channels. For example, the mentioned measured values of the considered bistatic channels can be complex and give an amplitude and a phase. Then, the product of two measured values corresponds to the product of their amplitudes and the addition of their phases. Ignoring the respective amplitudes, the mentioned measured values of the considered bistatic channels can also give only a phase. In this case, the product of two measured values corresponds to the addition of their phases.

[0031] The measurement vector can further have a monostatic component, which can correspond to the respective channels (transmit-receive channels) of the associated single radar sensor. That is, one component of the measurement vector corresponds to the given product of the bistatic channels in the case of bistatic and can correspond to the monostatic channels of the respective radar sensor in the case of monostatic. The components of the measurement vector are based on such measurement values: the measurement values are obtained in the respective channels for the detected radar target. For a single radar sensor, the components of the measurement vector can correspond to such measurement values: the measurement values are obtained in the respective channels of the radar sensor for the detected radar target.

[0032] As described above, the measurement values of two virtual bistatic channels (the channels corresponding to the bistatic paths from the m-th transmit element of the first radar sensor to the n-th receive element of the second radar sensor or from the m-th transmit element of the second radar sensor to the n-th receive element of the first radar sensor) may be redundant. In other words, the associated channels may be redundant. Due to the correspondence described above between the l-th bistatic component of the control vector assigned to the angle and the corresponding l-th bistatic component of the measurement vector, the spectral values in the d,v spectrum measured for the radar target in the two redundant bistatic channels can be analyzed and utilized with the corresponding bistatic components of the control vector assigned to the respective angle for angle estimation, where the bistatic components of the control vector can correspond to the product of the monostatic components of the control vectors of the individual radar sensors and can thus be determined without measuring the bistatic paths or the calibration matrix of the entire radar sensor network.

[0033] Therefore, a measurement vector with bistatic components corresponding to the mentioned product is particularly advantageous for angle estimation because it allows angle estimation including the bistatic paths without having to measure the bistatic paths or the calibration matrix of the entire radar sensor network.

[0034] In some embodiments, for each actual and virtual sensor, a corresponding measurement vector is obtained, where the measurement vector of the virtual sensor has bistatic components. In other embodiments, the measurement vector includes the channels of each of the actual and virtual sensors. The method can include: performing radar measurements to determine the azimuth of the radar target. The method or the performance of the radar measurements can include: calculating at least the range-velocity spectrum (d,v spectrum) in the bistatic channels.

[0035] For angle estimation, the following situation is fully utilized: the amplitude relationship and phase relationship of the signals obtained in at least different bistatic channels depend in a characteristic manner on the angle of the radar target. The phase change curve on the channels of the corresponding actual or virtual sensors is determined by the angle of the (ideal) radar target. Through correlation, the phase information of the individual components of the measurement vector enters the angle estimation by taking into account the corresponding control vector. The correlation of the measurement vector assigned to the detected radar target with the control vectors assigned to different angles and the determination of the estimated angle based on the result of the correlation can also be referred to as: checking how strongly the measurement vector assigned to the detected radar target is correlated with the control vectors assigned to different angles, and estimating the angle of the detected radar target based on the result of the check. The correlation of the measurement vector with the control vectors assigned to different angles can include multiplying the measurement vector with the control vectors assigned to different angles, and the result of the multiplication is the angle spectrum assigned to the detected radar target. For example, the control matrix can be multiplied with the measurement vector. In the simplest case, the angle estimation can be achieved by performing a maximum search in the corresponding angle spectrum. To perform the correlation, a so-called DML function (Deterministic Maximum Likelihood function) can be constructed, which gives how strongly the actually measured amplitude relationship and phase relationship for the radar target are correlated with the control vectors (i.e., the theoretically amplitude relationship and phase relationship for different angle hypotheses). Then, an angle hypothesis (in which the correlation is maximum) constitutes the best estimate value for the angle of the object. For the same bistatic virtual array, computationally efficient methods such as FFT or the so-called Matrix-Pencil-Methode can be used.

[0036] The determination of the estimated angle based at least on the product of the measurement vector and the control vector can in particular be based on the product of the measurement vector and a control vector from a certain number of control vectors, where the certain number of control vectors are assigned to different angles. Similarly, by determining the estimated angle based at least on the product of the measurement vector and the control vector, the phase information of the individual components of the measurement vector also enters into the estimation of the angle by taking into account the control vector. The determination of the estimated angle based at least on the product of the measurement vector and the control vector can in particular include: determining the estimated angle based at least on the product of the measurement vector and the control vector, where the control vector is assigned to an angle corresponding to an angle hypothesis. The angle hypothesis can in particular be determined based on the association of the measurement vector assigned to the detected radar target of at least one individual radar sensor with the control vectors assigned to different angles of the individual radar sensor. That is, starting from an angle hypothesis (using a measurement vector that only includes monostatic components, for example) that is the result of the association or the angle estimation of an individual radar sensor, the phase information of the measurement vector including bistatic components can be analyzed and utilized in order to achieve angle estimation while taking into account the bistatic channels. However, the angle hypothesis can also be determined, for example, based on the association of the measurement vector assigned to the detected radar target of at least one bistatic virtual sensor with the control vectors assigned to different angles of the bistatic virtual sensor, for example in a way that is referred to here as the "first association".

[0037] The bistatic component of the product of the respective measurement values corresponding to the bistatic channels of the radar sensor network of the measurement vector is also referred to below as the bistatic product component or the channel product component, and this product is also referred to below as the channel product.

[0038] In an advantageous configuration, the method according to the second aspect further includes: preparing a cooperative radar sensor network for bistatic angle estimation according to the method of the first aspect.

[0039] In particular, in the association of the measurement vector assigned to the detected radar target with the control vectors assigned to different angles, the control vectors can be assigned to the corresponding angles by the determined assignment relationship. In particular, the estimation of the angle of the detected radar target can be carried out based on the determined assignment relationship.

[0040] According to a third aspect of the present disclosure, at least one of the above-mentioned tasks is solved by a method for a collaborative radar sensor network having a plurality of individual radar sensors, the method for preparing the collaborative radar sensor network for bistatic angle estimation, wherein the method comprises: determining an assignment relationship that assigns a corresponding control vector to each of a plurality of angles, wherein the corresponding control vector has at least a bistatic component, and wherein the l-th bistatic component of the control vector corresponds to the following product: the corresponding first component of the transmit control vector of the corresponding first radar sensor among the radar sensors, wherein the first component corresponds to the m-th transmit antenna element of the first radar sensor; and the corresponding second component of the receive control vector of the corresponding second radar sensor among the radar sensors, wherein the second component corresponds to the n-th receive antenna element of the second radar sensor.

[0041] According to the third aspect, thus the bistatic component of the control vector is used, and the bistatic component does not correspond to the channel product of two bistatic channels, but corresponds to the product of the components of the transmit control vector and the receive control vector. The transmit control vector is a unidirectional control vector that gives the characteristics of the transmit antenna pattern of the associated transmit antenna element. The receive control vector is a unidirectional control vector that gives the characteristics of the receive antenna pattern of the associated receive antenna element. In addition, the method according to the third aspect corresponds to the method according to the first aspect. The following explanations supplement the above explanations of the first aspect, and the explanations of the first aspect are not repeated here.

[0042] The assignment relationship that assigns a corresponding control vector to each of a plurality of angles can again be, for example, a control matrix and can also be referred to as a calibration matrix or an antenna pattern.

[0043] The mentioned l-th bistatic component of the control vector is a bistatic component. It corresponds to the product generated by the corresponding first component of the transmit control vector of the corresponding radar sensor and the corresponding second component of the receive control vector of the corresponding second radar sensor. In particular, it can be a product. The product includes at least one addition of the relevant phases of the corresponding components of the considered transmit control vector and the considered receive control vector.

[0044] Therefore, in order to prepare the collaborative radar sensor network for bistatic angle estimation, a combination of the corresponding calibration values of the paths of the transmit antenna elements and the receive antenna elements assigned to the corresponding radar sensors can be performed.

[0045] In particular, the l-th component of the bistatic control vector can correspond to the product resulting from the following: the corresponding, measured first component of the transmission control vector of the corresponding first radar sensor in the radar sensor (in particular: the component determined in the calibration measurement), where the first component corresponds to the m-th transmission antenna element of the first radar sensor; and the corresponding, measured second component of the reception control vector of the corresponding second radar sensor in the radar sensor (in particular: the component determined in the calibration measurement), where the second component corresponds to the n-th reception antenna element of the second radar sensor.

[0046] When calibrating the corresponding transmission antenna elements (TX calibration) of the individual sensors related to calibration, the signals or electromagnetic waves emitted from the corresponding transmission antenna elements (TX antennas) can be received, for example, at different angles by the antennas (TRX antennas) of the calibration receiver synchronized with the individual sensors. Here, complex amplitude values and in particular phase values can be determined for all channels assigned to the corresponding paths from the transmission antenna elements to the calibration antenna (or to the calibration receiver). When calibrating the corresponding reception antenna elements (RX calibration), a signal synchronized with the individual sensors (i.e., a wave phase-coherent with the individual sensors) can be emitted from the same calibration antenna (TRX antenna), which is then received and detected by the reception antenna elements (i.e., all RX channels). Here, complex reception values are also determined for all reception antenna elements. After calibrating the individual sensors, there are the transmission phases and reception phases of all individual sensors, and the two-way calibration values for all monostatic but also bistatic, virtual channels can be determined computationally.

[0047] This method has the advantage that, by combining the corresponding components (or calibration values) of the transmission control vector and the reception control vector for the corresponding transmission antennas and reception antennas, in particular the corresponding phase values for the components of the bistatic control vector can be determined. Thus, the bistatic channels can be calibrated without the need for laborious measurements of the bistatic components of the installed radar sensor network. The transmission control vector or transmission calibration matrix or transmission antenna pattern and the reception control vector or reception calibration matrix or reception antenna pattern assigned to the corresponding angles of the individual radar sensors can be determined by one-way calibration measurements of the individual radar sensors. They can be measured for each individual radar sensor at the factory before installing the radar sensor network. This means that in the case of factory calibration separated into RX paths and TX paths, the phase values exist individually for each transmission antenna or reception antenna, i.e., not as two-way measurements. For example, the determination of the assignment can be carried out before actually using the radar sensor network, for example, after installing a cooperative radar sensor network.

[0048] Thus, the determination of the assignment relationship allows for the implementation of angle analysis of the bistatic paths of a cooperative radar sensor network or for preparing the cooperative radar sensor network for bistatic angle estimation without measuring the bistatic paths or the calibration matrix of the entire radar sensor network. The control vector with bistatic components can be calculated and the assignment relationship determined without the need for very laborious and expensive measurements of the components of the installed radar sensor network or the bistatic components of the overall system (if this is possible at all). Thus, the calibration costs can be significantly reduced. By also being able to use the bistatic paths during angle analysis, a much larger virtual antenna aperture can be spanned. Thus, by performing separate RX calibration and TX calibration for the individual sensors and their combinations for monostatic and bistatic paths, the calibration matrix for the full antenna aperture can be obtained. Based on the monostatic calibration of the individual sensors used, the calibration of the entire cooperative system can be derived.

[0049] According to a fourth aspect of the present disclosure, at least one of the above-mentioned tasks is solved by a method for a cooperative radar sensor network having a plurality of individual radar sensors for estimating the angle of a radar target, the method comprising: preparing the cooperative radar sensor network for bistatic angle estimation according to the method according to the third aspect; determining a measurement vector assigned to the detected radar target, wherein the measurement vector has at least bistatic components, wherein the l-th bistatic component of the measurement vector corresponds to the measurement value assigned to the radar target of a bistatic channel, the bistatic channel corresponding to the m-th transmit antenna element of a corresponding first radar sensor in the radar sensors and the n-th receive antenna element of a corresponding second radar sensor in the radar sensors; and estimating the angle of the detected radar target, wherein the estimated angle is determined based on the result of the association of the measurement vector with control vectors assigned to different angles or at least based on the product of the measurement vector and the control vectors.

[0050] In particular, when associating the measurement vector assigned to the detected radar target with control vectors assigned to different angles, the control vectors can be assigned to the corresponding angles by the determined assignment relationship. In particular, the angle of the detected radar target can be estimated based on the determined assignment relationship.

[0051] According to the fourth aspect, thus, a measurement vector with bistatic components is used, the bistatic components not corresponding to the channel product of two bistatic channels but only corresponding to the bistatic channels of the radar sensor network. Furthermore, the method according to the fourth aspect corresponds to the method according to the second aspect.

[0052] In particular, angle estimation is performed during the actual use of a radar sensor network. The following explanations supplement the above explanations regarding the first, second, third, and fourth aspects, and these explanations are not repeated here.

[0053] Advantageous developments and configurations of the present invention are described below. In addition, the following describes advantageous features, developments, and configurations of the present invention. These can be used separately in the methods according to each of the aspects described herein.

[0054] The method may include: storing an assignment relationship that assigns a corresponding control vector to each of a plurality of angles. For example, before a cooperative radar sensor network is put into operation, the assignment relationship can be stored on the factory side.

[0055] The method may in particular be a method for a cooperative radar sensor network for a vehicle, in particular a motor vehicle.

[0056] The "monostatic exploitation" of a bistatic, virtual sensor is achieved or carried out by the method described above. This means that now, the bistatic reflection can additionally be exploited. However, nevertheless, it is not yet possible to exploit the detected angles on a finer, coherently cooperative angular grid based on the total antenna aperture composed of the actual physical sensors and the bistatic sensors. For this purpose, not only must the virtual sensor also be spatially positioned between the physically existing sensors, but the existing physical sensors must also be spatially positioned relative to each other, and thus the relative phase relationship between the sub-apertures must be determined.

[0057] In an advantageous configuration, the method comprises: a first association of a first measurement vector of at least one bistatic, virtual sensor assigned to a detected radar target with a control vector of the bistatic, virtual sensor assigned to different angles, wherein the bistatic, virtual sensor corresponds to a configuration of a bistatic channel of a respective first radar sensor and a respective second radar sensor, and wherein the measurement vector of the bistatic, virtual sensor has a bistatic component; and / or a second association of a second measurement vector of at least one single radar sensor assigned to the detected radar target with a measurement vector of the single radar sensor assigned to different angles, wherein the estimation of the angle of the detected radar target comprises: determining a phase vector that assigns respective phases to at least one bistatic, virtual sensor and at least one single radar sensor, the respective phases being obtained at least based on the result of at least one of the first association and the second association; a third association of the determined phase vector with phase vectors assigned to different angles, the phase vectors assigned to different angles giving a phase relationship based on their spatial deviation relative to each other between at least one single radar sensor and at least one bistatic, virtual sensor; determining an estimated value with multiplicity for the angle of the radar target according to the result of the third association; and resolving the multiplicity of the estimated value for the angle of the radar target according to the result of the first or second association.

[0058] The previously mentioned association (determining the estimated angle based on the result of the previously mentioned association) may include the first association or be the first association.

[0059] The phase vector assigns respective phases to at least one bistatic, virtual sensor and at least one single radar sensor, and the respective phases are obtained based on at least the result of at least one of a first association and a second association. In some embodiments, in particular, the respective phases are obtained based on the product of the measurement vector of at least one single radar sensor or at least one bistatic, virtual sensor and the respective control vectors (the respective control vectors being assigned to angles corresponding to angle hypotheses), where the angle hypotheses are obtained based on the result of at least one of the first association and the second association. In particular, the respective components of the phase vector can correspond to the product generated by the measurement vector of the relevant actual or virtual sensor and the control vector assigned according to the angle hypothesis, where the angle hypothesis is determined based on the result of the relevant association. Here, the respective angle hypotheses for the respective virtual or actual radar sensors can be determined respectively based on the result of the respective first or second association, or the angle hypothesis can be determined based on one of the first or second associations; in the latter case, the same angle hypothesis can thus be used for multiple actual or virtual sensors, and the control vectors assigned based on the angle hypothesis of the relevant actual or virtual sensors are used respectively.

[0060] The method can include: determining an angle hypothesis based on the result of the first association; or determining an angle hypothesis based on the result of the second association.

[0061] In some embodiments, the method includes a first association and a second association, and the phase vector assigns respective phases, and the respective phases are obtained based on the results of the respective first and second associations. In particular, the phase vector assigns a phase to the respective bistatic, virtual sensor, and the phase is obtained based on the result of the relevant first association, and the phase vector assigns a phase to the respective single radar sensor, and the phase is obtained based on the result of the relevant second association.

[0062] In other embodiments, the method includes a first association or a second association, and the phase vector assigns respective phases, and the respective phases are obtained based on the result of the first association respectively, or the phase vector assigns respective phases, and the respective phases are obtained based on the result of the second association respectively. For example, the angle hypothesis can be determined according to the result of the first or second association, and the phase is obtained based on the angle hypothesis, and the phase vector assigns the phase to at least one bistatic, virtual sensor and at least one single radar sensor. The first association or the second association can be part of the respective angle estimation for the detected radar target and the relevant bistatic, virtual sensor or single radar sensor.

[0063] The first association and / or the second association may include calculating an angular spectrum assigned to the detected radar target. The angular hypothesis may correspond to the angle at which the angular spectrum has the maximum absolute value. The phase assigned to the relevant sensor may correspond to the phase of the corresponding complex number of the maximum absolute value of the relevant angular spectrum. The phase vector may be a vector whose components correspond to the phase values determined by the respective maxima of the angular spectrum.

[0064] The phase vectors assigned to different angles can be calculated from the known positions of the actual and virtual sensors relative to each other or from their positions determined as given below. These relative positions are also referred to as the spatial deviations relative to each other between the respective sensors.

[0065] The positions of the actual and virtual sensors relative to each other can create a position vector, the components of which can be understood as the mounting positions of the respective relevant sensors. The mounting positions of the respective sensors can be determined by measurement and / or by calculation. The geometric relationship between the relevant sensors relative to each other can result in a position estimation matrix from the position vector. The position estimation matrix can have the following form, where d is the position vector and α is the respective angle of the radar target, and λ is the radar wavelength:

[0066]

[0067] To calculate the position vector, in particular the position estimation matrix, in a similarly constructed antenna array of radar sensors, any reference position on the respective sensor can be used as a reference point for the respective sensor. For example, the respective center of gravity, in particular the phase center of the reference channel, the respective edge of the sensor, or the respective determined antenna element of the sensor can be selected.

[0068] Thus, the rows of the position estimation matrix each correspond to the phase vectors assigned to different angles, which each give the angle-dependent phase relationships resulting from their spatial deviations relative to each other between the relevant radar sensors. The measurement results for the physical as well as virtual sensors or sub-apertures are interpolated by the position vector to improve the accuracy of the total aperture.

[0069] The third association may include: determining an angular spectrum (spectrum of the position vector) by multiplying a phase vector with a position estimation matrix. Since only the positions of the sensors are considered for the third association and the relative spacing of the positions of the sensors (or the spacing of the sub-apertures) can be multiple times the wavelength of the received signal, this angular spectrum may have multi-valuedness. These spacings can be several 10λ long or longer. Here, the characteristic of the multi-valuedness is that for multiple angles, especially for periodic angles, values with the same amplitude appear, and these values correspond to the detected radar targets. Since only positions are considered here and sensor characteristics are not considered, this spectrum is not weighted by the envelope of a single sensor. It follows that the amplitude values for all possible angles are the same.

[0070] The resolution of the multi-valuedness can be understood as selecting one estimated value from the multi-valued estimated values of the angles for the radar target. For the common analysis and utilization of the two spectra of a single sensor and a position vector, a region around the peak can be selected in the single spectrum and superimposed with the same region in the position vector spectrum. Here, it is important that only a single unique "needle" of the position spectrum is located in the region to be analyzed and utilized. Based on the position of the needle of the position vector spectrum, an angle is selected in the single sensor spectrum: this angle corresponds to the accuracy of the total aperture for the best estimate of the angle. The resolution is carried out according to the results of the first and / or second association, especially according to the angular spectrum (the angular spectrum at least includes the surroundings around the maximum (in absolute terms) of the angular spectrum). Here, the surroundings are selected in such a way that exactly one of the multi-valued estimated values falls into the surroundings. In other words, by multiplying the results of the first and / or second association with the results of the third association, an angular refinement of the results of the first and / or second association can be achieved, and thus an angle can be determined with a refined angular resolution at which the radar target is detected.

[0071] Therefore, the spectrum of the single sensor angle estimation (virtually, bistatically or physically) and the multi-valued spectrum of the position vector of the virtual sensor can be used. In particular, for example, the corresponding actual or virtual radar sensors may have an angular resolution limited by the aperture of the relevant sensors; by analyzing and utilizing the phase vectors by means of the phase vectors assigned to individual angles, the phase information of a larger aperture from a cooperative radar sensor network can be analyzed and utilized and the angular resolution can be refined.

[0072] The method is particularly computationally efficient and enables a cost-effective implementation of angle estimation over the entire collaborative antenna aperture, since for example an analysis utilization of the angle spectrum for a limited angular region and / or a single radar sensor or virtual sensor suffices, and yet a high angular resolution is achieved through a third correlation. In particular, the angle analysis utilization can be performed at full aperture without the need for an explicit calibration matrix or control matrix for the full aperture to exist. The control vectors with bistatic components can be calculated and the assignments determined without the need for very time-consuming and expensive measurements of the bistatic components of the installed radar sensor network or the entire system (if this is possible at all). Thus, the calibration effort can be significantly reduced.

[0073] The method includes the determination of phase vectors. The phases can in particular each correspond to the product of the respective measurement vectors assigned to the detected radar target of a single sensor or a bistatic, virtual sensor and the respective control vectors assigned to the angles corresponding to the respective angle hypotheses, where the respective angle hypotheses are obtained according to the respective results of a first correlation and / or a second correlation.

[0074] Thus, by knowing the exact spatial positions of the virtual, bistatic sensors and virtual channels in space, angle estimation can be achieved based on the entire collaborative radar network.

[0075] The method described (in which a phase vector is determined and a third correlation is performed) also constitutes the invention independently of the first to fourth aspects. Thus, according to a fifth aspect of the present disclosure, there is provided a method for use in a cooperative radar sensor network having a plurality of individual radar sensors for estimating the angle of a radar target, the method comprising: determining a measurement vector assigned to the detected radar target, wherein the measurement vector has at least a bistatic component; and estimating the angle of the detected radar target, wherein the estimated angle is determined based on the result of the correlation of the measurement vector with a control vector assigned to a different angle, or at least based on the product of the measurement vector and the control vector, wherein the method comprises: a first correlation of a first measurement vector assigned to the detected radar target of at least one bistatic, virtual sensor with a control vector assigned to a different angle of the bistatic, virtual sensor, wherein the bistatic, virtual sensor corresponds to the configuration of the bistatic channel of a corresponding first radar sensor and a corresponding second radar sensor, and the measurement vector of the bistatic, virtual sensor has a bistatic component; and / or a second correlation of a second measurement vector assigned to the detected radar target of at least one individual radar sensor with a control vector assigned to a different angle of the individual radar sensor, wherein the estimation of the angle of the detected radar target comprises: determining a phase vector, which assigns respective phases to at least one bistatic, virtual sensor and at least one individual radar sensor, the respective phases being obtained at least based on the result of at least one of the first correlation and the second correlation; a third correlation of the determined phase vector with phase vectors assigned to different angles, the phase vectors assigned to different angles giving the phase relationship resulting from their spatial deviation relative to each other between at least one individual radar sensor and at least one bistatic, virtual sensor; determining an estimate of the angle of the radar target with multiplicity based on the result of the third correlation; and resolving the multiplicity of the estimate of the angle of the radar target based on the result of the first or second correlation. The method according to the fifth aspect can be used together with the methods according to the first to fourth aspects.

[0076] In the embodiment just described, the measurement vector of the bistatic, virtual sensor or the actual radar sensor is analyzed and utilized with the corresponding control vector. In other embodiments, measurement vectors or corresponding control vectors that include both monostatic components and bistatic components can be used.

[0077] In particular, according to the first, second, third or fourth aspect, the corresponding control vector assigned to the angle can consist of a monostatic component and a bistatic component, wherein the corresponding monostatic component is a component of the control vector of the relevant radar sensor in the radar sensors.

[0078] The control vectors assigned to the angles, in particular the corresponding rows of the control matrix, can include monostatic components and bistatic components. In other words, the corresponding control vectors can each include a monostatic path of the components assigned to the relevant radar sensor and a bistatic path of the components assigned to the relevant radar sensor. The angle grid can correspond to the calibration step of a single sensor.

[0079] The method has the advantage that the calibration matrix or the control matrix for a fully virtual aperture can be determined without measuring the full aperture or the bistatic channels. Here, the calibration matrix for the full antenna aperture can be determined by calculation based on the two-way calibration of a single sensor. In particular, the calibration matrix for the total aperture can be determined based on the positions of the virtual and physical sensors and the calibration matrices (assignment of control vectors) of the physical and virtual bistatic sensors. The calibration matrix for the full antenna aperture can also be determined based on an ideal, calculated calibration matrix, and the correction factor can be determined by an online calibration method. The calibration matrix is determined by the angular granularity of the single sensor calibration.

[0080] According to the first or second aspect, the component assigned to the bistatic path can be, for example, a multiplicative component, which corresponds to a combination of redundant bistatic paths of the channels assigned to the relevant radar sensors in particular.

[0081] In some embodiments, the method further includes: correcting the phase of the components of the corresponding control vector according to a corresponding phase correction, the corresponding phase correction corresponding to such a corresponding phase shift: for a corresponding angle, the corresponding phase shift is generated by the spatial deviation between the corresponding sensors relative to each other in at least one single radar sensor and at least one bistatic, virtual sensor, wherein the bistatic, virtual sensor corresponds to the configuration of the bistatic channels of the corresponding first radar sensor and the corresponding second radar sensor. Therefore, the phase progression is corrected based on the positioning of the sensor / virtual sensor at the corresponding positions of the control vector or the control matrix.

[0082] The phase correction can be determined and / or considered as an assignment relationship between the angle and the phase correction vector or as a phase correction matrix.

[0083] Therefore, by knowing the exact spatial positions of the virtual, bistatic sensors and the virtual channels in space, angle estimation based on the entire cooperative radar network is achieved. In particular, a (corrected) calibration matrix for the entire aperture can be created.

[0084] For a corresponding angle, the phase offset can in particular be determined by the phase relationship between two (actual and / or virtual) sensors, in particular by the phase relationship resulting from the spatial deviation of at least one single radar sensor and at least one bistatic, virtual sensor relative to each other. In particular, for a sensor, the corresponding phase correction of the components of the control vector assigned to an angle can be the same, that is to say, all components of the control vector of the sensor are corrected with the same phase correction, which depends on the angle and on the sensor.

[0085] The corresponding phase correction can include a corresponding phase correction value of the form e -j·2·π·k resulting from the geometric relationship between the relevant sensors, in particular from their respective spatial deviations, where k is an integer. The correction factor can be determined by means of online calibration or on-site calibration.

[0086] The corresponding phase correction can be determined based on the measurement of the detected radar target. Here, this measurement is also referred to as a stitch measurement (Stichmessung). In particular, a correlation can be established between the angle and the phase position for a single stitch measurement, and it can be analytically transferred to all other angles by means of trigonometry and knowledge of the deviation of the sensors. Thus, the phase deviation (including, for example, the phase deviation due to the line length) between individual radar sensors can be corrected. Alternatively, by knowing the position of the virtual channels on the total aperture (or the exact position of the virtual or physical sensors), the phase correction or the phase correction vector or the phase correction matrix can be calculated purely analytically.

[0087] In the case of the bistatic multiplication component of the control vector, the phase offset can be twice the phase difference resulting from the path length difference caused by the angle-based geometry.

[0088] Not only in the embodiments according to one of the first to fifth aspects, which include the third correlation step of the phase vector, but also in the embodiments according to one of the first to fourth aspects, which include the correction of the phase of the components of the corresponding control vector, the spatial deviation between two corresponding sensors relative to each other can be determined as given below. In particular, the individual radar sensors can thus be positioned accurately relative to each other, and thus the virtual, bistatic sensors can be positioned accurately relative to the individual radar sensors.

[0089] In some embodiments, the method further includes: determining a spatial deviation relative to each other between two corresponding sensors among a plurality of individual radar sensors and at least one bistatic, virtual sensor based on signals of detected radar targets from two sensors that are respectively assigned to different angles, wherein the determining includes: examining the correlation between the phase (especially: the measured phase) of the signal assigned to a corresponding radar target of one of the two corresponding sensors (especially: the signal at the spatial position of the virtual antenna element) and the extrapolated, spatially phase change curve of the signal assigned to the radar target of the other of the two corresponding sensors; and determining the spatial deviation relative to each other between the two corresponding sensors based on the result of the examination, for example, based on the highest correction quality among a plurality of detected radar targets and respectively a plurality of deviation hypotheses or correction positions. In some embodiments, the two corresponding sensors include two of the plurality of individual radar sensors. In some embodiments, the two corresponding sensors include one of the individual radar sensors and one of the at least one bistatic, virtual sensor.

[0090] The method may further include: determining a spatial deviation relative to each other between two corresponding sensors including a bistatic, virtual sensor and one of the two corresponding individual radar sensors based on the spatial deviation determined relative to each other between two individual radar sensors. This can be done by calculation. Here, the bistatic, virtual sensor includes at least one bistatic channel that is assigned to two individual radar sensors. In other words, the bistatic, virtual sensor corresponds to at least one bistatic channel (or a plurality of bistatic channels) of two individual sensors. Thus, if the relative positions of the actual sensors relative to each other have been determined, the position of the corresponding virtual sensor can be accurately calculated therefrom.

[0091] Therefore, for each of a plurality of detected radar targets (test targets) assigned to different angles, it is examined how strongly the phase of the signal assigned to the corresponding radar target of the sensor (e.g., at the corresponding virtual antenna position) correlates with the extrapolated, spatially phase change curve of the signal assigned to the same radar target of other radar sensors. Based on the result of the examination, a deviation can be determined at which the extrapolated, spatially phase change curve is maximally consistent with the measured phase.

[0092] In particular, the extrapolated spatially phase change curves are those that extend beyond the range of the sensor or its actual or virtual antenna array. The phase change curves can be extrapolated periodically, in particular sinusoidally or generally as a harmonic oscillation. When extrapolating, it is fully utilized that, for example, via a plurality of virtual antenna positions or receiving antenna elements of a relevant physical sensor, in particular when at least one spacing between two virtual antenna elements is less than or equal to λ / 2, the phase change curve on the sensor can be unambiguously determined. This is the case for all such sensors: in the sensor, at least one spacing between at least two of the virtual antenna elements is designed to be λ / 2. Typically, this is the case for all common radar sensors. On this premise, for different test angles, the phase can be "unrolled" spatially beyond a single sensor, that is, extrapolated as a harmonic oscillation, in particular up to a reference point, such as the reference point of a bistatic virtual sensor with a known phase. Since this spatial phase progression depends on the angle, this process must be repeated for different test angles. The phase at the reference point is known from the measurement. Now, the spatial position can be determined on the spatially unrolled phase at which the phase coincides with the phase sought at the reference point. Thereby, for different test angles, a large number of hypotheses for the position of the reference point are generated. Correspondingly, the point in the spatially unrolled phase at which most or all of the hypotheses for the test angle / test target coincide is used as the reference point or position hypothesis for determining the deviation.

[0093] Furthermore, by extrapolating the spatially phase change curves from both sides (e.g., from the first sensor towards the second sensor and from the second sensor towards the first sensor), the confidence of the selected position hypothesis can be further enhanced, or the hypothesis established from one side can be verified from the other side.

[0094] Each of the two corresponding sensors can be assigned the phase of the signal received by the corresponding sensor. This phase respectively corresponds to the relevant component of the measurement vector assigned to the detected radar target. The signals assigned to the relevant sensors can respectively be assigned phase change curves on the channels or transmit antenna elements or receive antenna elements. The phase centroid of the relevant channels (relevant transmit and receive antenna elements) can be assumed as the position of the channels.

[0095] The relevant sensors can respectively receive a large number of signals, and the large number of signals are respectively assigned to radar targets assigned to different angles.

[0096] The spatially extrapolated phase progressions of signals received from one of the relevant sensors (the signals being respectively assigned to relevant radar targets having different angles) can be different. In particular, the respective extrapolated spatially phase change curves can be angle-dependent.

[0097] The phase change curve of the signal assigned to the radar target of the other of the two sensors can be generated from the respective measurement vectors, the components of the respective measurement vectors being assigned to respective spatial positions (relative to the sensor). The spatial position can correspond to the phase centroid of the relevant channel or the relevant channel product.

[0098] For the respective bistatic channels, the phase and the phase change curve can be determined from the signal according to the respective multiplicative components of the measurement vector, the signal corresponding to the respective measured value of the bistatic channel or the signal corresponding to the respective product of the measured values of the bistatic channel.

[0099] Each extrapolated spatially phase change curve of the signal assigned to the radar target of the other of the two sensors can coincide at periodic positions with the phase measured by one of the two sensors of the signal assigned to the detected radar target. Here, the period depends on the angle of the radar target.

[0100] The check for correlation can include checking the consistency and / or similarity between the signal received by one of the two respective sensors, in particular its phase, and the extrapolated spatially phase change curve of the other of the two respective sensors.

[0101] The check for correlation can in particular be carried out for the signals assigned to the respective radar targets of a plurality of channels of one of the two sensors. Thereby, a large number of signals can be checked with respect to their correlation with the extrapolated phase change curves.

[0102] The check for correlation can in particular include: checking the correlation between the spatially phase change curve of the signal assigned to the respective radar target of one of the two respective sensors and the extrapolated spatially phase change curve of the signal assigned to the radar target of the other of the two respective sensors. Thus, for the respective radar target, the correlation between the measured phase change curve of one sensor and the extrapolated phase change curve of the other sensor is checked.

[0103] The phase change curve of the signal assigned to the radar target of one of the two sensors can equally be generated from the corresponding measurement vector, the components of which are assigned to the corresponding spatial positions (relative to the sensor). The spatial position can correspond to the phase center of gravity of the relevant channel or the relevant channel product.

[0104] Based on the result of the inspection, the spatial deviation relative to each other between two corresponding sensors can be determined. In particular, the spatial deviation relative to each other between two corresponding sensors can be determined based on the result corresponding to the convolution operation. For example, the spatial deviation relative to each other between two corresponding sensors can be determined by determining the maximum value of the result of the convolution operation (or multiple convolution operations for the corresponding radar target or its angle).

[0105] The position of the virtual antenna element or channel of the bistatic, virtual sensor can be determined more accurately, for example, based on the roughly known position of the channel, by resolving the multi-valuedness of the signal path length due to possible phase jumps of k·2·π based on the monostatic angle estimation of the individual sensor and the phase of the complex-valued measurement variable of the virtual channel. Such multi-valuedness can be caused, for example, by the larger spacing of the virtual channels. The position of the virtual antenna element or channel of the bistatic, virtual sensor can also be determined, for example, by TDM (time division multiplexing)-MIMO calculation and knowledge of the exact positions of the TX and RX antennas on the physical sensor.

[0106] The method according to different aspects can include: improving the angular resolution of the assignment relationship between the control vector and the angle by interpolating the control vector. Thus, the angular intermediate step can be refined by interpolation. Thus, the angular granularity of the assignment relationship or the bistatic calibration matrix relative to the monostatic angular accuracy can be improved. This is achieved by increasing the total aperture relative to the individual sensor and correctly including the phase of the measured values of the bistatic channels.

[0107] In some embodiments, the method further includes: refining the angular granularity of the assignment relationship by interpolation, where the assignment relationship assigns a corresponding control vector to each of a plurality of angles. That is, interpolation of the control vector or the control matrix is performed. In particular, here, the angular resolution (improvement) of the angular region for the angle is increased. This can be done especially in the first or third aspect, but can also be done in the assignment of the control vector according to the second or fourth aspect.

[0108] In some embodiments, when estimating the angle of a detected radar target, the angle is estimated with an angular granularity less than or equal to 0.5°, preferably less than or equal to 0.1°. In particular, in the third association, the angular granularity of the assignment relationship between the phase vector and different angles can be less than or equal to 0.5°, preferably less than or equal to 0.1°. In particular, in each of the first to fifth aspects, the assignment of the control vector to different angles can have the mentioned angular granularity.

[0109] In some embodiments, in the third association, the assignment relationship between different angles and the phase vector has a smaller angular granularity compared to the assignment relationship between the control vector of the bistatic, virtual sensor and different angles in the case of the first association and / or compared to the assignment relationship between the control vector of a single radar sensor and different angles in the case of the second association.

[0110] These above-mentioned embodiments with improved or smaller angular granularity have particular advantages: Angle estimation is achieved with a particularly high angular resolution without having to measure the entire control matrix at this angular resolution. In particular, the control matrix does not have to be determined by calibration measurements of the system with a high angular resolution.

[0111] Furthermore, the subject matter of the present invention is a collaborative radar sensor network, in particular a collaborative radar sensor network for a vehicle or a motor vehicle, in which one of the methods described above is implemented. Thus, according to a further aspect of the present disclosure, a collaborative radar sensor network having a plurality of individual radar sensors is provided for use, the collaborative radar sensor network including control and analysis utilization means, the control and analysis utilization means being arranged to implement the methods according to each of the aspects described herein. The control and analysis utilization means can be connected to the radar sensors, for example. Description of the Drawings

[0112] Hereinafter, the preferred embodiments of the present invention will be explained in more detail based on strongly simplified schematic diagrams.

[0113] Shown here:

[0114] Figure 1 A schematic diagram showing a first embodiment of a method for estimating the angle of a radar target for a collaborative radar sensor network having a plurality of individual radar sensors;

[0115] Figure 2 A schematic diagram showing a second embodiment of the method and the collaborative radar sensor network;

[0116] Figure 3Schematic diagram showing details of a method according to a first or second embodiment relating to the determination of an assignment relationship, wherein each of a plurality of angles is assigned a corresponding control vector;

[0117] Figure 4 Schematic diagram showing details of a method according to a first or second embodiment relating to the determination of a spatial deviation between corresponding sensors from one another; and

[0118] Figure 5 Schematic diagram showing details of a second embodiment of a method relating to angle estimation.

[0119] In the following description of the embodiments, the same or similar reference numerals are used for elements shown in different drawings and acting similarly, and repeated descriptions of these elements are omitted. Detailed implementation

[0120] Figure 1 Schematic diagram of a first embodiment of a method 100 for estimating the angle α of a radar target 300 of an S30 radar using a phase - coherent cooperative radar sensor network 200 having a plurality of individual radar sensors S 1 ,S 2 。 For the sake of simplicity, as schematically shown in Figure 3 a, the radar sensor network 200 includes two individual radar sensors S Figure 3 a, the radar sensor network 200 includes two individual radar sensors S 1 ,S 2 which each have a transmitting antenna element TX 11 ,TX 21 and two receiving antenna elements RX 11 ,RX 12 ,RX 21 ,RX 22 。 Here, the first digit in the index refers to the respective number of the individual radar sensors S 1 ,S 2 , and the second digit refers to the respective number of the corresponding receiving or transmitting antenna element. In practice, a plurality of transmitting antenna elements and more than two receiving antenna elements can be provided for each radar sensor, and more than two radar sensors can be provided. The transmitting antenna elements TX 11 ,TX 21 and the receiving antenna elements RX 11 ,RX 12 ,RX 21 ,RX 22 are in the respective radar sensors S 1 ,S 2are arranged in different positions in such a direction that, in said direction, the respective radar sensors S 1 , S 2 or the cooperative radar sensor network 200 is angle-resolving. Depending on the participating radar sensors S 1 , S 2 the respective combinations of transmit antenna elements TX 11 , TX 21 and receive antenna elements RX 11 , RX 12 , RX 21 , RX 22 are assigned radar signal paths, in particular respective channels. As shown in Figure 3 b, according to the first or second aspect of the invention, the respective monostatic channels TX 11 ·RX 11 , TX 11 ·RX 12 , TX 21 ·RX 21 , TX 21 ·RX 22 correspond to monostatic radar signal paths (in which monostatic radar signal paths, the transmit antenna elements TX 11 , TX 21 and the receive antenna elements RX 11 , RX 12 , RX 21 , RX 22 belong to the same radar sensors S 1 , S 2 of the radar sensor network 200). In order to prepare the cooperative radar sensor network 200 for bistatic angle estimation, in step S10 the determination of the assignment relationship is carried out, which can assign a respective control vector A α to each of a plurality of angles α (azimuth angles) (at which the respective radar target 300 can be detected). In other words, the determination S10 of the assignment relationship is the determination of the control matrix A. The control matrix A has y rows, in particular control vectors A α , where y is a natural number, which natural number corresponds to the number of the plurality of angles α. Each row of the y rows is assigned to a different angle α. The respective control vectors A α of the control matrix A include x components A αx , where x is a natural number. The l-th component of the respective control vector A α is the bistatic component A 1-2 αl , where here, A 1-2 α is A αpartial vector that combines the bistatic components of the two stations. It corresponds to the corresponding first monostatic component TX of the control vector of the first radar sensor S 1 ·RX 11 ·RX 11 TX 11 ·RX 12 and the corresponding second monostatic component TX of the control vector of the corresponding second radar sensor S 2 ·RX 21 ·RX 21 TX 11 ·RX 22 The product of complex numbers TX 11 ·RX 11 ·TX 21 ·RX 21 TX 11 ·RX 12 ·TX 21 ·RX 22 . The determination of the affiliation relationship S10 is carried out without measuring the bistatic components of the installed radar sensor network 200, and is carried out before the actual use of the radar sensor network 200 or after the installation of the collaborative radar sensor network 200. In addition to the bistatic components of the partial vector A 1-2 α , the corresponding control vector A α also has partial vectors A 1 α , A 2 α monostatic components corresponding to individual radar sensors S 1 S 2 corresponding monostatic components TX of the control vector of 11 ·RX 11 TX 11 ·RX 12 TX 21 ·RX 21 TX 21 ·RX 22 . Assign the corresponding bistatic channel V of the virtual sensor S 1-2 αl to the corresponding l-th bistatic component A of the control vector A 1-2 ; V 11,21 or V 21,11 ; V 11,22 ; V 21,12 The product TX 11 ·RX 21 ·TX 21 ·RX 11 or TX 11 ·RX22 TX 21 RX 12 In this case, the transmitting or (after a comma) receiving radar sensor S is given in the index. 1 , S 2 The virtual sensor S is numbered and the corresponding radar sensor is given the corresponding receiving antenna element or transmitting antenna element number. 1-2 The dual station channel V 11,21 ; V 21,11 ; V 11,22 ; V 21,12 The redundancy that usually applies is:

[0121] TX 1m RX 1n TX 2m RX 2n =TX 1m RX 2n TX 2m RX 1n ,

[0122] Where m and n are natural numbers. Figure 3 In b, this relationship is shown by way of example for m=1 and n=1 or n=2, wherein m corresponds to the corresponding transmitting antenna element of the sensor and n corresponds to the corresponding receiving antenna element of the sensor. In other words, the product of the two given single-station channels of a single sensor is equal to the product of the two redundant double-station channels. This is taken into account in the analysis using the correspondingly constructed measurement vectors. Here and below, only the case of m=1 and n=1 or n=2 is also shown by way of example for the measurement vectors.

[0123] According to the third or fourth aspect, alternatively, the determination S10 of the assignment relationship may also be the determination of a control matrix A, in which the first component A of the bi-station of the control vector A is 1-2 αl Corresponding to the radar sensor S 1 or S 2 The corresponding measured first component TX of the transmission control vector of the corresponding first radar sensor in 11 or TX 21 and radar sensor S 2 or S 1 The corresponding receiving control vector RX of the second radar sensor 21 , RX 12 or RX 11 , RX 12 The corresponding measured second component produces the product TX 11 RX 21 , TX11 · RX 22 or TX 21 · RX 11 , TX 21 · RX 12 。Therefore, the components of the unidirectional (transmit or receive) control vector are combined with each other. Therefore, the correspondingly constructed measurement vector is analyzed and utilized, and the measurement vector includes bistatic components.

[0124] According to a second aspect, the method includes the step of calculating a measurement vector X of S12 assigned to the detected radar target 300. The measurement vector X has x components X x . The measurement vector X has at least bistatic components, which are jointly formed as a partial vector X 1-2 . The l-th bistatic component X 1-2 l of the measurement vector X corresponds to the product of the measurement value assigned to the radar target 300 by the first channel TX 11 · RX 21 , TX 21 · RX 11 of the bistatic and the measurement value assigned to the radar target 300 by the second channel TX 21 · RX 11 , TX 21 · RX 12 · TX 11 · RX 21 · TX 21 · RX 11 , TX 11 · RX 22 · TX 21 · RX 12 . The first channel TX 11 · RX 21 , TX 21 · RX 11 corresponds to the m-th transmit antenna element TX 1 of the corresponding first radar sensor S 11 in the radar sensor and the n-th receive antenna element RX 2 of the corresponding second radar sensor S 21 , RX 22 . The second channel TX 21 · RX 11 , TX 21 · RX 12 corresponds to the m-th transmit antenna element TX 2 of the second radar sensor S 21 and the n-th receive antenna element RX 1 of the first radar sensor S11 ,RX 12 。

[0125] The measurement vector can have a monostatic component, which is jointly the partial vector X 1 ,X 2 。The measurement vector X 1-2 The bistatic component corresponds to the given product TX of the bistatic channels 11 ·RX 21 ·TX 21 ·RX 11 ,TX 11 ·RX 22 ·TX 21 ·RX 12 。The monostatic component X of the measurement vector X 1 ,X 2 corresponds to the monostatic channels TX of the corresponding radar sensor S 1 ,S 2 ,S 11 ·RX 11 ,TX 11 ·RX 12 ,TX 21 ·RX 21 ,TX 22 ·RX 22 。As described above, the measurements of the two bistatic virtual channels TX 11 ·RX 21 ·TX 21 ·RX 11 ,TX 11 ·RX 22 ·TX 21 ·RX 12 are redundant.

[0126] Due to the corresponding relationship between the bistatic l-th component A of the control vector A assigned to the angle α as described above α and the corresponding bistatic l-th component X of the measurement vector X 1-2 α1 According to the first or second aspect, the corresponding bistatic component A of the control vector X assigned to the corresponding angle α can be used 1-2 l to process the two redundant bistatic channels TX 1-2 α ·RX 11 ·RX 21 ·TX 21 ·RX 11 ,TX 11 ·RX 22 ·TX 21 ·RX 12The product of the spectrum values in the d and v spectra measured for the radar target 300 is analyzed and utilized for angle estimation.

[0127] exist Figure 4 In step S50, schematically shown in FIG. 1 , a spatial offset d (S 1-2 -S 2 ), the two corresponding sensors comprising a single radar sensor S 1 , S 2 One and at least one dual-station virtual sensor S 1-2 The premise for this is that a single radar sensor S 1 , S 2 It can be unambiguously determined that the radar sensor S 1 and S 2 This is the case for all sensors in which the virtual antenna element V 11,21 ; V 11,22 The distance d(V 11,21 -V 11,22 ) is designed to have a maximum of λ / 2. The relevant sensors each receive a large number of signals, which are assigned to the corresponding radar targets 300 with different angles α. The two sensors are each assigned a phase of the signal received by the corresponding sensor. As in Figure 4 b-II shows, by way of example, the associated sensor S 2 The signals received by one of the sensors are respectively transmitted through the relevant sensor S 2 Channel or transmit or receive antenna element RX 21 , RX 22 Assign the extrapolated spatial phase progression (Phasenfortschritt) a1,…a t , wherein radar targets 300 are numbered from 1 to t. Different angles α are respectively associated with phase progressions a1, ... a of the associated radar targets 300. t Due to the different angular dependencies of the corresponding radar targets 300. Figure 4 As shown in bI, two sensors S are generated by the corresponding measurement vectors 1-2 The corresponding phase curves b1 . . . b2 of the signal of the other sensor in FIG. 3 , which is assigned to the corresponding radar target 300 , are shown in FIG. t .

[0128] Determining S50 includes checking S52 two corresponding sensors S 2The extrapolated spatial phase change curves a1, … a of one of them t and the two corresponding sensors S 1-2 The corresponding measured phase change curves b1 … b of the other one of the sensors t The association. Based on the sensor S 2 The extrapolated spatial phase progress a1, … a of one of them t Determine the spatial position at which the corresponding (extrapolated) phase assigned to the sensor S 2 of one of them corresponds to the corresponding (measured) phase of the other one of the sensors S 1-2 Thus, for different angles α, a large number of hypotheses for the position d(S 1-2 -S 1-2 -S 2 ) of the other one of the sensors S Position d(S 1-2 -S 2 )(At this position, among the extrapolated spatial phase change curves a1, … a t , most of the hypotheses (in the dashed rectangle in Figure 4 b-II) are consistent with the corresponding measured phase change curves b1 … b 1-2 of the other one of the two corresponding sensors S t (In the dashed rectangle in Figure 4 b-I)) corresponds to the searched position d(S 1-2 -S 2 ). For this purpose, the association check S52 includes a convolution operation *, where the corresponding phase change curves a1, … a 1-2 assigned to the relevant sensors S 2 *b1 … b t are convolved. The phase change curves b1 … b t of the signal received by the other one of the two corresponding sensors S 1-2 are inverted and scaled for calculation. In the step of determining the spatial deviation d(S t ) between the two corresponding sensors S 1-2 , S 2 according to the result of the check 52, by determining the corresponding maximum value of the result of the convolution operation a1, … a 1-2 *b1 … b 2 to determine the spatial deviation d(S t ) between the two corresponding sensors S t 1-2 , S 2 1-2 -S 1-2 -S 2)。In this way, the deviations of all real and virtual sensors relative to each other can be determined.

[0129] In step S40, according to the corresponding phase correction C α to correct the phase of the components of the corresponding control vector A α wherein the correction corresponds to the corresponding phase shift. The corresponding phase correction C includes a phase correction value in the form of e α obtained from the geometric relationship between the relevant sensors S 1 , S 2 , S 1-20 . In the case of the bistatic multiplicative components TX -j·2·π·k ·RX α ·TX 11 ·RX 11 ·TX 21 ·RX 21 , TX 11 ·RX 12 ·TX 21 ·RX 22 of the control vector A, the phase shift can be twice the phase difference generated by the path difference Δl(S 1-2 ) generated based on the geometry for the angle α. For the corresponding sensors S 1 , S 2 , S 1-2 , the components or partial vectors A 1 , A 2 , A 1-2 of the control vector A assigned to the angle α for the corresponding sensors S α , S 1 α , A 2 α , A 1-2 α are equal, that is, all components of the partial vectors A 1 , A 2 , A 1-2 of the control vector A for the corresponding sensors S α , S 1 α , A 2 α , A 1-2 α are corrected with the same phase correction C 1 α , C 2 α , C 1-2 α . The corresponding phase corrections C 1 α , C2 α , C 1-2 α depending on the angle α and on the sensor S 1 , S 2 , S 1-2 .

[0130] In the step of associating S20 the measurement vector X with the control vector A assigned to different angles α α , full use is made, by taking into account the phase correction C α of the corresponding control vector A corrected thereby α , and the phase information of the individual component X x of the measurement vector X enters into the estimation of the angle α. For this purpose, the measurement vector X is multiplied by the corresponding phase-corrected control vector A α (the control vector A α being assigned to the angle α corresponding to the angle hypothesis) or by the corrected control matrix A. The result is the angle spectrum assigned to the detected radar target, from which the angle α can be estimated. In the simplest case, the estimation S30 of the angle α is carried out by a maximum search in the corresponding angle spectrum obtained by the association S20. Thus, the phase information of the bistatic components of the measurement vector X is analyzed and utilized in order to carry out the angle estimation α by taking into account the bistatic channels TX 11 ·RX 21 , TX 21 ·RX 11 , TX 11 ·RX 22 , TX 21 ·RX 12 .

[0131] Figure 2 shows a schematic illustration of a second embodiment of a method 100 for estimating S30 the angle α of a radar target 300 for a collaborative radar sensor network 200 having a plurality of individual radar sensors S 1 , S 2 . The radar sensor network 200 includes control and analysis means 280, which are connected to the radar sensors S 1 , S 2 and are designed to carry out the method 100. The method 100 according to the second embodiment includes the above-described determination S10 of the assignment relationship, the calculation S12 of the measurement vector X assigned to the detected radar target 300, the determination S50 of the spatial deviation d(S 1-2 -S 2) and the step of estimating the angle α of S30 based on the result of the association S20. The differences from the first embodiment are described below. The association S20 includes a first measurement vector X of at least one virtual sensor S of the bistatic station assigned to the detected radar target 300 1-2 and a control vector A of the virtual sensor S of the bistatic station assigned to different angles α 1-2 of the first association S22. The virtual sensor S of the bistatic station 1-2 corresponds to the corresponding first radar sensor S according to the first and second aspects or according to the third and fourth aspects 1-2 α and the bistatic channel TX of the corresponding second radar sensor S 1-2 ·RX 1 ,TX 2 ·RX 11 ,TX 21 ·RX 21 ,TX 11 ·RX 11 ,TX 22 ·RX 21 ·RX 12 configuration. In addition, the association S20 includes at least one single radar sensor S 1 ,S 2 a second measurement vector X of the detected radar target 300 assigned to 1 ,X 2 and a second association S24 of the control vectors A of the single radar sensor S 1 ,S 2 assigned to different angles α 1 α ,A 2 α Therefore, first, the monostatic and bistatic sensors are analyzed and utilized individually in step S22 or S24.

[0132] As shown in Figure 5 , according to the fifth aspect (which can be part of the first to fourth aspects), the estimation S30 of the angle α of the detected radar target 300 includes the determination S32 of a phase vector, and the phase vector assigns corresponding phases to at least one virtual sensor S of the bistatic station 1-2 and at least one single radar sensor S 1 ,S 2 The corresponding phases The corresponding phases are obtained respectively based on at least one of the results of the first association S22 and the second association S24. Assigned to the relevant sensor S 1 ,S 2 ,S 1-2The phase of the corresponding complex value corresponds to the phase of the corresponding maximum value of the absolute value of the relevant angular spectrum As in Figure 5 The phase vector is a vector whose components correspond to the phase values ​​determined by the corresponding maximum values ​​of the angle spectrum.

[0133] The estimation S30 includes a third association S34 of the determined phase vector with phase vectors assigned to different angles α. These phase vectors indicate the phase vectors of at least one individual radar sensor S 1 , S 2 and at least one dual-station virtual sensor S 1-2 Due to their spatial deviation d(S 1-2 -S 2 ) produces a phase relationship. Figure 5 The corresponding sensor S 1 , S 2 , S 1-2 The spatial deviation between them is d(S 1 -S 1-2 ), d(S 1-2 -S 2 ), d(V 11,21 -V 11,22 ) calculates the phase vectors associated with different angles α. For this purpose, the spatial deviations d(S 1-2 -S 2 ) creates a position vector whose components are the corresponding associated sensors S 1 , S 2 , S 1-2 To calculate the position vector, use the corresponding sensor S 1 , S 2 , S 1-2 Any reference position on the sensor S, in particular the phase center of gravity of the reference channel, is used as the reference position for the corresponding sensor S 1 , S 2 , S 1-2 The reference point of the sensor S 1 , S 2 , S 1-2 The position vectors are used to generate a position estimation matrix P relative to each other's geometric relationship. The position estimation matrix P has the following form, where d is the position vector, α is the corresponding angle of the radar target 300, and λ is the radar wavelength:

[0134]

[0135] Thus, the rows of the position estimation matrix P correspond respectively to the phase vectors assigned to different angles α, and the phase vectors respectively give, in the relevant radar sensors S 1 , S 2 , S 1-2 the angle-dependent phase relationships resulting from their spatial deviations d(S1 - S 1 -2 ), d(S 1-2 - S 2 ), d(V 11,21 - V 11,22 ) from each other.

[0136] The third association S34 includes determining an angle spectrum by multiplying the phase vectors with the position estimation matrix P. Since for the third association S34 only the deviations d(S 1 , S 2 , S 1-2 relative to each other, d(S 1 - S 1-2 ), d(S 1-2 - S 2 ), d(V 11,21 - V 11,22 ) of the sensors S are considered, and these deviations can be many times larger than the wavelength λ of the received signal, the corresponding angle spectrum has multi-valuedness, as shown in Figure 5 c. Here, the multi-valuedness is characterized in that for a periodic angle α, values with the same amplitude or phase appear, and these values correspond to the detected radar target 300.

[0137] The estimation S30 further includes determining an estimated value S36 with multi-valuedness for the angle α of the radar target 300 according to the result of the third association S34, as shown in Figure 5 d.

[0138] The angle α of the radar target 300 is determined by the step of resolving the multi-valuedness of the estimated value for the angle α according to the corresponding angle spectrum of the first association S22 or the second association S24, S38. Here, as shown in Figure 5 b, the surroundings α - U around the maximum value of the angle spectrum are selected such that the estimated values of the multi-valued estimated values are as in Figure 5It falls into the surrounding environment α-U shown in c. Here, the refined angular resolution of the angle α (at which the radar target 300 is detected) is determined by multiplying the result of the first correlation S22 and / or the second correlation S24 by the result of the third correlation S34. Therefore, the phase correction S40 is not necessary. Thus, the angular resolution is improved compared to the angular grid of the first or second correlation S22, S24 (the angular grid corresponding to the angular grid of the calibration step of a single sensor). Here, as a result of the first or second correlation, only the product of the measurement vector and the relevant control vector needs to be measured in order to estimate the angle α with a fine angular resolution.

[0139] However, in all embodiments, additionally, the angular intermediate step can be taken into account for the correlation by interpolating the control matrix in step S10, or, in step S34, an angular resolution finer than the angular resolution in the correlation steps S22 or S24 can be used for the assignment of the phase vector to the angle, so that the angular resolution can be further improved, for example, increased to 0.5° or 0.1°.

Claims

1. A method (100) for a cooperative radar sensor network (200) having a plurality of individual radar sensors (S 1 , S 2 ), which is used to prepare the cooperative radar sensor network (200) for bistatic angle estimation, wherein, The method (100) comprises: Determine (S10) the allocation relationship (A), where the allocation relationship is that each of a plurality of angles (α) is allocated a corresponding control vector (A α ), wherein the corresponding control vector (A α ) has at least a bistatic component (A 1-2 α ). wherein, the l-th component (A α ) of the bistatic of the control vector (A 1-2 αl ) corresponds to the product (TX 1m ·RX 1n ·TX 2m ·RX 2n ) generated by the following: - a respective first radar sensor (S 1 ) of the control vector corresponding to the first component of the single station (TX 1m RX 1n ), wherein the first component of the single station corresponds to the first radar sensor (S 1 )’s mth transmit antenna element (TX 1m ) to the first radar sensor (S 1 )’s nth receiving antenna element (RX 1n ) of a single station; and - a corresponding second radar sensor (S 2 ) of the control vector corresponding to the second component of the single station (TX 2m RX 2n ), wherein the second component of the single station corresponds to the second radar sensor (S 2 )’s mth transmit antenna element (TX 2m ) to the second radar sensor (S 2 )’s nth receiving antenna element (RX 2n ) of a single station.

2. Method (100) for a cooperative radar sensor network (200) having a plurality of individual radar sensors (S 1 , S 2 ) for estimating the angle (α) of a radar target (300), wherein The method (100) comprises: Calculate (S12) a measurement vector (X) assigned to the detected radar target (300), wherein the measurement vector (X) has at least a bistatic component (X 1-2 ), wherein the l-th bistatic component (X 1-2 l ) of the measurement vector (X) corresponds to the product (TX 1m ·RX 2n ·TX 2m ·RX 1n ) resulting from the following: - Measurements of the bistatic channels (TX 1m ·RX 2n ) assigned to the radar target (300), the bistatic channels corresponding to the m-th transmit antenna element (TX 1 ) of the corresponding first radar sensor (S 1m ) in the radar sensor and the n-th receive antenna element (RX 2 ) of the corresponding second radar sensor (S 2n ) in the radar sensor; and - Measurements assigned to the radar target (300) for the bistatic channels (TX 2m · RX 1n ), the bistatic channels corresponding to the m-th transmit antenna element (TX 2 ) of the second radar sensor (S 2m ) and the n-th receive antenna element (RX 1 ) of the first radar sensor (S 1n ), and Estimate the angle (α) of the detected radar target (300), where the estimated angle (α) is determined based on the result of the association (S20) between the measurement vector (X) and the control vector (A α ) assigned to different angles (α), or at least determined based on the product of the measurement vector (X; X 1 ; X 2 ; X 1-2 ) and the control vector (A α ; A 1 α ; A 2 α ; A 1-2 α ).

3. The method according to claim 2, wherein, The method (100) further comprises: Preparing (S10) the collaborative radar sensor network (200) for bistatic angle estimation according to the method as claimed in claim 1.

4. The method (100) according to claim 3, Among them, The method (100) comprises: - At least one bistatic, virtual sensor (S 1-2 ) of a first measurement vector (X 1-2 ) assigned to the detected radar target (300) and a first correlation (S22) of the control vector (A 1-2 ) assigned to different angles (α) of the bistatic, virtual sensor (S 1 -2 α ), wherein the bistatic, virtual sensor (S 1-2 ) corresponds to the configuration of the bistatic channel of the respective first radar sensor (S 1 ) and the respective second radar sensor (S 2 ), wherein the measurement vector (X 1-2 ) of the bistatic, virtual sensor (S 1-2 ) has the bistatic components (X 1-2 x ); and / or - at least one single radar sensor (S 1 , S 2 ) of a second measurement vector (X 1 , X 2 ) assigned to the detected radar target (300) and the control vector (A 1 , S 2 ) of the single radar sensor (S 1 α , A 2 α ) assigned to different angles (α) of the second association (S24), wherein, the estimation (S30) of the angle (α) of the detected radar target (300) comprises: Determine (S32) a phase vector that assigns a respective phase to the at least one bistatic, virtual sensor (S 1-2 ) and the at least one single radar sensor (S 1 , S 2 ), wherein the respective phases are obtained based at least on the result of at least one of the first association (S22) and the second association (S24); ​ The third correlation (S34) of the determined phase vector with the phase vectors assigned to different angles (α), the phase vectors assigned to different angles (α) giving the phase relationship between the at least one single radar sensor (S 1 , S 2 ) and the at least one bistatic, virtual sensor (S 1-2 ) due to their spatial deviation (d(S 1-2 -S 2 )) from each other; Determining (S36) a multi-valued estimated value for the angle (α) of the radar target (300) according to the result of the third association (S34); and Resolving the multi-valued nature of the estimated value for the angle (α) of the radar target (300) according to the result of the first association (S22) or the second association (S24).

5. The method (100) according to any one of claims 1 to 3, wherein The corresponding control vector (A α ) belonging to the angle (α) is composed of the single - station components (A 1 α , A 2 α ) and the bistatic components (A 1-2 α ). Among them, the corresponding single - station components (A 1 α , A 2 α ) are the components of the control vectors of the relevant radar sensors in the radar sensors (S 1 , S 2 ).

6. The method (100) according to claim 5, wherein, The method (100) further comprises: Correct (S40) the phase of the components of the respective control vector (A α ) according to the respective phase correction (C α ). The phase correction corresponds to a respective phase offset which, for the respective angle (α), is caused by a spatial deviation d(S 1 , S 2 ) between the respective sensors in at least one single radar sensor (S 1-2 ) and at least one bistatic, virtual sensor (S 1-2 -S 2 ), where the bistatic, virtual sensor (S 1-2 ) corresponds to a configuration of a bistatic channel of a respective first radar sensor (S 1 ) and a respective second radar sensor (S 2 ).

7. The method (100) according to any one of claims 4 or 6, wherein The method (100) further comprises: Based on the signals of the detected radar targets (300) from multiple angles (α) respectively assigned to the two sensors, determining (S50) the spatial deviation (d(S 1 , S 2 )) between two corresponding sensors (S 1-2 ) in the sensors including the multiple individual radar sensors (S 2 , S 1-2 ) and the at least one bistatic, virtual sensor (S 1-2 -S 2 ), wherein the determination includes: - Check (S52) the phase of the signal assigned to the respective radar target (300) of one of the two respective sensors (S 1-2 ) and the correlation of the extrapolated spatially varying phase curve (a1,... a 2 ) of the signal assigned to the radar target (300) of the other of the two respective sensors (S t ); and - Determine (S54) the spatial deviation (d(S 2 , S 1-2 )) between each other between the two corresponding sensors (S 1-2 -S 2 ) based on the result of the inspection (S52).

8. Method (100) for a cooperative radar sensor network (200) with a plurality of individual radar sensors (S 1 , S 2 ), which is used to prepare the cooperative radar sensor network (200) for bistatic angle estimation, wherein, The method (100) comprises: Determine (S10) the assignment relationship (A), where the assignment relationship assigns a corresponding control vector (A α ) to each of a plurality of angles (α), and wherein the corresponding control vector (A α ) has at least a bistatic component (A 1-2 α ). wherein, the l-th component (A α ) of the bistatic of the control vector (A 1-2 αl ) corresponds to the product resulting from the following: - The corresponding first component (TX1m) of the transmission control vector of the corresponding first radar sensor (S 1 , S 2 ) in the radar sensor, where the first component (TX 1 ) corresponds to the m-th transmit antenna element of the first radar sensor (S 1m ); and 1 ) of the first radar sensor; and - the corresponding second component (RX 1 , S 2 ) of the reception control vector of the corresponding second radar sensor (S 2 ) in the radar sensor (S 2n ), wherein the second component corresponds to the nth reception antenna element of the second radar sensor (S 2 ).

9. Method (100) for a collaborative radar sensor network (200) with a plurality of individual radar sensors (S 1 , S 2 ), for estimating the angle (α) of a radar target (300), wherein, The method (100) comprises: Preparing the collaborative radar sensor network (200) for bistatic angle estimation according to the method as claimed in claim 8, Determining a measurement vector (X) associated with a detected radar target (300), wherein the measurement vector (X) has at least a bistatic component (X 1-2 ), wherein the first component (X 1-2 l ) corresponds to the dual-station channel (TX 1m RX 2n TX 2m RX 1n ) is assigned to the radar target (200), the channels of the dual station corresponding to the respective first radar sensors (S 1 ; S 2 )’s mth transmit antenna element (TX 1m ; RX 2m ) and the radar sensor (S 1 , S 2 ) in the corresponding second radar sensor (S 2 ; S 1 )’s nth receiving antenna element (RX 2n ; RX 1n ),and Estimate the angle (α) of the detected radar target (300), wherein the estimated angle (α) is determined based on the result of the association (S20) of the measurement vector (X) with the steering vectors (A α ) assigned to different angles (α), or at least based on the product (A 1 ; X 2 ; X 1-2 ) of the measurement vectors (X; X α ; A 1 α ; A 2 α ; A 1-2 α ) and the steering vectors (A y ·X).

10. The method (100) according to claim 9, Among them, The method (100) comprises: - At least one bistatic, virtual sensor (S 1-2 ) of a first measurement vector (X 1-2 ) assigned to the detected radar target (300) and a first association (S22) of the control vector (A 1-2 ) assigned to different angles (α) of the bistatic, virtual sensor (S 1 -2 α ), wherein the bistatic, virtual sensor (S 1-2 ) corresponds to the configuration of the bistatic channel of the respective first radar sensor (S 1 ) and the respective second radar sensor (S 2 ), wherein the measurement vector (X 1-2 ) of the bistatic, virtual sensor (S 1-2 ) has the bistatic components (X 1-2 x ); and / or - at least one individual radar sensor (S 1 , S 2 ) of a second measurement vector (X 1 , X 2 ) assigned to the detected radar target (300) and a control vector (A 1 , S 2 ) of the individual radar sensor (S 1 α , A 2 α ) assigned to different angles (α) of a second correlation (S24), wherein, the estimation (S30) of the angle (α) of the detected radar target (300) comprises: Determine (S32) a phase vector, the phase vector being for the at least one bistatic, virtual sensor (S 1-2 ) and the at least one single radar sensor (S 1 , S 2 ) to which a respective phase (α) is assigned, the phases being obtained respectively at least based on the result of at least one of the first association (S22) and the second association (S24); The third association (S34) of the determined phase vectors with the phase vectors assigned to different angles (α), the phase vectors assigned to different angles (α) giving the phase relationship resulting from their spatial deviation (d(S 1 , S 2 )) between the at least one single radar sensor (S 1-2 ) and the at least one bistatic, virtual sensor (S 1-2 -S 2 )); Determining (S36) a multi-valued estimated value for the angle (α) of the radar target (300) according to the result of the third association (S34); and Resolving the multi-valued nature (S38) of the estimated value for the angle (α) of the radar target (300) according to the result of the first association (S22) or the second association (S24).

11. The method (100) according to any one of claims 8 or 9, wherein The corresponding control vector (A α ) associated with the angle (α) consists of the components of the single-static (A 1 α , A 2 α ) and the components of the bistatic (A 1-2 α ), where the corresponding single-static components are the components of the control vector of the relevant radar sensor among the radar sensors (S 1 , S 2 ).

12. The method (100) according to claim 11, wherein, The method (100) further comprises: Correct (S40) the phase of the components of the respective control vector (A α ) according to the respective phase correction (C α ). The phase correction corresponds to a respective phase offset which, for the respective angle (α), is caused by a spatial deviation d(S 1 , S 2 ) between the respective sensors in at least one single radar sensor (S 1-2 ) and at least one bistatic, virtual sensor (S 1-2 -S 2 ), where the bistatic, virtual sensor (S 1 -2 ) corresponds to a configuration of a bistatic channel of a respective first radar sensor (S 1 ) and a respective second radar sensor (S 2 ).

13. The method (100) according to any one of claims 10 or 12, wherein, The method (100) further comprises: Based on the signals of the detected radar targets (300) from the two sensors, which are respectively assigned to different angles (α), determine (S50) in the sensors including the multiple individual radar sensors (S 1 , S 2 ) and the at least one bistatic, virtual sensor (S 1-2 ) the spatial deviation (d(S 2 , S 1-2 )) between two corresponding sensors (S 1-2 -S 2 )) in space, wherein the determination includes: - Check (S52) the phase of the signal assigned to the respective radar target (300) of one of the two respective sensors (S 1-2 ) and the correlation with the extrapolated spatially varying phase curve (a1,... a ) of the signal assigned to the radar target (300) of the other of the two respective sensors (S 2 ); and t ) - Determine (S54) a spatial deviation (d(S 2 , S 1-2 )) between the two respective sensors (S 1-2 - S 2 ) based on the result of the inspection (S52).

14. The method according to any one of the preceding claims, Among them, The method further includes: refining the angular granularity of the assignment relationship (A) by interpolation, where the assignment relationship assigns a corresponding control vector (A α ) to each of a plurality of angles (α). or, wherein, when estimating the angle (α) of the detected radar target (300), the angle (α) is estimated with an angular granularity less than or equal to 0.5°, preferably less than or equal to 0.1°. Alternatively, in the third association (S34), the assignment relationship between the different angles and the phase vectors has a smaller angular granularity than the assignment relationship between the control vectors (A 1-2 ) of the bistatic, virtual sensor (S 1-2 α ) and the different angles (α) in the first association and / or than the assignment relationship between the control vectors (A 1 , S 2 ) of the single radar sensors (S 1 α , A 2 α ) and the different angles (α) in the second association.

15. A cooperative radar sensor network (200) having a plurality of individual radar sensors (S 1 , S 2 ), which includes a control and analysis utilization device (280) configured to implement the method (100) according to any one of the preceding claims.

Images