CN118033562A

CN118033562A - Single-point radar cross section method for radar simulation

Info

Publication number: CN118033562A
Application number: CN202310095625.XA
Authority: CN
Inventors: A·约菲; M·塞弗; M·斯特福
Original assignee: Aptiv Technologies Ltd
Current assignee: Aptiv Technologies Ltd
Priority date: 2022-02-11
Filing date: 2023-02-07
Publication date: 2024-05-14

Abstract

Single point radar cross section for radar simulation. Techniques and systems for implementing a more accurate single point Radar Cross Section (RCS) method for radar simulation without introducing complexity are described herein. For automotive radar applications, the variation, angle and range of RCS radial modes are considered, and simulation accuracy is improved by combining multipath phenomena due to ground reflection due to analysis of calculated multipath effects and near-field RCS effects. The described techniques are performed without the use of full-scale ray tracing or other computationally demanding techniques.

Description

Single-point radar cross section method for radar simulation

Technical Field

The invention relates to a single-point radar cross-section method for radar simulation.

Background

Radar is a useful device for detecting and tracking objects. Thus, radar provides many advantages for autonomous driving applications or driver assistance applications. During engineering and development, radar simulations may be run to evaluate the performance of a radar under various conditions. To reduce complexity, these simulations may represent the target response as a single point radar cross section (RCS: radar Cross Section). The single point RCS may be an excessive simplification of the actual radar response from the target, which reduces the accuracy and usefulness of the radar simulation.

Disclosure of Invention

Described herein are techniques and systems for implementing a more accurate single-point RCS method for radar simulation without introducing complexity. For automotive radar applications, the variation in angle and distance of the RCS radial pattern is considered and simulation accuracy is improved by combining multipath phenomena due to ground reflections (e.g., without using full-scale ray tracing or other computationally demanding techniques) due to analytical calculated multipath effects.

In one example, a method includes implementing a more accurate single-point RCS method for radar simulation without introducing complexity. For example, an actual antenna response representing RCS fluctuations caused by target distance may be simulated, including for close range. The distance-dependent road impact on the target RCS is then considered in a computationally efficient manner using existing simulation techniques. For fast environment simulation, information obtained from both road multipath simulation and near field simulation is combined to create a more accurate angle and range related RCS map associated with a single point representation of the target. Accuracy may be further improved by optionally representing important targets (e.g., cars and trucks) as many RCS points rather than a single-point RCS. For multiple RCS points, the simulation may optionally be programmed to take into account the occlusion and directionality caused between two parts of the same target.

The method may include determining an antenna response obtained with an electromagnetic sensor located in free space over a plurality of target distances. The method may further include determining multipath effects and near field effects on the electromagnetic sensor when the electromagnetic sensor is mounted on the vehicle for a plurality of target distances. The method may then include: based on the antenna responses with multipath effects and near-field effects, a response map of expected responses is generated for the electromagnetic sensor given a plurality of single point locations proximate to the vehicle, each expected response taking into account a corresponding angle and distance-related ground effect determined for the single point location. Further, the method may include outputting the response map to a sensor simulator configured to test for detection accuracy of the simulated object located at the particular single point location given the expected response defined by the response map for the particular single point location.

In some examples, each expected response includes a target radar cross-section adjusted for a corresponding angle and distance-dependent ground effect determined for the single point target location. The target radar cross-section for each expected response may depend on the target angle and target distance between the vehicle and the single point target location. Each expected response may have a target signal-to-noise ratio that is greater than a detection threshold of the electromagnetic sensor. In other words, as described herein, the plurality of target distances are one or more relative positions or distances from the electromagnetic sensor, including an expected distance or range of targets detected with the electromagnetic sensor when the electromagnetic sensor is mounted on the vehicle. Examples of the plurality of target distances may include a first distance (e.g., equal to or less than 20 meters) and a second distance (e.g., between 20 and 30 meters). The plurality of target distances may define a target distance as zero or near zero, or any other potential detection distance of the sensor, up to a maximum detectable target distance (e.g., 1000 meters).

In another example, a system includes at least one processor configured to perform the methods outlined above as well as the other methods set forth herein. A system is also described, comprising: means for performing the methods outlined above and other methods set forth herein; a non-transitory computer-readable medium, such as a computer-readable storage medium comprising instructions that, when executed, cause at least one processor to perform the methods outlined above and other methods set forth herein. A computer system including a simulation software program (e.g., a radar simulator) may be configured to utilize a response map of performance output from the above-described methods and other methods set forth herein. The computer software product may be configured to execute on computer hardware to cause the computer hardware to perform the methods outlined above and other methods set forth herein.

The present disclosure presents techniques for implementing a single point RCS method for radar simulation and is further described in the detailed description and figures. This document is mainly written in the context of radar sensors and radar simulators. In addition to radar, the techniques and systems described herein may be used to improve simulation of target responses to any high frequency sensor or sensing technology (e.g., microwave applications, wi-Fi, cellular phone applications). This summary is not intended to identify essential features of the claimed subject matter, nor is it intended to be used to determine the scope of the claimed subject matter.

Drawings

Details of one or more aspects for implementing a more accurate single-point RCS approach for radar simulation without increasing the degree of complexity are described in this document with reference to the following drawings, in which like numerals are used throughout to designate like or similar drawing features:

FIG. 1 illustrates an example environment for simulating components of a radar system using a more accurate single-point RCS approach in accordance with the techniques of this disclosure;

FIG. 2 illustrates a process for applying a more accurate single point RCS method in radar simulation in accordance with the techniques of the present invention; and

Fig. 3-1 through 3-8 illustrate example results of applying a more accurate single point RCS method in radar simulation in accordance with the techniques of the present invention.

Detailed Description

Radar is increasingly used for environmental awareness in automotive applications (e.g., autonomous driving control, semi-autonomous driving control, advanced safety warning, collision avoidance) due to its favorable capabilities in many driving situations. It is a challenge to develop a computer-based radar algorithm with performance specifications that desire to quickly process radar signals into various forms of radar data that are generated to support the perception tasks performed on the vehicle. The model may be used to increase radar processing speed; however, especially when employing machine learning techniques, deployment of the model may require some initial adjustment or training to reliably work in a wide range of scenarios. For reasons of time and/or cost, it is impractical for most radar developments to perform an exhaustive list of real world tests to generate large complex training data. Rather, simulations may generate these large data sets, which may then be used to evaluate, train, or tune the radar to perform in many different driving situations.

Various simulation methods may be used to train or tune the radar model. For example, full-wave based finite element methods (FEM: FINITE ELEMENT methods), moment methods (MOM: method of Moments), and incident and bouncing ray (SBR: shooting and Bouncing Rays) methods or other ray tracing techniques are some existing ways of generating large datasets of target responses (also referred to herein as "expected responses") to support large-scale automotive radar simulations. Although these different simulation techniques have increased speed in large part due to advances in computing technology, execution may still be too slow or cumbersome when used outside of an expensive computer laboratory.

To reduce complexity, radar simulations may represent the expected response (i.e., target response) as a simpler single-point RCS. In the real world, the RCS of a target depends on a number of factors. In addition to angular fluctuations, the RCS may vary according to target distance, mounting height of the radar, and height profile (elevation profile) of the road surface between the radar and target distance. For simplicity, the single point RCS associated with the target is generally represented as a distance-invariant variable, which is allowed to vary only as a function of angle (e.g., azimuth angle relative to the radar and target). Using these inaccurate single-point RCS estimates rather than through the complete geometry (e.g., height, width) of the target to represent that the target is an overall overcomplete of the actual radar response from the target; the RCS is inaccurately reported as dependent only on azimuth to the target. This simplifications allow many target responses to be simulated quickly, however, with much lower accuracy than other simulation techniques that can address other conditions affecting the target's RCS.

For example, small variations of the RCS at different reflection positions on the target can be considered by some simulations to improve the accuracy of the single-point RCS representation. The target may have multiple surfaces that may reflect radar signals differently from one surface to the next; the single scattering point may or may not reflect from a single concentration point on the target, but may reflect from an end or other portion of the target. This may shift the start of the radar response not only to depend on the angle, but also on the characteristics of the target surface at the target location. The accuracy of the simulated RCS can be improved by frequency scanning at the simulation angle and performing a fast fourier transform to obtain a range profile, and then using the distance from the range profile as the reflection position of the target.

However, even with such improvements, existing representations of single-point RCS representations in radar simulation may be inaccurate because the RCS simulation results still assume far-field conditions are met. For driving applications, some of the most important targets (e.g., trucks, cars) may have significantly high RCS and large dimensions in some directions. Ignoring distance-based fluctuations in its RCS (as is done with existing simulations) is not sufficient to accurately simulate radar response, especially at close distances. The far field conditions of the radar may not be properly represented and instead the simulation may be limited in its usefulness, e.g., effectively simulating radar performance over a minimum distance of only a few hundred meters. These and other drawbacks of the prior art simulation techniques appear not to be recognized or even understood outside the present disclosure. Some simulations consider the variation of RCS over close range, but do not consider or apply corrections to account for the variation of RCS to evaluate different near field and far field conditions. Indeed, rather than considering far field effects, these other simulations may use RCS values of 10 meters or less as representative RCS values for all distances, thereby limiting their technical effectiveness.

In addition to distance, existing simulations may not address other situations that cause RCS fluctuations in the real world, e.g., the target RCS representation of existing simulations does not take into account multipath effects caused by ground reflection of the road. That is, to minimize the complexity of the simulation, the presence of road reflection (which always appears to be present in the driving scene) is typically completely ignored, although the presence of road reflection causes significant changes in RCS behavior. It is therefore desirable to provide a faster way of modeling the change in target RCS values in many cases to provide more accurate training or tuning data for simulating radar performance in large or complex scenarios. For this reason, a simpler way to simulate the real target response is desired for tuning or training the radar model in the simulation, in particular to keep the target represented as a substantial way of a single point (scattering center) RCS map.

This document describes techniques and systems for implementing a more accurate single-point RCS method for radar simulation without increasing complexity compared to previous radar simulation techniques. As described herein, the radar simulator may be executed on a Central Processing Unit (CPU) and/or a Graphics Processing Unit (GPU) of a computer to generate a target RCS response from a traffic scene that accounts for angular variations other than distance. The RCS response may be determined using as inputs the scene information (e.g., the position, orientation, and velocity of all targets) and the basic parameters of the transmitter/receiver (position, orientation, field of view, and operating frequency) and may be used to generate an output from the radar simulation that takes into account the RCS response. The distance variations and multipath effects are analytically applied to the RCS pattern associated with the simulated target, which results in a radial RCS map that is dependent on distance and other conditions, rather than on angle alone. The response map may define expected responses, each including a target RCS adjusted for the corresponding angle and distance dependent ground effect of the single point target location simulation. The target RCS for each expected response may depend on the target angle and the target distance between the vehicle and the single point target location. Each expected response may have a target signal-to-noise ratio (SNR) that is greater than a detection threshold of the electromagnetic sensor.

For complex objects like cars, it is not an accurate or sufficient simulation to consider a single RCS value for all possible distances, as the size of the car results in or otherwise creates a near field region that extends a long distance encompassing the entire distance of the radar. As a result, the recorded RCS defined in the anechoic chamber is different from the RCS in far field conditions. To improve the accuracy of the simulation, the RCS pattern conveyed by the example response graphs generated in accordance with the described techniques has recorded fluctuations in the RCS at several distances, which when output to the sensor simulator, enables the simulation to better convey various distance-based fluctuations in the simulated antenna response for electromagnetic sensors subjected to the simulation test. This results in radar simulation that applies more realistic behavior to the target RCS, rather than overly simplifying the simulation by ignoring fluctuations in the RCS due to conditions other than just angle. In addition to being dependent on angle, the RCS defined for the target is also dependent on distance. In some cases, the RCS may depend on other considerations, such as target surface and multipath reflection caused by the road.

To achieve a more accurate single point radar RCS method, the techniques and systems may employ the following criteria. For all targets, not only cars and trucks, but also road multipath conditions are simulated, which greatly improves the accuracy of radar simulation. For at least some of the most important emulated targets (e.g. cars and trucks), distance-dependent fluctuations in the RCS of the target may be taken into account by assuming that the emulated radar is almost always at least partially in near-field mode. With the simulated road multipath conditions, distance-related fluctuations in its RCS can then be considered, where each target is represented by its own multi-dimensional (e.g., two-dimensional) RCS profile, the values of which are accurately related based at least in part on changes in target angle (e.g., azimuth) and target distance.

FIG. 1 illustrates an example simulation environment 100 in which components of a radar system 104 are simulated using a more accurate single-point RCS approach, in accordance with the techniques of the present invention. In environment 100, radar simulator 102 interfaces with electromagnetic simulator 134 and radar system 104, radar system 104 being simulated as being integrated within vehicle 106.

Radar system 104 may be of any type, including continuous wave or pulsed radar, frequency modulated or phase modulated radar, single Input Single Output (SISO) or Multiple Input Multiple Output (MIMO) radar, or some combination thereof, including Frequency Modulated Continuous Wave (FMCW) MIMO radar, with or without a Code Division Multiple Access (CDMA) scheme. The radar system 104 includes a MIMO antenna 114, at least one Monolithic Microwave Integrated Circuit (MMIC) 116, one or more processors 118, and a computer-readable storage medium (CRM) 120. The MMIC 116 includes circuitry and logic for transmitting and receiving radar signals via the MIMO antenna 114 and detecting the target 108. These components enable the transmission of the transmitted signal 110 and, in response, the detection of a corresponding received signal 112, which received signal 112 may be processed into radar data for output to the vehicle 106 to perform perceptually relevant tasks. The MMIC 116 may include amplifiers, mixers, switches, analog-to-digital converters, or filters for conditioning the radar signals. The MMIC 116 also includes logic units that perform in-phase/quadrature (I/Q) operations, such as modulation or demodulation, but are not limited thereto. The MMIC 116 may include at least one transmitter and at least one receiver, or a combined transceiver. In some cases, MIMO antenna 114 enables radar system 104 to form a field of view by forming steered or non-steered and wide or narrow beams. Steering and shaping may be achieved by analog beamforming or digital beamforming. The transmit subarray of MIMO antenna 114 may have, for example, a steered omnidirectional radiation pattern, or may produce a wide steerable beam to illuminate a large spatial volume. To achieve target angular accuracy and angular resolution, the receive subarray of MIMO antenna 114 may include multiple receive antenna elements to generate hundreds or thousands of narrowly steered beams using a technique known as digital beamforming. The radar system 104 may effectively monitor the environment 100 of the target 108. CRM 120 includes radar software 122, radar software 122 is configured to analyze radar received signal 112, detect or track target 108 based on radar received signal 112, and/or determine one or more characteristics (e.g., position or velocity) of detected or tracked target 108.

The radar simulator 102 is configured to model hardware and/or software of the radar system 104, e.g., to enable evaluation of the performance of the radar system 104 and for various simulation environments. Throughout the development cycle of radar system 104, radar simulator 102 may be used, for example, to evaluate different system designs (e.g., different hardware configurations or different modes of operation), evaluate different versions of radar software 122, or verify requirements. The use of the radar simulator 102 enables rapid discovery of design or implementation issues within the radar system 104 during the testing phase prior to designing, integrating, and performing field testing. The radar emulator 102 performs operations to emulate hardware of a radar system (e.g., portions of the MIMO antenna 114 and/or the MMIC 116), radar software 122 executed by the processor 124, or a combination thereof. The radar simulator 102 may also consider non-ideal characteristics of the radar system 104 or the environment 100, such as noise or nonlinearity. In particular, the radar simulator 102 may simulate phase noise, waveform nonlinearities, and uncorrelated noise within the MMIC 116. With these capabilities, the radar simulator 102 may have a similar noise floor and dynamic distance as the radar system 104. The radar emulator 102 includes at least one processor 124 and a computer readable storage medium 126.CRM 126 may be implemented by one or more memory devices that enable persistent and/or non-transitory data storage. In some cases, the processor 124 and CRM 126 are packaged together within an integrated circuit or on a printed circuit board. In other cases, the processor 124 and the CRM 126 may be implemented separately and operably coupled together such that the processor 124 may access instructions stored by the CRM 126.CRM 126 includes a radar hardware emulator 130 and a radar software emulator 132.

The radar hardware simulator 130 and the radar software simulator 132 execute to provide assurance that the radar system 104 can reliably help perform perceptually relevant tasks performed on the vehicle 106. The radar hardware simulator 130 and the radar software simulator 132 may simulate the radar system 104 to determine radar responses to be detected by the radar system 104 when one or more simulated targets 108 are present in the simulated radar field of view at different angles and distances. In general, the radar hardware emulator 130 and the radar software emulator 132 may define the radar system 104 as being configured to any type of mobile platform, including mobile machines or robotic devices. Although illustrated as a car, the vehicle 106 may represent a truck or other type of ground vehicle, including other types of motor vehicles (e.g., motorcycles, buses, tractors, semi-trailer trucks, or construction equipment), non-motor vehicles (e.g., bicycles), rail vehicles (e.g., trains or trams), and the like. The simulated mounting location of the radar system 104 may coincide with the top surface of the vehicle 106, and in other examples, the mounting location is on the underside or side of the vehicle 106. The radar hardware simulator 130 and the radar software simulator 132 may take into account the variations in far field and near field effects caused by the reflection planes on the ground or road surface where the vehicle 106 is located. For example, different mounting locations of the radar system 104 may be tested with the radar system 104 distributed at different mounting locations to simulate an instrumented field of view anywhere from 0 to 360 degrees around the vehicle 106.

In this example, processor 124 executes instructions for performing the operations of radar hardware emulator 130 and radar software emulator 132. The radar hardware emulator 130 and the radar software emulator 132 may be implemented using software executing on the processor 124, either alone or in combination with operations performed by other hardware (executing as firmware), or in combination thereof. The radar hardware simulator 130 and/or the radar software simulator 132 include interfaces for receiving environmental response data provided by the electromagnetic simulator 134. As one example, the environmental response data may include a single point RCS of the target. In general, radar hardware simulator 130 and/or radar software simulator 132 transform the environmental response data into a form that can be used by radar simulator 102. For example, the environmental response data is adjusted by the radar simulator 102 to account for the exact antenna response of the MIMO antenna 114. The radar hardware simulator 130 models the MMIC 116. Specifically, the radar hardware simulator 130 performs operations that simulate waveform generation, modulation, demodulation, multiplexing, amplification, frequency conversion, filtering, and/or analog-to-digital conversion operations performed by the MMIC 116 to determine an expected response. In other words, the radar hardware emulator 130 performs the operations of the MMIC 116 that occur between the processor 118 and the MIMO antenna 114. The radar hardware simulator 130 may consider the dynamic range of the radar system 104, as well as the presence of nonlinear effects and noise (e.g., in a driving environment). The radar hardware emulator 130 may be used to verify different hardware configurations and modes of operation of the radar system 104. The radar software simulator 132 models the radar software 122 of the radar system 104. In particular, the radar software emulator 132 performs digital baseband processing operations that emulate the operations performed by the processor 118. These operations may include fourier transforms (e.g., fast fourier transforms), noise floor estimation, clutter map generation, constant false alarm rate thresholding, object detection, and object position estimation (e.g., digital beamforming). In some cases, radar software emulator 132 includes a version of radar software 122. In this way, the radar simulator 102 may be used to verify software requirements and evaluate different software versions of the radar system 104. During operation, the radar simulator 102 accepts electromagnetic response data from the electromagnetic simulator 134 and generates a radar report. The radar report may be used to evaluate the performance of the radar system 104 for a given simulation environment. By providing accurate electromagnetic response data to train or tune the model evaluated by the radar simulator 102, the electromagnetic simulator 134 may improve the accuracy of radar simulation and provide assurance that the radar system 104 may manage a wide range of driving scenarios.

Electromagnetic emulator 134 includes at least one processor 136 and a computer readable storage medium 138. The CRM 138 may optionally include a radiation emitter 140 and a radiation tracker 142. The CRM 138 also includes an Electromagnetic (EM) module 144. The ray emitter 140, ray tracker 142, and EM module 144 may be implemented using hardware, software, firmware, or a combination thereof. In this example, the processor 136 executes instructions for performing the operations of the radiation emitter 140, the radiation tracker 142, and the EM module 144. If included in electromagnetic simulator 134, ray emitter 140 enables electromagnetic simulator 134 to adaptively emit electromagnetic rays during simulation. The ray tracker 142 enables the electromagnetic simulator 134 to track rays emitted during the simulation, including reflections of rays from the target in the simulation. The EM module 144 determines an electromagnetic response for each tracked ray and may perform other operations to adjust the determined response to account for other conditions. The radiation emitter 140, the radiation tracker 142, and the EM module 144 are discussed in more detail below.

Electromagnetic simulator 134 may generate simulated radar responses determined for each simulated target 108. Each simulation target 108 may be associated with a target RCS output from EM module 144. The target RCS output is input to a radar simulator to evaluate performance or to help tune or train a model performed by the radar system 104 before the radar system 104 is physically deployed in the real world. Electromagnetic simulator 134 simulates a radar response having a signal-to-noise ratio greater than a detection threshold of radar system 104. Each simulated target 108 is composed of one or more types of material that reflect radar signals. Depending on the application, the target 108 may represent an object of interest or clutter. In some cases, the target 108 is a moving target, such as another vehicle, a person, or an animal. In other cases, the target 108 is a stationary target, such as a continuous or discontinuous road barrier (e.g., traffic cone, concrete barrier, guardrail, or fence), tree, or parked vehicle. The electromagnetic simulator 134 may be specifically designed to simulate the most impacting driving target, including trucks, cars, or other large objects that appear in the driving lanes of the roadway without warning. The target RCS generated by electromagnetic simulator 134 is dependent on a number of conditions, including target azimuth and target distance. Unlike other simulators that report a target RCS that relies solely on angle, electromagnetic simulator 134 supports more accurate simulation and evaluation without additional complexity. The radar response generated by electromagnetic simulator 134 may take into account multipath reflections caused by the ground, which is also unique compared to other simulations.

FIG. 2 illustrates a process 200 for applying a more accurate single point RCS method in radar simulation in accordance with the techniques of the present invention. For ease of description, process 200 is performed in the context of the elements of fig. 1. Process 200 may be performed by electromagnetic simulator 134, radar simulator 102, or a combination of both. The steps of process 200 may be repeated, rearranged, omitted, or otherwise modified depending on the application. Likewise, the sensor simulator is commonly referred to as a radar simulator, however, the technique may be beneficial to other sensor simulator types. As explained with reference to process 200, when configured as a radar simulator, the sensor simulator is configured to test the performance of radar hardware or radar software in detecting a simulated object given the expected response of the simulated object as defined by the response map for that particular single point location.

As described above, far-field RCS simulation of large objects is not suitable for radar simulation itself due to near-field effects occurring near the target. Existing simulations ignore the existence of roads or cannot take into account their impact on the RCS in an appropriate manner. To take this into account and facilitate more accurate radar simulation, process 200 may be performed.

In step 202, an antenna response obtained with an electromagnetic sensor located in free space is determined over a plurality of target distances. For example, an actual antenna response representing RCS fluctuations caused by target distance may be simulated, including for close range. The simulation may be performed using existing RCS simulation techniques based on the SBR method described above or variations thereof to collect antenna responses simulated for each target distance and for a plurality of angular intervals arranged around the electromagnetic sensor.

For example, the far field RCS value in a one degree angular step around a target (e.g., a car) may be calculated using commercially available modeling software (e.g., HFSS), which results in the pattern 300-1 shown in FIG. 3-1. The simulated far field RCS of the car in fig. 3-1 includes a peak value of about 46 decibels per square meter (dBsm) within the pattern 300-1. Note that this does not correspond to the target RCS calculated by previous simulation techniques, as the measurement is made in the vicinity of the target (e.g., at a distance of less than 10 meters) and thus in the near field of the target, which tends to produce a peak slightly above 20dBsm, which is calculated using the following equation 1 (as illustrated in the Antenna Theory document (e.g., antenna Theory: ANALYSIS AND DESIGN, 4 th edition, wiley, 2016) of planis, constantaine a:

In the above equation 1, D _F is the distance from the start of the far field region, D is the longest dimension of the target object, and λ is the wavelength of the radar-transmitted signal 110. For a car of length 5 meters, calculating the RCS at a frequency of 76 gigahertz (GHz) results in a far-field zone distance of approximately 12 points 7 kilometers (km). Thus, the antenna at a distance of 10 meters from the car is clearly in the near field region. Parameter scans may be performed by the ray emitter 140 and ray tracker 142 in distance and angle to create a more accurate RCS radial map for the target representation in free space. That is, the traditional approach to emulating the target RCS in the far field is replaced by an emulation of a small antenna response at the target distance, which provides an accurate distance-dependent RCS for high RCS targets, where road reflection is still not considered. By adjusting the ray emitter and ray tracker to step through the distance and angle adjustments, the antenna response can be collected at angular intervals for each target distance.

To verify this step, a simulation with an actual antenna may be performed to measure RCS over distances from one meter to thirty kilometers. For example, the antenna may be designed as a seventy-six GHz single embedded feed patch antenna for both the transmitter and receiver, which are shifted in the direction of little radiation. Patch antennas with far field electric field (E-field) distribution coverage are shown in fig. 3-2, and corresponding gain patterns at zero degrees and ninety degrees in cross-section are shown in fig. 3-3. In fig. 3-3, the inner curved arc is used for a vertical plane parallel to the feed direction of the antenna shown in fig. 3-2. The outer arc in fig. 3-3 represents a plane perpendicular to the feed line direction. Commercial simulation software (such as that provided in HFSS) provides scattering parameters between antennas that are directly related to induced voltages on one antenna caused by the electric field generated by the other antenna. Consider equation 2, where E21 is the electric field at antenna 2 caused by voltage V1 applied to antenna 1, S21 is the scattering parameter/coupling between antenna 1 and antenna 2, and G is a constant that accounts for antenna characteristics (e.g., gain and impedance), as follows:

E21 =g·s21·v1 equation 2

Thus, in preparation for calculating the RCS from the scattering parameters, the following equations in equations 3, 4, 5 and 6 can be used:

In the above equation, r is the distance, E _21sc is the electric field generated from Rx at Tx taking into account scattering from the target surface, while E _21inc is the direct coupling between the two antennas due to low side waves, and E _31inc is the electric field at the distance r where the target is located, representing the incident field in free space. And E _31ref is the reference signal taken at one meter by the exact same antenna to avoid the need to directly measure the incident field and eliminate the constant G and voltage V1 from equation 2. To verify this RCS simulation step, a five percent sphere with a radius of one meter was simulated as it had an analytical formula of the comparison result. The theoretical result is 0.0079 square meters, while the simulation result is 0.0081 square meters, with an error of 2%, which is acceptable for many simulation applications. Thereafter, the RCS can be determined by simulation over distance using a car model.

The RCS of the resulting range profile is shown in FIGS. 3-4, including enlarged views. According to fig. 3-4, the rcs does converge to a high value above 40 dBsm. However, from an enlarged view, the RCS hardly fluctuates within the first 500 meters; at close distances up to 100 meters, the RCS fluctuates around 20dBsm, which is much closer to the measurement that is typically made within a distance close to the target. In this way, the RCS can be simulated in angular steps (e.g., three degrees) around the target car and for several distances. Fig. 3-5 show example RCS patterns for a range of 10 meters to 30 kilometers. The maximum RCS at long distances is almost 50dBsm, while at short distances the maximum RCS is slightly higher than 20dBsm, which better corresponds to the measured value.

In step 204, multipath effects and near field effects on the electromagnetic sensor are determined for a plurality of target distances when the electromagnetic sensor is mounted on the vehicle. For example, existing simulation techniques are used to consider distance-dependent road effects on a target RCS in a computationally efficient manner. For example, with a multi-dimensional RCS map defined for a target, multipath effects can be analytically applied (e.g., using a four-path model) to account for ground reflections without modeling the ground, which is otherwise computationally complex and time consuming. To include the effects of ground reflections, a four-path model is analytically applied to the RCS values for each angle and distance from the simulation results.

For example, a multipath model may be used to account for potentially multiple propagation paths of electromagnetic signals obtained by electromagnetic sensors at each target distance. By using the multipath model, a combination of response fluctuations caused by a plurality of propagation paths can be determined at each target distance. The multipath model may be a four-path model that considers signal propagation from the transmitter to the object and then reflects to the receiver in four different paths as shown in fig. 3-6.

Four paths of the four path model consider: direct-direct (dd) propagation path; direct-indirect (di) propagation paths; an indirect-direct (id) propagation path; and an indirect-indirect (ii) propagation path. The signal at the receiver is determined according to equations 7, 8, 9 and 10 as follows:

E _{Rx_total}＝ E_{Rx_dd}+ E_{Rx_di}+ E_{Rx_id}+ E_{Rx_ii} equation 7

E _{Rx_p} is the signal of the receiver on path p, L _1p and L _2p are the forward path length and the backward path length, respectively, both geometrically calculated as shown in fig. 3-6, E _0Tx is the signal strength one meter from the transmitter, k is the wave number 2 pi/λ, and Γ _i is the reflection coefficient of each path, which is equal to 1 for dd, equal to the ground reflection coefficient Γ _g for di and id, and equal to (Γ _g)² for ii, while a _T describes the RCS of the target in equation 11:

By multiplying the return signal in free space by r ², item a _T·E_0Tx is estimated from the simulation results and then applied in equation 7, assuming it is equal in the four paths for the application.

At step 206, a response map of expected responses of the electromagnetic sensor is generated given a plurality of single point locations in the vicinity of the vehicle based on the antenna responses having the multipath effect and the near field effect. A set of response fluctuations of the electromagnetic sensor may be determined at each target distance based on the antenna response. The multipath and near-field effects are used to account for antenna response and/or response fluctuations to produce an expected response to account for corresponding angle and distance dependent ground effects simulated for any single point location. For example, information obtained from the road multipath simulation and the near field simulation are combined to create a more accurate angle and distance dependent RCS map associated with a single point representation of the target of the rapid environmental simulation. For example, a response map is generated by applying the set of response fluctuations determined for each target distance at step 202 and the multipath effects and near-field effects determined for that target distance at step 204. The combination of response fluctuations caused by the multiple propagation paths at each target distance is used to determine the respective angle and distance dependent ground effects to be applied to the expected response defined in the response map.

At step 212, the response map is output to a sensor simulator configured to test the detection accuracy of the simulated object at a particular single point location given the expected response defined by the response map for that particular single point location. For example, the response map output from the electromagnetic simulator may be an RCS radial map that accounts for multipath effects, which may be used as an input to the radar hardware simulator 130 or the radar software simulator 132 to support rapid execution of radar simulations with little complexity or computational resources required to perform raytracing, but still enable considerable accuracy in simulating radar responses.

For example, the response map output at step 212 may be provided to an environmental response data interface of the sensor simulator, which converts the expected response defined by the response map into environmental response data, which is used to simulate electromagnetic echoes from the simulation object. Using the response maps, the sensor simulator can uniquely consider the corresponding angle and distance dependent ground effects defined by the response maps for each simulated object position.

In response to outputting the response map to the sensor simulator at step 212, the sensor simulator may be executed to test whether the hardware or software of the electromagnetic sensor accurately detects the simulated object at a particular single point location by processing the expected response defined by the response map for that particular single point location. The test results performed by the sensor simulator may be output (e.g., to a file, for display) to indicate whether the performance of the hardware or software meets the criteria for controlling the vehicle based on the sensor data output from the electromagnetic sensor.

To improve accuracy, previous steps are performed to extract the contours of the target car to move the scattering centers of the RCS from the origin of the target to the surface of the car. Frequency scanning is performed at a distance of ten meters in each angular step to obtain an impulse response, and then inverse fourier transform (IFFT) is performed to generate a range profile. The distance of the first significant peak is then stored in a look-up table for correction. Figures 3-7 show the contours resulting from this step; fig. 3-8 illustrate a single-point RCS multi-dimensional (e.g., angle and distance dependent) pattern with multipath ground effect, and overlapping of range profiles to convey how the RCS fluctuates in angle and based on other conditions (e.g., distance).

In step 208, accuracy may optionally be further improved in the response map by representing important targets (e.g., cars and trucks) as many RCS points rather than a single-point RCS. This enables the response map to be used as an input to a sensor simulator configured to test whether hardware or software of the electromagnetic sensor accurately detects a simulated object present at a plurality of single point locations, rather than detecting only a simulated object present at one location, given an expected response defined by the response map for each of the plurality of single point locations. For example, in addition, if the above point target simulation example is adjusted to account for the size of the target, the response map may indicate that movement in the reflection position is near a surface point of the target from which an expected response including an accurate RCS dependency for a given distance and/or angle may be applied.

At step 210, with multiple RCS points, the simulation may optionally be programmed to take into account occlusion and directionality caused between two portions of the same target. The response map may be further modified to enhance this type of sensor simulator to be able to test whether the hardware or software of the electromagnetic sensor accurately detects a simulated object having a first surface that is obscured by or has directionality affected by a second surface. For example, by grouping the expected response defined by the response map for each of the plurality of single point locations on the first surface separately from the expected response defined by the response map for each of the plurality of single point locations on the second surface, the simulated object has a different surface that appears at a different single point location than just one location. Given the expected responses defined by the response maps for each of the plurality of single point locations, the plurality of expected responses at different surfaces may be combined. For example, it may be considered to extract several scattering centers to represent the target, to improve the accuracy of the response map.

The process 200 may also include executing the sensor simulator on the same processor or a different processor executing steps 202 through 212, including providing the response map as input to the sensor simulator. For example, the sensor simulator may test whether the hardware or software of the electromagnetic sensor accurately detects the simulated object that appears at multiple single point locations, including when the simulated object has a first surface that is occluded or has directionality that is affected by a second surface.

Some further embodiments include the following:

Embodiment 1: a method, the method comprising the steps of: determining antenna responses obtained with electromagnetic sensors disposed in free space over a plurality of target distances; determining, for the plurality of target distances, a multipath effect and a near field effect on the electromagnetic sensor when the electromagnetic sensor is located on a vehicle; generating a response map of expected responses of the electromagnetic sensor given a plurality of single point locations in the vicinity of the vehicle, each expected response taking into account a corresponding angle and distance dependent ground effect determined for the single point location, based on the antenna responses having the multipath effect and the near field effect; and outputting the response map to a sensor simulator configured to test the detection accuracy of the simulated object located at the particular single point location given the expected response defined by the response map for that particular single point location.

Embodiment 2: the method of any of the preceding embodiments, wherein the expected responses each comprise a target radar cross-section adjusted for a corresponding angle and distance dependent ground effect determined for the single point target location.

Embodiment 3: a method according to any one of the preceding embodiments, wherein the target radar cross-section of each expected response is dependent on the target angle and target distance between the vehicle and the single point target location.

Embodiment 4: the method of any of the preceding embodiments, wherein each of the expected responses has a target signal-to-noise ratio that is greater than a detection threshold of the electromagnetic sensor.

Embodiment 5: the method of any preceding embodiment, wherein determining the antenna response comprises: antenna responses determined for each target distance and for a plurality of angular intervals disposed about the vehicle are collected.

Embodiment 6: the method of any preceding embodiment, wherein collecting the antenna response comprises: the ray emitter and ray tracker are adjusted to step through the distance and angle adjustments to collect antenna responses in the angular interval for each target distance.

Embodiment 7: the method according to any of the preceding embodiments, further comprising the steps of: determining a set of response fluctuations of the electromagnetic sensor at each of the target distances based on the antenna responses; and generating a response map by applying a set of response fluctuations determined for each target distance for which multipath effects and near field effects are emulated.

Embodiment 8: the method of any of the preceding embodiments, wherein determining multipath effects and near field effects on the electromagnetic sensor while on the vehicle comprises: the multipath model is used to consider a plurality of propagation paths of the electromagnetic signal obtained by the electromagnetic sensor at each target distance to determine a combination of response fluctuations caused by the plurality of propagation paths at each target distance.

Embodiment 9: the method of any preceding embodiment, wherein the multipath model comprises a four-path model to consider a direct-direct propagation path, a direct-indirect propagation path, an indirect-direct propagation path, and an indirect-indirect propagation path.

Embodiment 10: the method according to any of the preceding embodiments, wherein a combination of response fluctuations caused by multiple propagation paths at each target distance is used to determine a corresponding angle and distance dependent ground effect to be applied to an expected response defined in the response map.

Embodiment 11: the method of any preceding embodiment, wherein the electromagnetic sensor comprises radar and the sensor simulator comprises a radar simulator.

Embodiment 12: the method of any preceding embodiment, wherein the radar simulator is configured to test the performance of radar hardware or radar software when detecting a simulated object given an expected response of the simulated object as defined by a response map of a particular single point location.

Embodiment 13: the method of any of the preceding embodiments, wherein outputting the response map to the sensor simulator comprises: the response map is provided to an environmental response data interface of the sensor simulator, which converts an expected response defined by the response map into environmental response data to enable the sensor simulator to simulate electromagnetic echoes from simulated objects by uniquely considering corresponding angle and distance dependent ground effects defined by the response map for each simulated object location.

Embodiment 14: the method of any of the preceding embodiments, wherein outputting the response map to the sensor simulator comprises outputting the response map to configure the sensor simulator to test detection accuracy of the simulated object at a particular single point location by processing the expected response defined by the response map for the particular single point location.

Embodiment 15: the method according to any of the preceding embodiments, further comprising the steps of: the response map is output to another sensor simulator configured to test detection accuracy of another simulated object appearing at the plurality of single point locations based on an expected response defined by the response map for each of the plurality of single point locations.

Embodiment 16: the method according to any of the preceding embodiments, further comprising the steps of: the response map is output to another sensor simulator configured to test whether the electromagnetic sensor accurately detects a simulated object having a first surface that is obscured by or has directionality affected by a second surface based on a set of expected responses defined by the response map for each single point location on the first surface separate from the expected responses defined by the response map for each single point location on the second surface.

Embodiment 17: a system comprising at least one processor configured to perform the method of any of the preceding embodiments.

Embodiment 18: a system comprising means for performing the method according to any of the preceding embodiments.

Embodiment 19: a computer-readable storage medium comprising instructions that, when executed, cause at least one processor to perform the method of any of the preceding embodiments.

Embodiment 20: a computer system comprising a radar simulation software program configured to perform the method according to any one of the preceding embodiments.

Embodiment 21: a computer software product configured to be executed on computer hardware to cause the computer hardware to perform the method of any of the preceding embodiments.

While various embodiments of the present disclosure have been described in the foregoing description and shown in the accompanying drawings, it is to be understood that the disclosure is not so limited, but may be practiced in various ways within the scope of the appended claims. From the foregoing description, it will be apparent that various changes may be made without departing from the scope of the disclosure or the scope defined by the following claims. Problems associated with simulating environmental electromagnetic responses may occur in other systems. Thus, while improving the electromagnetic response of a radar simulator is described, the techniques described previously may be applied to other systems that simulate other electromagnetic sensors.

The use of "or" and grammatical-related terms denotes a non-exclusive alternative without limitation unless the context clearly indicates otherwise. As used herein, a phrase referring to "at least one" in a list of items refers to any combination of these items, including individual members. As an example, "at least one of: a. b or c "is intended to cover a, b, c, a-b, a-c, b-c, and a-b-c, as well as any combination of multiples of the same element (e.g., a-a-a, a-a-b, a-a-c, a-b-b, a-c-c, b-b-b, b-b-c, c-c, and c-c, or any other order of a, b, and c).

Claims

1. A method, the method comprising the steps of:

Determining an antenna response obtained with an electromagnetic sensor located in free space over a plurality of target distances;

determining, for the plurality of target distances, a multipath effect and a near field effect on the electromagnetic sensor when the electromagnetic sensor is located on a vehicle;

Generating a response map of expected responses of the electromagnetic sensor given a plurality of single point locations in the vicinity of the vehicle, each expected response taking into account a corresponding angle and distance dependent ground effect determined for that single point location, based on the antenna responses having the multipath effect and the near field effect; and

The response map is output to a sensor simulator configured to test the detection accuracy of a simulated object located at a particular single point location given the expected response defined by the response map for that particular single point location.

2. The method of claim 1, wherein the expected responses each comprise a target radar cross-section adjusted for a corresponding angle and distance dependent ground effect determined for the single point target location.

3. The method of claim 2, wherein the target radar cross-section of each of the expected responses is dependent on a target angle and a target distance between the vehicle and the single point target location.

4. A method according to claim 2 or 3, wherein the expected responses each have a target signal to noise ratio greater than a detection threshold of the electromagnetic sensor.

5. The method of any of claims 1 to 4, wherein the step of determining the antenna response comprises the steps of:

The antenna responses determined for each of the target distances and for a plurality of angular intervals disposed about the vehicle are collected.

6. The method of claim 5, wherein the step of collecting the antenna response comprises the steps of:

The ray emitter and ray tracker are adjusted to step through the distance and angle adjustments to collect antenna responses in the angular interval for each of the target distances.

7. The method according to any one of claims 1 to 6, further comprising the step of:

determining a set of response fluctuations of the electromagnetic sensor at respective ones of the target distances based on the antenna responses; and

The response map is generated by applying the set of response fluctuations determined for each of the target distances, and the multipath effect and the near field effect are simulated for that target distance.

8. The method of claim 7, wherein determining multipath effects and near field effects on the electromagnetic sensor when the electromagnetic sensor is located on the vehicle comprises:

A multi-path model is used to consider a plurality of propagation paths of electromagnetic signals obtained by the electromagnetic sensor at respective ones of the target distances to determine a combination of the response fluctuations caused by the plurality of propagation paths at the respective ones of the target distances.

9. The method of claim 8, wherein the multipath model comprises a four-path model for taking into account direct-direct propagation paths, direct-indirect propagation paths, indirect-direct propagation paths, and indirect-indirect propagation paths.

10. The method of claim 8 or 9, wherein a combination of the response fluctuations caused by the plurality of propagation paths at each of the target distances is used to determine the corresponding angle and distance dependent ground effect to be applied to the expected response defined in the response map.

11. The method of any of claims 1-10, wherein the electromagnetic sensor comprises radar and the sensor emulator comprises a radar emulator.

12. The method of claim 11, wherein the radar simulator is configured to test performance of radar hardware or radar software in detecting the simulated object given the expected response of the simulated object defined by the response map for that particular single point location.

13. The method of any of claims 1 to 12, wherein the step of outputting the response map to the sensor simulator comprises at least one of:

Providing the response map to an environmental response data interface of the sensor simulator, the environmental response data interface converting the expected response defined by the response map into environmental response data to enable the sensor simulator to simulate electromagnetic echoes from simulated objects by uniquely considering the corresponding angle and distance dependent ground effects defined by the response map for each simulated object location;

outputting the response map to configure the sensor simulator to test the detection accuracy of the simulated object at the particular single point location by processing the expected response defined by the response map for that particular single point location;

Outputting the response map to configure the sensor simulator to test detection accuracy of another simulation object appearing at a plurality of single point locations based on the expected response defined by the response map for each of the plurality of single point locations; or alternatively

The response map is output to configure the sensor simulator to test whether the electromagnetic sensor accurately detects a simulated object having a first surface that is obscured by or has directionality affected by a second surface, the test being based on a set of expected responses defined by the response map for each of a plurality of single point locations on the first surface that are separate from expected responses defined by the response map for each of the plurality of single point locations on the second surface.

14. A computer-readable storage medium comprising instructions that, when executed, cause at least one processor to perform the method of any one of claims 1 to 13.

15. A system comprising at least one processor configured to perform the method of any one of claims 1 to 13.