Antenna array for a radar sensor

Abstract

Method for avoiding interference between radar signals of several radar sensors (12 - 22) comprising the following steps: - detecting potentially interfering signals, - entering the interference signals into a multidimensional resource allocation map (26) which has at least the dimensions of frequency and time and indicates the allocation of resources by the interference signals, and - applying an avoidance algorithm (32) to minimize interference by changing the resource allocation by the radar sensors (12 - 22), characterized in that the method is applied to an integrated radar system with several radar sensors (12 - 22) that are coordinated in their function and that a coordinated avoidance strategy (38) for the entire integrated system is developed by means of the avoidance algorithm (32) taking into account the coordination of the radar sensors.

Description

[0001] The invention relates to a method for avoiding interference between radar signals of several radar sensors, comprising the following steps: - detecting potentially interfering signals, - entering the interference signals into a multidimensional resource allocation map, which has at least the dimensions of frequency and time and which indicates the allocation of resources by the interference signals, and - applying an avoidance algorithm to minimize interference by changing the resource allocation of the radar sensors.

[0002] Furthermore, the invention relates to a radar system with multiple radar sensors in a motor vehicle. State of the art

[0003] Radar sensors are increasingly used in motor vehicles to detect the traffic environment. The data collected by these radar sensors forms the basis for various assistance functions that support the driver in controlling the vehicle or ultimately replace the driver entirely.

[0004] With the increasing prevalence of such radar sensors, the probability of interference occurring between radar signals transmitted by different sensors during reception of radar echoes also rises. Such disruptive interference can significantly hinder or even prevent the correct evaluation of the radar signal. Therefore, various strategies have been developed to avoid such interference.

[0005] German patent application DE 10 2019 219 640 A1 describes a method called "spectrum sensing" that allows interference to be detected and characterized by analyzing the spectrum of the signal received by a radar sensor. If interference is detected in a measurement cycle of the radar sensor, this interference can be avoided in subsequent measurement cycles by shifting the frequency and / or the transmission time of the signal sent by the radar sensor to a range that is not yet occupied by signals from interfering sources. The parameters, such as the frequency or the transmission time, which can be changed within certain limits to avoid interference, are collectively referred to here as "resources".

[0006] From DE 10 2021 202 142 A1 a method is known in which several (ideally all) vehicles are connected to each other via a communication network, through which information about the frequencies and modulation schemes used by the radar sensors can be exchanged in order to avoid interference in advance.

[0007] German patent application DE 10 2019 215 358 A1 describes a method in which resource allocation is varied depending on the direction of travel of the vehicle. This makes it possible, for example, to ensure that the radar sensors of vehicles meeting each other always operate in different frequency bands.

[0008] In known interference avoidance methods, the avoidance strategy is based on a single radar sensor. However, newer driver assistance systems increasingly employ radar systems that include multiple radar sensors whose functions are coordinated. This coordination can, for example, involve synchronizing the measurement cycles and / or frequency modulation sequences of the radar sensors to enable cooperation between them. For instance, the cooperation of two sensors located at a relatively large distance from each other can... When multiple radar sensors are installed in the same vehicle, a large synthetic aperture for angle estimation of radar targets is achieved. However, the coordination can also consist of managing the resource usage of the different radar sensors in such a way as to avoid interference between the radar sensors of the same vehicle. Disclosure of the invention

[0009] The object of the invention is to further improve interference avoidance in radar systems.

[0010] This problem is solved according to the invention by applying the aforementioned method to an integrated radar system with several radar sensors whose function is coordinated with each other, and by developing a coordinated avoidance strategy for the entire integrated system using the avoidance algorithm, taking into account the coordination of the radar sensors.

[0011] In the method according to the invention, the avoidance strategy is therefore not directed at a single radar sensor, but at the entire system. This makes it possible to maintain the functional coordination between the various radar sensors of the system even when it is necessary to switch to other resources due to detected interference. This allows, for example, a situation to be managed in which two radar sensors belonging to the system operate in different frequency bands to avoid interference between the signals of these two sensors, and in which interference occurs in the received signal of the sensor operating at the higher frequency, necessitating a switch to a different frequency. An avoidance strategy that focuses only on this one sensor could then consist of lowering the center frequency of the frequency band.However, this would mean that the frequency separation between the two sensors could no longer be maintained. Therefore, the inventive method develops a coordinated workaround strategy, which, for example, consists of also lowering the frequency of the second sensor (insofar as this is possible without causing interference with another source of disturbance), so that the frequency separation is maintained or a compromise solution is achieved in which the overall interference remains as small as possible.

[0012] Another advantage of the method is that the received signals from the system's multiple sensors can be used to detect interference. This allows for a more precise characterization of the frequency response of the interfering signals, enabling more targeted evasive maneuvers.

[0013] Advantageous embodiments and further developments of the invention are set out in the dependent claims.

[0014] The radar system to which the method is applied can also include cooperative radar sensors, such as those described, for example, in DE 10 2021 207 215 A1. In such a cooperative radar sensor system, the sensors transmit and receive signals simultaneously at a frequency specified by a common local oscillator. Since a coordinated avoidance strategy is already employed in the method according to the invention, the cooperation can also be maintained when the avoidance strategy is applied. The avoidance strategy then aims not only to prevent interference between one of the sensors and a source of interference, but also considers possible interference between other sensors of the cooperative network and other sources of interference.

[0015] As with conventional methods of this type, the avoidance algorithm can generate a resource allocation map that specifies the frequencies and reception times for all detected interference signals. Since the radar sensors belonging to the system have different installation locations and orientations, a given interference signal will typically only affect a portion of the available sensors. In the simplest case, an interference signal is entered into the resource allocation map if it is receivable by at least one of the sensors. Optionally, for each interference signal, the resource allocation map can also indicate which sensors of the system are affected by this signal. This provides additional flexibility for developing avoidance strategies.

[0016] The resource allocation map can also include the signals currently being transmitted by the system, specifically the frequencies and transmission times of all signals sent by any sensor belonging to the system during the current measurement cycle. Naturally, the measurement cycles of the various sensors should be synchronized. When developing a backup strategy, the resource allocation map is then searched for available slots (frequencies and times) that could be used as alternatives.

[0017] In addition to frequency and time, the resource occupancy map can also include other dimensions, such as azimuth and elevation angles, which indicate the direction from which the signals are received. Other conceivable dimensions could relate to the polarization or encoding of the received signals.

[0018] In an advantageous embodiment, the fallback algorithm is configured to extrapolate the resource allocation map into the future, so that the resource allocation is predicted, for example, for the next two or three measurement cycles, and a fallback strategy can then be developed on this basis.

[0019] The sources of interference will generally be radar sensors from other vehicles. Since different types of radar sensors are used in modern vehicles, such as FMCW radars (with subtypes like Chirp Sequence and Slow Chirp) and digitally modulated radar sensors like OFDM or PMCW, it is useful to classify the detected sources of interference by type so that future behavior can be better predicted.

[0020] The method used to evaluate interference sources also depends on the types of radar sensors employed in the system. For example, if the integrated system uses chirp sequence radars, the following parameters may be relevant for the evaluation logic that analyzes the interference signals: ramp number, start and end times of an interval of consecutive samples perturbed by interference within a frequency ramp, ramp slope and center frequency of the transmitted signal, interference duration, and interference characteristics. The ramp slope of the center frequency can then be used to estimate the start and end times of a perturbed sampling interval. The ramp steepness of an FMCW jammer can be determined from the interference duration and characteristics.

[0021] If the system is formed by digitally modulated radar sensors, the evaluation logic can, for example, perform a short-time Fourier transform (STFT) of the baseband signal to estimate the frequency and power of an interfering signal as a function of time.

[0022] To improve and increase the accuracy of interfering signal detection, additional measures can be optionally implemented. For example, the sensors of the integrated system can have additional receiving channels for detecting and spectrally characterizing interfering signals outside the frequency range of their own sensor. Furthermore, in addition to the radar sensors used for object location, the system can also include additional RF sensors for detecting interfering signals, thus supplementing the detection and characterization of the interfering signals received by the radar sensors with data provided by the RF sensors. Additionally, valuable information for interference avoidance can be obtained through vehicle-to-vehicle (V2V) communication.

[0023] The choice of avoidance strategies naturally depends on the type of radar sensors belonging to the system. For example, with an FMCW radar, in addition to the transmission frequency and timing, the ramp slope can also be varied, provided this is compatible with the sensor's required range resolution. Reversing the sign of the ramp slope can also contribute to interference avoidance under certain circumstances. In the case of digitally modulated radar sensors, the avoidance strategy may also include changing the encoding scheme.

[0024] The fallback algorithm and, if necessary, the evaluation logic can be optimized through machine learning and, if necessary, further improved through self-learning while the system is running.

[0025] In one embodiment, the avoidance algorithm can include a neural network that predicts the resource occupancy map and / or develops the actual avoidance strategy. When predicting the resource occupancy map, the network can not only extrapolate the periodicities of the interfering signals into the future, but also learn to anticipate interference avoidance measures that similarly structured radar systems in other vehicles use to react to the current interference situation.

[0026] The following section explains an exemplary embodiment in more detail with reference to the drawing. The drawing shows: Fig. 1 a motor vehicle with a radar system to which the method according to the invention is applicable; Fig. 2 a block diagram of the radar system in which the method according to the invention is implemented; Fig. 3 an example of a traffic situation to illustrate the method according to the invention; and Fig. 4. A simplified example of a resource allocation map.

[0027] In Fig. Figure 1 shows a top view of a motor vehicle 10 equipped with an integrated radar system. The radar system comprises a total of six radar sensors 12 to 22, which are symbolically represented here as stylized radar beams. These radar sensors are, for example, FMCW radars that transmit a sequence of ramped frequency-modulated radar signals within each measurement cycle. The sequences of transmitted signals are interrupted by pauses during which the received radar echoes are electronically evaluated. The measurement cycles of all six radar sensors 12 to 22 are synchronized. The two sensors 12 and 14, which face forward in the direction of travel, are cooperative sensors whose frequency-modulated transmission signals are synchronized, so that, for example, an echo of a signal transmitted by sensor 12 can be received and evaluated by sensor 14.Since the two sensors are a large horizontal distance apart, a large synthetic aperture is achieved, enabling an accurate azimuth angle estimation. The same applies to the two sensors 18 and 20 mounted at the rear of the vehicle. The two laterally arranged sensors 16 and 22 operate in frequency bands that differ from those of sensors 12, 14, 18, and 20, so that, for example, the echo of a signal sent by sensor 12 does not cause disruptive interference at sensor 22.

[0028] In Fig. Figure 2 shows the radar system with sensors 12 to 22 and an associated evaluation and control unit as a block diagram. Only those components of the evaluation and control unit that serve to avoid interference are shown here. Other components, which serve to evaluate the signals received from the radar sensors for object detection, are not shown here.

[0029] In an evaluation logic 24, the signals (baseband signals) supplied by the radar sensors 12 to 22 are evaluated using known methods to detect interference caused by external sources of disturbance. The interference signals detected in the current measurement cycle are entered into a resource allocation map 26, for example, a 4-dimensional map, which indicates the signal strength as a function of frequency, time, azimuth angle, and elevation angle. Additionally, the resource allocation map 26 can specify for each interference signal which of the six radar sensors 12 to 22 are affected by the signal. If an interference source causes interference in two or more sensors, the characterization of the interference source can be improved by fusing the data supplied by these sensors. For example, the direction of incidence of the interference signal can be determined more precisely in this way.If the frequency of the interference signal is ramp-modulated, and, for example, the upper part of the frequency ramp overlaps with the frequency band of sensor 12 while the lower part of the frequency ramp overlaps with the frequency band of sensor 22, then the lower and upper limits of the frequency bands, as well as the ramp duration of the interference signal, can be determined by fusing the data from the two sensors. Further data that characterizes the nature of the interference signals in more detail can also be entered in the resource allocation map 26.

[0030] In the example shown, the radar system additionally features an RF sensor 28 which is able to receive external signals even outside the frequency bands used by the system sensors 12 to 22 and thus complete the signal image in the resource occupancy map 26.

[0031] Furthermore, the radar system in this example features a communication interface 30 for vehicle-to-vehicle communication. Via this communication network, vehicles in a shared environment can exchange information about the parameters and interference avoidance strategies of their radar systems, thereby further completing the situational awareness and improving predictability.

[0032] The resource occupancy map 26 created for the current measurement cycle is then passed to a fallback algorithm 32, which serves to develop a fallback strategy to avoid interference. The core of this fallback algorithm is a neural network 34, which is trained to first generate a prediction 36 for the temporal development of the resource occupancy over the next measurement cycles based on the current resource occupancy map 26, and then, based on this prediction 36, to develop a predictive and coordinated fallback strategy 38 that then instructs the radar sensors 12 to 22 to change their operating parameters in such a way as to eliminate current interference and prevent future interference.

[0033] Via interface 30, the avoidance algorithm 32 can also receive information about the actions planned in the radar systems of other vehicles, so that these actions can be incorporated into prediction 36. Likewise, data from the vehicle's own sensors 40, which characterize the dynamics of the vehicle itself (e.g., the steering angle), are incorporated into prediction 36. This allows, for example, changes in direction of the vehicle and corresponding changes in angles to be taken into account in prediction 36 within the resource allocation map.

[0034] The fallback strategy 38 is coordinated in that changes to the operating parameters in sensors 12 to 22 are always made in such a way that the necessary coordination between these sensors is maintained. In particular, this ensures that sensors 12 and 14, as well as sensors 18 and 20, can continue to cooperate with each other and that the necessary frequency separation between sensors 16 and 22 on the one hand and the other sensors on the other, is maintained. The action options of the neural network 34 are thus limited to actions that do not violate any coordination rules.

[0035] In another embodiment, the avoidance algorithm 32 can be programmed to display the resource allocation by the sensors 12 to 22 of its own system in the projection 36 of the 4-dimensional resource allocation map and then coordinate these resource allocations in the map so that the overlaps with interference signals are minimized.

[0036] To illustrate the working methods of the in Fig. The radar system shown in section 2 is in Fig. Figure 3 shows a traffic situation in which vehicle 10 encounters another vehicle 42 equipped with a radar system consisting of four radar sensors 44-50. It is assumed that radar sensor 44 and sensors 12 and 14 operate in the same frequency band, while sensors 16 and 22 operate in a lower frequency band. Sensor 44 causes strong interference at sensor 12, while sensor 14 is already outside the radar beam of sensor 44, so that practically no interference occurs there. To avoid interference at sensor 12, the operating range of sensors 12 and 14 could be shifted to a higher frequency range, provided it is not occupied by other sources of interference. The alternative would be to shift the frequency of sensors 12 and 14 to lower frequencies.However, reflections could then lead to interference between sensors 12 and 22, as well as between sensors 14 and 16, because the frequency difference has decreased. In this case, the operating range of sensors 16 and 22 would therefore also have to be shifted to lower frequencies.

[0037] In the Fig. In the situation shown in Figure 3, interference can also occur between sensor 46 on the one hand and sensors 12 and 14 on the other. Even if this interfering signal is initially relatively weak, the frequency of the signal from sensor 46 can be determined by sensors 12 and 14. If it turns out that this frequency is very close to the frequency of sensor 22, the frequency of sensor 22 can be shifted preventively to avoid strong interference when the other vehicle 42 passes the other vehicle 10. When the two vehicles are at the same height, sensor 46 will no longer cause interference at sensor 12, thus freeing up resources for new evasive strategies.

[0038] In Fig. Figure 4 shows examples of possible contents of the resource occupancy map 26 and the prediction 36. For simplicity, only the dimensions of time t and frequency f are shown. Time t0 denotes the present, i.e., the boundary between the current resource occupancy map 26 and the prediction 36. In the current map, periodically recurring signals 52 and 54 from two interference sources are visible. Signal 52 consists of a rising frequency ramp followed by a falling frequency ramp, while signal 54 is a chirp sequence in a lower frequency band. Based on the periodicity, it can be predicted when these signals will appear in the prediction 36. A sensor belonging to the system, for example, sensor 12 (and thus also sensor 14), is expected to send a chirp sequence 56 with a linearly increasing center frequency in the next measurement cycle.The duration of the sequence and the frequency deviation of the individual chirps are determined by requirements of the measurement function; however, it is permissible to shift the entire sequence 56 in the time dimension t and / or in the frequency dimension f within certain limits. If necessary, the rate of rise of the center frequency may also be varied within certain limits. In the example shown, a workaround strategy was found in which the chirp sequence lies in a time gap between the successive repetitions of signals 52 and 54 and was positioned in the frequency dimension such that it does not overlap with either signal 52 or 54, even though the duration of the chirp sequence 56 is greater than the time gap between the repetitions of signals 52 and 54. The signals to be transmitted by sensors 16 and 22 are shown in [reference missing]. Fig. 4 are not shown, but are positioned so that the necessary frequency difference to the signals of sensors 12 and 14 is maintained. QUOTES INCLUDED IN THE DESCRIPTION

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

[0000] DE 10 2019 219 640 A1

[0005] DE 10 2021 202 142 A1

[0006] DE 10 2019 215 358 A1

[0007] DE 10 2021 207 215 A1

[0014]

Claims

[1] A method for avoiding interference between radar signals from multiple radar sensors (12 - 22), comprising the following steps: - detecting potentially interference-causing noise signals (52, 54), - entering the interference signals into a multidimensional resource occupancy map (26) which has at least the dimensions frequency (f) and time (t) and indicates the occupancy of resources by the interference signals (52, 54), and - applying an avoidance algorithm (32) to minimize interference by changing the resource occupancy by the radar sensors (12 - 22), characterized by that the method is applied to an integrated radar system with several radar sensors (12 - 22) whose functions are coordinated with one another, and that a coordinated avoidance strategy (38) for the entire integrated system is developed by means of the avoidance algorithm (32), taking into account the coordination of the radar sensors. [2] Method according to claim 1, wherein the radar system comprises at least two cooperating radar sensors (12, 14; 18, 20) between whose transmission frequencies there is a predetermined relationship. [3] Method according to claim 1 or 2, wherein the measuring cycles of the radar sensors (12 - 22) belonging to the system are synchronized with each other. [4] Method according to claim 3, wherein a prediction (36) of the resource occupancy for future measurement cycles is generated on the basis of the resource occupancy map (26) for the current measurement cycle and the development of the alternative strategy (38) is carried out on the basis of the prediction (36). [5] Method according to claim 4, wherein the avoidance algorithm (32) is optimized by machine learning. [6] A method according to claim 5, wherein the avoidance algorithm includes a neural network (34). [7] Method according to claim 6, wherein the neural network (34) also generates the prediction (36). [8] Method according to one of the preceding claims, in which measurement data from an RF sensor (28) are also included in the resource occupancy map (26). [9] Method according to one of the preceding claims, in which information obtained via an interface (30) for communication from vehicle to vehicle also flows into the resource occupancy map (26) and / or the prediction (36). [10] Method according to claim 9, wherein the radar system is installed in a vehicle (10) and data from an intrinsic sensor system (40) which measures the dynamic state of the vehicle (10) are also included in the prediction (36). [11] Motor vehicle with a radar system with several radar sensors (12 - 22), characterized by that a method according to one of claims 1 to 10 is implemented in the radar system. [12] A software product comprising program code executable by a processor unit which, when loaded into a processor unit of a radar system, causes the radar system to carry out a method according to any one of claims 1 to 10.

Images