Method and system for separating multiple radar signals based on hardware accelerator

Abstract

The invention discloses a method and system for separating multiple radar signals based on a hardware accelerator, and relates to the technical field of hardware accelerators, and the method comprises the steps: determining a corresponding separation mechanism based on each pseudo-random code, the model of a millimeter wave radar and the models of multiple transmitters; the identity mark of each radar signal is determined according to each pseudo-random code, a hardware accelerator and a separation mechanism, and each radar signal is separated, so that the separation precision of each radar signal is improved; therefore, the early warning range of the intelligent vehicle under autonomous driving is determined based on the sensing range of the millimeter wave radar, the camera shooting range of the intelligent vehicle and each signal receiving path; according to the early warning range of the intelligent vehicle, the number of the radar signals and the service life of the intelligent vehicle, the hardware accelerator is triggered to autonomously optimize the signal separation level of each radar signal, and the separation effect of the hardware accelerator on each radar signal in different environments is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of hardware accelerators, and in particular, to a method and system for separating multiple radar signals based on a hardware accelerator. Background Art

[0002] With the development of technology, intelligent vehicles are gradually applied to people's lives and perform intelligent driving on streets or roads. In the prior art, millimeter-wave radars are installed on intelligent vehicles and receive radar signals emitted by transmitters. However, multiple transmitters emit multiple radar signals, and multiple radar signals are fed back to the same millimeter-wave radar. Only pseudo-random codes are used for separation, which affects the separation accuracy of each radar signal. Summary of the Invention

[0003] The purpose of the present invention is to overcome the deficiencies of the prior art. The present invention provides a method and system for separating multiple radar signals based on a hardware accelerator.

[0004] An embodiment of the present invention provides a method for separating multiple radar signals based on a hardware accelerator, including: when a millimeter-wave radar moves with the movement of an intelligent vehicle, receiving multiple radar signals emitted by each transmitter based on the millimeter-wave radar; determining signal reception paths according to the millimeter-wave radar and each transmitter, and determining corresponding pseudo-random codes according to the multiple radar signals, corresponding signal reception paths, and a pseudo-random code generator; determining corresponding separation mechanisms based on each pseudo-random code, the model of the millimeter-wave radar, and the models of multiple transmitters; determining identity markers of each radar signal according to each pseudo-random code, the hardware accelerator, and the separation mechanism, and separating each radar signal; determining a warning range of the intelligent vehicle during autonomous driving based on the sensing range of the millimeter-wave radar, the imaging range of the intelligent vehicle, and each signal reception path; triggering the hardware accelerator to autonomously optimize the signal separation level of each radar signal according to the warning range of the intelligent vehicle, the number of radar signals, and the service life of the intelligent vehicle.

[0005] An embodiment of the present invention provides a system for separating multiple radar signals based on a hardware accelerator. The system for separating multiple radar signals based on a hardware accelerator is applied to the above method for separating multiple radar signals based on a hardware accelerator. The system for separating multiple radar signals based on a hardware accelerator includes:

[0006] A radar signal module, configured to receive multiple radar signals emitted by each transmitter based on a millimeter-wave radar when the millimeter-wave radar moves with the movement of an intelligent vehicle;

[0007] A pseudo-random code module, which is used to determine a signal reception path according to a millimeter-wave radar and each transmitter, and determine a corresponding pseudo-random code according to multiple radar signals, the corresponding signal reception path, and a pseudo-random code generator;

[0008] A separation module, which is used to determine a corresponding separation mechanism based on each pseudo-random code, the model of the millimeter-wave radar, and the models of multiple transmitters;

[0009] An identity marking module, which is used to determine the identity marking of each radar signal according to each pseudo-random code, a hardware accelerator, and the separation mechanism, and separate each radar signal;

[0010] A warning range module, which is used to determine the warning range of an intelligent vehicle during autonomous driving based on the sensing range of the millimeter-wave radar, the imaging range of the intelligent vehicle, and each signal reception path;

[0011] An optimization module, which is used to trigger the hardware accelerator to autonomously optimize the signal separation level of each radar signal according to the warning range of the intelligent vehicle, the number of radar signals, and the service life of the intelligent vehicle.

[0012] In an embodiment of the present invention, by the method in the embodiment of the present invention, when the millimeter-wave radar moves along with the movement of the intelligent vehicle, multiple radar signals emitted by each transmitter are received based on the millimeter-wave radar; a signal reception path is determined according to the millimeter-wave radar and each transmitter, and a corresponding pseudo-random code is determined according to multiple radar signals, the corresponding signal reception path, and a pseudo-random code generator; a corresponding separation mechanism is determined based on each pseudo-random code, the model of the millimeter-wave radar, and the models of multiple transmitters; the identity marking of each radar signal is determined according to each pseudo-random code, a hardware accelerator, and the separation mechanism, and each radar signal is separated, taking into account the overall consideration of each pseudo-random code, the hardware accelerator, and the separation mechanism, and improving the separation accuracy of each radar signal.

[0013] Therefore, the warning range of the intelligent vehicle during autonomous driving is determined based on the sensing range of the millimeter-wave radar, the imaging range of the intelligent vehicle, and each signal reception path; the hardware accelerator is triggered to autonomously optimize the signal separation level of each radar signal according to the warning range of the intelligent vehicle, the number of radar signals, and the service life of the intelligent vehicle, so as to facilitate the regulation of the signal separation level of each radar signal by the hardware accelerator, and ensure the separation effect of the hardware accelerator on each radar signal in different environments. Description of the Drawings

[0014] Figure 1 It is a schematic diagram of an application scenario of a method for separating multiple radar signals based on a hardware accelerator in an embodiment;

[0015] Figure 2It is a schematic flow chart of the method for separating multiple radar signals based on a hardware accelerator in an embodiment of the present invention;

[0016] Figure 3 It is a schematic diagram of the structural composition of the system for separating multiple radar signals based on a hardware accelerator in an embodiment of the present invention;

[0017] Figure 4 It is a schematic diagram of the signal processing of the hardware accelerator in the method for separating multiple radar signals based on a hardware accelerator in an embodiment of the present invention;

[0018] Figure 5 It is a schematic diagram of the hardware accelerator in the method for separating multiple radar signals based on a hardware accelerator in an embodiment of the present invention. Detailed implementation manners

[0019] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention.

[0020] Embodiment 1

[0021] The method for separating multiple radar signals based on a hardware accelerator provided by this application is applied to an application environment as Figure 1 shown. Among them, the computer 102 communicates with the server 104 through the network. Among them, the terminal 102 is but not limited to various personal computers, servers, and hardware accelerators, and the server 104 is implemented by an independent server or a server cluster composed of servers.

[0022] Embodiment 2

[0023] Please refer to Figures 1 to 5 , a method for separating multiple radar signals based on a hardware accelerator, which is applied to a scenario of separating multiple radar signals based on a hardware accelerator; the method for separating multiple radar signals based on a hardware accelerator includes:

[0024] Step S11: When the millimeter-wave radar moves with the movement of the intelligent vehicle, receive multiple radar signals emitted by each transmitter based on the millimeter-wave radar;

[0025] Step S12: Determine the signal reception path according to the millimeter-wave radar and each transmitter, and determine the corresponding pseudo-random code according to the multiple radar signals, the corresponding signal reception path, and the pseudo-random code generator;

[0026] Step S13: Determine the corresponding separation mechanism based on each pseudo-random code, the model of the millimeter-wave radar, and the models of multiple transmitters;

[0027] Step S14: Determine the identity tags of each radar signal according to each pseudo-random code, the hardware accelerator, and the separation mechanism, and separate each radar signal;

[0028] Step S15: Determine the warning range of the intelligent vehicle during autonomous driving based on the sensing range of the millimeter-wave radar, the imaging range of the intelligent vehicle, and each signal reception path;

[0029] Step S16: Trigger the hardware accelerator to autonomously optimize the signal separation level of each radar signal according to the warning range of the intelligent vehicle, the number of radar signals, and the service life of the intelligent vehicle.

[0030] In step S11, when the millimeter-wave radar moves with the intelligent vehicle, multiple radar signals emitted by each transmitter are received based on the millimeter-wave radar.

[0031] In the specific implementation process of the present invention, the specific steps are as follows:

[0032] S111: The millimeter-wave radar is installed on the intelligent vehicle. The millimeter-wave radar moves with the intelligent vehicle and enters various environments;

[0033] S112: Determine the sensing range of the millimeter-wave radar according to the position of the millimeter-wave radar, the moving speed of the millimeter-wave radar, and the signal interference level of the environment;

[0034] S113: Respond to corresponding multiple transmitters based on the position of the millimeter-wave radar and the sensing range of the millimeter-wave radar. The multiple transmitters emit multiple radar signals in the response state. At this time, the multiple radar signals are received by the millimeter-wave radar.

[0035] In the embodiment of the present application, the millimeter-wave radar is installed on the intelligent vehicle. The millimeter-wave radar moves with the intelligent vehicle and enters various environments; determine the sensing range of the millimeter-wave radar according to the position of the millimeter-wave radar, the moving speed of the millimeter-wave radar, and the signal interference level of the environment, which comprehensively considers the position of the millimeter-wave radar, the moving speed of the millimeter-wave radar, and the signal interference level of the environment, realizes the accuracy of the sensing range of the millimeter-wave radar, and accurately controls the sensing range of the millimeter-wave radar.

[0036] At this time, millimeter-wave radars are usually installed at the front or rear of intelligent vehicles, and the specific location depends on their designed uses. For example, if it is used for forward collision warning, it is installed on the front bumper of the vehicle; if it is used for blind spot monitoring, it is installed near the rearview mirror of the vehicle. As the intelligent vehicle moves, the millimeter-wave radar will also move accordingly, which means that the radar needs to be able to adapt to different road conditions, speeds, and direction changes. Intelligent vehicles travel in various environments, including urban streets, highways, rural roads, tunnels, bridges, etc. Each environment has an impact on the performance of the radar, such as obstacles like buildings, trees, and other vehicles blocking or reflecting radar signals.

[0037] The location where the millimeter-wave radar is located, the moving speed of the millimeter-wave radar, and the signal interference level of the environment are introduced. At this time, the sensing range of the radar is affected by its installation location. For example, if the radar is installed at a higher position on the vehicle, its vertical sensing range will be wider; if it is installed at a lower position, it will focus more on the horizontal sensing range. As the vehicle speed changes, the sensing range of the radar also needs to be adjusted accordingly. When driving at high speed, the radar needs a farther sensing range to detect potential obstacles in advance; when driving at low speed or parking, more attention is paid to precise detection at close range. The signal interference level in the environment has a direct impact on the sensing range of the radar. For example, in an urban environment, due to the presence of a large number of radio devices (such as cell phone base stations, Wi-Fi routers, etc.), the radar faces high signal interference, resulting in a reduced sensing range. At this time, the radar needs to adopt more advanced signal processing algorithms to reduce the interference impact.

[0038] Specifically, when an intelligent sedan is driving at high speed on a highway, the millimeter-wave radar needs to have a farther sensing range to detect vehicles and obstacles ahead in advance. At this time, the radar system will automatically adjust parameters such as its transmission power and receiving gain to ensure high detection accuracy even in the high-speed driving state. At the same time, if the vehicle enters an area with severe radio interference (such as near a large radio transmitting tower), the radar system will adopt interference suppression algorithms to reduce the interference impact and thus maintain a stable sensing range.

[0039] In addition, a sensing range matching table is pre-made. This sensing range matching table comprehensively considers the installation location of the radar, its moving speed, and the signal interference level of the environment it is in, and provides a preset sensing range value for each combination.

[0040] Example of the sensing range matching table:

[0041] Radar position Moving speed (km / h) Signal interference level Induction range (meters) On the front bumper 60 Low 150 On the front bumper 120 Low 180 Near the rearview mirror 60 Medium 120 Near the rearview mirror 120 High 90 Roof-mounted 60 Low 200

[0042] In this example, when the radar is installed on the front bumper of a vehicle, the vehicle is traveling at a speed of 60 km / h, and the environmental signal interference level is low, the sensing range of the radar is preset to 150 meters.

[0043] Therefore, based on the location of the millimeter-wave radar and the corresponding multiple transmitters in response to the sensing range of the millimeter-wave radar, the multiple transmitters transmit multiple radar signals in the response state. At this time, the multiple radar signals are received by the millimeter-wave radar, and the millimeter-wave radar and the transmitters interact with each other, and the influence of the location of the millimeter-wave radar and the sensing range of the millimeter-wave radar is fully considered to ensure the response of the multiple transmitters, and then multiple radar signals are output.

[0044] At this time, after the millimeter-wave radar is installed on the intelligent vehicle, its specific location (such as near the front bumper, roof, rearview mirror, etc.) determines its field of view direction and the obstacle detection range; the transmitters around the intelligent vehicle include the radar systems of other intelligent vehicles, roadside traffic monitoring devices, wireless communication base stations, etc.; the millimeter-wave radar identifies different transmitters by analyzing the received signal characteristics (such as frequency, waveform, modulation method, etc.); once a specific transmitter is identified, the radar decides whether to respond according to its location and sensing range; the response includes adjusting the transmission power, changing the scanning mode, optimizing the signal processing algorithm, etc. to ensure effective reception of the signal from the transmitter.

[0045] In the response state, the multiple transmitters will emit radar signals, and these signals include different types of radar signals such as continuous wave (CW), frequency-modulated continuous wave (FMCW), and pulsed radar; the millimeter-wave radar receives these signals through its antenna array and performs preliminary signal processing such as filtering, amplification, and down-conversion; the millimeter-wave radar receives these signals through its antenna array and performs preliminary signal processing such as filtering, amplification, and down-conversion.

[0046] Specifically, assume that an intelligent car is traveling on a highway, and the millimeter-wave radar is installed on the front bumper of the vehicle; at this time, the sensing range of the radar is set to 200 meters ahead (considering the high-speed driving state of the vehicle and environmental factors).

[0047] The radar is located at the front of the vehicle and can cover a relatively wide area in front; the sensing range is set to 200 meters to ensure that obstacles in front can be detected in advance when driving at high speed; the radar is located at the front of the vehicle and can cover a relatively wide area in front; the sensing range is set to 200 meters to ensure that obstacles in front can be detected in advance when driving at high speed; the radar system on the truck emits FMCW signals. After the radar system of the intelligent car receives these signals, preliminary processing such as filtering and amplification is performed; subsequently, through the fast Fourier transform (FFT) algorithm, the radar system calculates information such as the relative distance, relative speed, and azimuth angle between the truck and the car, and this information is used in the path planning and obstacle avoidance decision-making of the intelligent car.

[0048] In step S12, the signal reception path is determined according to the millimeter-wave radar and each transmitter, and the corresponding pseudo-random code is determined according to multiple radar signals, the corresponding signal reception paths, and the pseudo-random code generator.

[0049] In the specific implementation process of the present invention, the specific steps are as follows:

[0050] S121: Interact between the millimeter-wave radar and each transmitter.

[0051] S122: Determine the signal reception path according to the interaction between the millimeter-wave radar and each transmitter.

[0052] S123: Determine the pseudo-random code generator based on the matching of the millimeter-wave radar, each transmitter, and the intelligent vehicle.

[0053] S124: Interact between multiple radar signals, the corresponding signal reception paths, and the pseudo-random code generator.

[0054] S125: Determine the corresponding pseudo-random code according to the interaction between multiple radar signals, the corresponding signal reception paths, and the pseudo-random code generator. At this time, each pseudo-random code changes dynamically with the change of the usage time of the pseudo-random code generator.

[0055] In the embodiment of the present application, interact between the millimeter-wave radar and each transmitter; determine the signal reception path according to the interaction between the millimeter-wave radar and each transmitter, realizing the interaction between the millimeter-wave radar and each transmitter and ensuring the accuracy of the signal reception path.

[0056] At this time, the millimeter-wave radar system actively or passively communicates with multiple surrounding transmitters, which are radar systems of other vehicles, radar devices of traffic infrastructure (such as smart streetlights, traffic signals), or roadside units (RSUs) dedicated to vehicle-to-vehicle communication (V2X); the content of the interaction includes signal synchronization, data exchange, status update, etc.; the millimeter-wave radar needs to synchronize the time reference with the transmitter to ensure the temporal consistency of the transmitted and received signals, which is crucial for achieving accurate distance measurement and speed estimation; the radar system shares environmental perception data, such as the position and speed of detected obstacles, with the transmitter, and this data exchange helps enhance the situational awareness of the entire transportation system.

[0057] After establishing the interaction with the transmitter, the millimeter-wave radar system determines the optimal signal reception path based on the characteristics of the received signals, which include signal strength, phase, angle of arrival (AoA), angle of departure (AoD), etc.; the purpose of determining the optimal reception path is to optimize the signal quality and reduce the impact of multipath effects and interference.

[0058] The radar system selects the path with the highest signal strength as the main reception path because this path usually has the best signal-to-noise ratio; by comparing the phase differences of the signals on different paths, the radar system identifies the signal distortion caused by multipath effects and selects the path with the best phase consistency; using the beamforming ability of the antenna array, the radar system measures the angle of arrival and angle of departure of the signal, thereby determining the optimal reception direction of the signal.

[0059] Specifically, when the car radar interacts with the truck radar, it will find that there are multiple signal paths; one path is propagated through the direct line of sight (LoS), while the other path is propagated through the reflection of roadside buildings; the car radar determines that the direct line-of-sight path has the highest signal quality and the lowest interference level by analyzing the received signal strength, phase, and angle of arrival; therefore, the car radar selects this path as the main path for communicating with the truck radar; at the same time, it also uses the signals on other paths as redundant information to improve the robustness of communication.

[0060] Furthermore, a pseudo-random code generator is determined based on the matching of the millimeter-wave radar, each transmitter, and the intelligent vehicle, realizing the matching of the millimeter-wave radar, each transmitter, and the intelligent vehicle, accommodating the overall consideration of the millimeter-wave radar, each transmitter, and the intelligent vehicle, and ensuring the adaptability of the pseudo-random code generator.

[0061] At this time, a pseudo-random code generator is determined according to the specific matching relationship among the millimeter-wave radar, each transmitter, and the intelligent vehicle. The function of this generator is to generate a unique pseudo-random code sequence, which will be used in the modulation or demodulation process of the radar signal to achieve signal differentiation and identification.

[0062] First, the system needs to establish the matching relationship among the millimeter-wave radar, the transmitter, and the intelligent vehicle. This matching is based on the vehicle's identity identification (such as license plate number, VIN code), location information, speed data, etc. The establishment of the matching relationship is achieved through vehicle-to-vehicle communication (V2V) or vehicle-to-infrastructure communication (V2I). For example, the intelligent vehicle exchanges information with the surrounding transmitters through dedicated short-range communication technology (DSRC) or cellular vehicle-to-everything communication technology (C-V2X).

[0063] Once the matching relationship is established, the system will configure the pseudo-random code generator according to this matching information. The configuration of the generator involves selecting a specific pseudo-random code algorithm, setting initialization parameters (such as seed value), determining the code length, etc. After the configuration is completed, the pseudo-random code generator starts to work and generates a unique pseudo-random code sequence, which will be used to modulate the radar signal or to demodulate the received signal at the receiving end. Optionally, the generation of the pseudo-random code sequence is dynamic, that is, it will change continuously over time, and this dynamic change helps to enhance the security and anti-interference ability of the signal.

[0064] Specifically, assume that there are two intelligent vehicles A and B driving on a highway. Vehicle A is equipped with a millimeter-wave radar system, while vehicle B is equipped with a radar transmitter (which is part of another millimeter-wave radar system or a device specifically used for vehicle-to-vehicle communication). When vehicles A and B approach to a certain distance, they establish a connection through V2V communication technology. The radar system of vehicle A obtains the identity identification (such as VIN code) and location information of vehicle B.

[0065] At the same time, the transmitter of vehicle B also obtains the relevant information of vehicle A. In this way, a matching relationship based on identity and location is established between the two.

[0066] Based on the established matching relationship, the radar system of Vehicle A selects a pseudo-random code algorithm (such as the LFSR algorithm) and sets the corresponding initialization parameters (such as a seed value based on the VIN code of Vehicle B and the current time); the radar system of Vehicle A sends this configuration information to the transmitter of Vehicle B for synchronous configuration at the receiving end; after the configuration is completed, the radar system of Vehicle A starts to generate a unique pseudo-random code sequence and uses this sequence to modulate the transmitted radar signal; when the transmitter of Vehicle B receives the signal from Vehicle A, it uses the same pseudo-random code generator configuration information to generate a matching pseudo-random code sequence and uses it to demodulate the received signal; since the pseudo-random code sequence is unique and dynamically changing, even if there are other interfering signals or noises, the transmitter of Vehicle B can accurately identify the signal from Vehicle A.

[0067] Therefore, interactions are carried out among multiple radar signals, the corresponding signal reception paths, and the pseudo-random code generator; the corresponding pseudo-random codes are determined based on the interactions among multiple radar signals, the corresponding signal reception paths, and the pseudo-random code generator. At this time, each pseudo-random code changes dynamically with the usage time of the pseudo-random code generator, realizing the interactions among multiple radar signals, the corresponding signal reception paths, and the pseudo-random code generator, ensuring the overall consideration of multiple radar signals, the corresponding signal reception paths, and the pseudo-random code generator, and achieving precise control of the pseudo-random code.

[0068] At this time, complex interactions occur among multiple radar signals (from different emission sources, such as the radar systems of other intelligent vehicles, radar devices of traffic infrastructure, etc.), their respective corresponding signal reception paths, and the pseudo-random code generator. This interaction aims to ensure the correct reception, decoding, and identification of radar signals, while using pseudo-random codes to enhance the anti-interference ability and security of the signals; at the same time, multiple radar signals propagate in space and reach the receiving end (such as the millimeter-wave radar system of an intelligent vehicle) through different paths; the receiving end determines the optimal signal reception path based on the received signal characteristics (such as intensity, phase, arrival angle, etc.); when the pseudo-random code generator receives signals from different emission sources, it generates corresponding pseudo-random code sequences according to preset algorithms and parameters, and these pseudo-random code sequences are used to demodulate or decode the received signals to achieve signal differentiation and identification; multiple components within the system (radar signals, reception paths, pseudo-random code generator) carry out real-time and dynamic interactions to adapt to the changing communication environment.

[0069] The system determines the pseudo-random code corresponding to each radar signal based on the interaction results in the previous steps. This pseudo-random code is the key for decoding and identifying the received radar signals. The system identifies different transmitters based on the characteristics of the received radar signals (such as modulation mode, frequency, waveform, etc.) and the determined signal reception path. For each identified transmitter, the system generates the corresponding pseudo-random code sequence according to its preset pseudo-random code algorithm and parameters. The pseudo-random code generator continuously changes the generated pseudo-random code sequence over time. This dynamic change helps enhance the security and anti-jamming ability of the signal. All components within the system (radar signal transmitter, receiver, pseudo-random code generator) synchronize this dynamic change according to the preset rules and protocols.

[0070] Specifically, in the scenario of an intersection, the communication system of the intelligent traffic light identifies these two transmitters based on the characteristics of the radar signals received from Vehicle A and Vehicle B and the determined signal reception path. For Vehicle A and Vehicle B, the communication system of the intelligent traffic light generates the corresponding pseudo-random code sequences according to their respective preset pseudo-random code algorithms and parameters. These pseudo-random code sequences are used for demodulating or decoding the radar signals received from Vehicle A and Vehicle B. Over time, the pseudo-random code generator of the intelligent traffic light continuously changes the generated pseudo-random code sequence. At the same time, the radar systems of Vehicle A and Vehicle B also adjust the configuration of their pseudo-random code generators according to the synchronization signals received from the intelligent traffic light to ensure communication synchronization and correct decoding with the intelligent traffic light. This dynamic change enhances the security and anti-jamming ability of the entire communication system.

[0071] In step S13, determine the corresponding separation mechanism based on each pseudo-random code, the model of the millimeter-wave radar, and the models of multiple transmitters;

[0072] In the specific implementation process of the present invention, the specific steps are as follows:

[0073] S131: Obtain each pseudo-random code;

[0074] S132: Interact on each pseudo-random code, the model of the millimeter-wave radar, and the models of multiple transmitters;

[0075] S133: Determine the first separation parameter based on each pseudo-random code and the model of the millimeter-wave radar, and determine the second separation parameter based on each pseudo-random code and the models of multiple transmitters;

[0076] S134: Determine the corresponding separation mechanism according to the first separation parameter, the second separation parameter, and the hardware accelerator.

[0077] In an embodiment of the present application, various pseudo-random codes are obtained; the interaction of various pseudo-random codes, the model of the millimeter-wave radar, and the models of multiple transmitters is realized, achieving the multiple interaction of various pseudo-random codes, the model of the millimeter-wave radar, and the models of multiple transmitters;

[0078] At this time, various pseudo-random codes are obtained. These pseudo-random codes have been generated in the previous steps and are used to modulate radar signals; the model information of the millimeter-wave radar and the transmitter is obtained. This information is stored in the system's database or configuration file and is also obtained from the radar / transmitter itself through a real-time communication interface; the model information includes the manufacturer, model name, serial number, etc. of the radar / transmitter, and this information is crucial for subsequent signal processing and identification; the system interacts the pseudo-random codes with the corresponding radar / transmitter model information and establishes an association relationship between them. This association relationship exists in the form of a data table, a mapping file, or a memory structure, so that the system can quickly access and process it; optionally, over time, the system will receive information about new radar / transmitters joining or existing radar / transmitters leaving; in this case, the system needs to dynamically update the association relationship between the pseudo-random codes and the radar / transmitter model information.

[0079] Specifically, assume that in an intelligent transportation monitoring system, there are two intelligent vehicles (Vehicle A and Vehicle B) and a traffic signal equipped with millimeter-wave radar systems. These radar systems have different models and they are all continuously transmitting radar signals.

[0080] The system first identifies that the signal emitted by the radar system of Vehicle A has a pseudo-random code PA, the signal emitted by the radar system of Vehicle B has a pseudo-random code PB, and the signal emitted by the radar system of the traffic signal has a pseudo-random code PS; the system obtains through querying the database or configuration file that the model of the radar system of Vehicle A is RadarA, the model of the radar system of Vehicle B is RadarB, and the model of the radar system of the traffic signal is RadarS; the system establishes an association relationship between the pseudo-random codes and the radar / transmitter model information; for example, it creates a data table that lists each pseudo-random code (PA, PB, PS) and its corresponding radar / transmitter model (RadarA, RadarB, RadarS).

[0081] Assume that later a new intelligent vehicle C joins the system and it is equipped with a millimeter-wave radar system with a model of RadarC; after the system receives the information that Vehicle C has joined, it dynamically updates the association relationship between the pseudo-random codes and the radar / transmitter model information; for example, it assigns a new pseudo-random code PC to the radar system of Vehicle C and adds a new record in the data table to associate PC with RadarC.

[0082] Therefore, determine the first separation parameter based on each pseudo-random code and the model of the millimeter-wave radar, and determine the second separation parameter according to each pseudo-random code and the models of multiple transmitters; determine the corresponding separation mechanism based on the first separation parameter, the second separation parameter, and the hardware accelerator, taking into account the overall consideration of the first separation parameter, the second separation parameter, and the hardware accelerator, ensuring the precise control of the separation mechanism;

[0083] At this time, determine two key separation parameters according to the pseudo-random code and the models of the millimeter-wave radar and the transmitter: the first separation parameter and the second separation parameter, which will be used in the subsequent signal separation and processing process.

[0084] First separation parameter: The system first analyzes the relationship between each pseudo-random code and the corresponding model of the millimeter-wave radar; according to the radar model, the system obtains some inherent characteristics of the radar, such as operating frequency, beam width, scanning mode, etc., which will be used as the first separation parameter to help the system distinguish signals from different radars at the receiving end.

[0085] Second separation parameter: The system analyzes the relationship between each pseudo-random code and the corresponding model of the transmitter; the model of the transmitter reveals its output power, modulation method, signal waveform and other characteristics, which are used as the second separation parameter to further help the system accurately identify the signal source at the receiving end.

[0086] After determining the first separation parameter and the second separation parameter, the system stores these parameters in an appropriate database or configuration file; with the addition of new radars and transmitters or the update of existing equipment, the system needs to dynamically update these parameters with the addition of new radars and transmitters or the update of existing equipment.

[0087] Combine the previously determined first and second separation parameters, and the available hardware accelerator, to determine the optimal signal separation mechanism; the system first evaluates the available hardware accelerators, such as FPGA (Field Programmable Gate Array), GPU (Graphics Processing Unit) or dedicated digital signal processing (DSP) chips; when selecting a hardware accelerator, the system will consider its processing power, power consumption, cost, and whether it supports the required signal processing algorithms.

[0088] According to the selected hardware accelerator and the known separation parameters, the system designs an efficient signal separation mechanism, which involves a variety of signal processing techniques. At the same time, the design goal of the signal separation mechanism is to minimize the interference between signals while ensuring the accurate identification and extraction of signals; at the same time, the system implements the designed separation mechanism on the selected hardware accelerator; then, verify the effectiveness of the mechanism through simulation or actual testing; according to the test results, the system needs to adjust and optimize the separation mechanism.

[0089] Specifically, assume that in an intelligent transportation monitoring system, there are two intelligent vehicles (Vehicle A and Vehicle B) equipped with different models of millimeter-wave radar systems (RadarA and RadarB), and these radar systems are continuously transmitting radar signals. At the same time, there is a traffic signal in the system, which is equipped with a transmitter (TransmitterS) for communicating with the vehicles.

[0090] The system first identifies that the signal transmitted by the radar system (RadarA) of Vehicle A has a pseudo-random code PA, and the operating frequency of RadarA is 24 GHz with a beam width of 30 degrees. The signal transmitted by the radar system (RadarB) of Vehicle B has a pseudo-random code PB, the operating frequency of RadarB is 77 GHz with a beam width of 10 degrees. The signal transmitted by the transmitter (TransmitterS) of the traffic signal has a pseudo-random code PS, and its output power is 10 W, using FSK modulation. Based on this information, the system determines the first separation parameters (such as operating frequency, beam width) and the second separation parameters (such as output power, modulation).

[0091] The system selects an FPGA as the hardware accelerator because it provides flexibility and high performance. According to the determined separation parameters and the capabilities of the FPGA, the system designs a signal separation mechanism, which includes filtering algorithms for the different operating frequencies and beam widths of RadarA and RadarB, and an FSK demodulation algorithm for TransmitterS. The system implements the designed separation mechanism on the FPGA and verifies its effectiveness through simulation tests.

[0092] In addition, assume the following weight assignment and score calculation: Weight assignment: operating frequency (0.4), beam width (0.3), output power (0.2), modulation (0.1), hardware accelerator compatibility (0.5, an additional consideration factor that does not directly correspond to the separation parameters but affects the final selection).

[0093] Score calculation (taking RadarA as an example): Operating frequency score: 0.4 * (normalized value corresponding to 24 GHz, assumed to be 0.8) = 0.32; Beam width score: 0.3 * (normalized value corresponding to 30°, assumed to be 0.6) = 0.18; Hardware accelerator compatibility score: 0.5 * (compatibility score between FPGA and RadarA, assumed to be 0.9) = 0.45; (Note: Output power and modulation are not applicable to RadarA, so the scores are 0); Total score: 0.32 + 0.18 + 0 + 0 + 0.45 = 0.95 (assuming scores for other inapplicable factors are 0).

[0094] Assume that scores are also calculated for RadarB and TransmitterS, and if the score of RadarA is the highest, the system will select the separation mechanism (band - pass filtering + beamforming) and the hardware accelerator (FPGA) corresponding to RadarA for processing.

[0095] In step S14, identity tags of each radar signal are determined based on each pseudo - random code, the hardware accelerator, and the separation mechanism, and each radar signal is separated.

[0096] In the specific implementation process of the present invention, the specific steps are as follows:

[0097] S141: Collect multiple working parameters of the hardware accelerator, and determine the signal separation level of the hardware accelerator according to the multiple working parameters of the hardware accelerator and the response model of the hardware accelerator.

[0098] S142: Associate each pseudo - random code, the signal separation level of the hardware accelerator, and the separation mechanism.

[0099] S143: Determine the first identity feature according to each pseudo - random code and the signal separation level of the hardware accelerator, and determine the second identity feature according to each pseudo - random code and the separation mechanism.

[0100] S144: Determine the identity tags of each radar signal based on the first identity feature, the second identity feature, and each pseudo - random code.

[0101] S145: Trigger the separation of each radar signal according to the identity tags of each radar signal.

[0102] In the embodiment of the present application, collecting multiple working parameters of the hardware accelerator and determining the signal separation level of the hardware accelerator according to the multiple working parameters of the hardware accelerator and the response model of the hardware accelerator is compatible with the multiple working parameters of the hardware accelerator. Considering the overall situation of the multiple working parameters of the hardware accelerator and the response model of the hardware accelerator ensures the accuracy of the signal separation level of the hardware accelerator.

[0103] At this time, the system will conduct comprehensive monitoring and data collection on the hardware accelerator; the working parameters of the hardware accelerator include but are not limited to:

[0104] Processing speed: The number of signals or amount of data that a hardware accelerator can process per second; Power consumption: The energy consumption of a hardware accelerator when performing tasks; Temperature: The operating temperature of a hardware accelerator, which reflects its heat dissipation performance and potential thermal stability issues; Resource utilization: The usage of internal resources (such as logic units, memory, cache, etc.) of a hardware accelerator; Error rate: The proportion of errors or anomalies that occur when a hardware accelerator processes signals; Latency: The processing time of a signal from input to output. These parameters are usually obtained through built-in sensors, monitoring tools, or specialized diagnostic software of the hardware accelerator.

[0105] After collecting the working parameters of the hardware accelerator, the system determines its signal separation level based on these parameters and the specific model of the hardware accelerator (i.e., its design and configuration); Since the parameter ranges of different hardware accelerators are different, the system first needs to normalize these parameters so that they can be compared on a unified scale; The system uses a pre-trained model or algorithm to evaluate the performance of the hardware accelerator. This model is trained based on a large amount of historical data and can accurately predict the signal separation ability of the hardware accelerator under different working conditions; Based on the evaluation results of the model, the system assigns a signal separation level to the hardware accelerator. This level is a quantitative value (such as a score from 1 to 10) and also a classification label (such as levels like "low", "medium", "high").

[0106] Specifically, assume that in an intelligent transportation monitoring system, there are two available hardware accelerators: one is a high-performance FPGA (Field Programmable Gate Array), and the other is a medium-performance DSP (Digital Signal Processor); The system needs to determine the signal separation levels of these two hardware accelerators in order to allocate resources for subsequent radar signal processing tasks.

[0107] For the FPGA, the system collects that its processing speed is 10,000 signals per second, power consumption is 50 watts, temperature is 60 degrees Celsius, resource utilization is 80%, error rate is 0.01%, and latency is 1 millisecond.

[0108] For the DSP, the system collects that its processing speed is 5,000 signals per second, power consumption is 30 watts, temperature is 50 degrees Celsius, resource utilization is 60%, error rate is 0.02%, and latency is 2 milliseconds;

[0109] The system normalizes these parameters so that their values are all in the range of 0 to 1. For example, the processing speed is normalized by dividing by the maximum processing speed, the power consumption is normalized by dividing by the maximum allowable power consumption, and so on; The system uses a model trained based on historical data to evaluate the performance of the FPGA and DSP. This model considers multiple factors such as processing speed, power consumption, temperature, resource utilization, error rate, and latency and gives a comprehensive score.

[0110] Based on the evaluation results of the models, the FPGA obtained a score of 9.5 (close to the full score of 10), and was determined to have a "high" level of signal separation ability; the DSP obtained a score of 6.8 and was determined to have a "medium" level of signal separation ability.

[0111] Furthermore, associate each pseudo-random code, the signal separation level of the hardware accelerator, and the separation mechanism; determine the first identity feature based on each pseudo-random code and the signal separation level of the hardware accelerator, determine the second identity feature based on each pseudo-random code and the separation mechanism, and introduce the first identity feature and the second identity feature.

[0112] At this time, establish a comprehensive association relationship, closely linking the unique identifier (pseudo-random code) of each radar signal, the signal separation level of the hardware accelerator responsible for processing the signal, and the signal separation mechanism actually applied; the pseudo-random code serves as a unique identity tag for the radar signal, ensuring that each signal can be accurately identified; the signal separation level reflects the efficiency level of the hardware accelerator when processing a specific signal, and this level is comprehensively evaluated based on the real-time working parameters and model of the hardware accelerator; the separation mechanism refers to the specific signal processing algorithm or process, which determines how the signal will be analyzed and processed; the system will create a data table or database that details the information of the hardware accelerator, signal separation level, and separation mechanism corresponding to each pseudo-random code.

[0113] In step S142, the system will integrate this information into a data table as follows:

[0114] Pseudo-random code Hardware accelerator Signal separation level Separation mechanism PA FPGA1 Advanced Beamforming PB DSP2 Intermediate Direction estimation

[0115] The system will further generate two unique identity features for each radar signal based on the previously established association relationship: the first identity feature and the second identity feature; the first identity feature: combines the pseudo-random code and the signal separation level of the hardware accelerator, and is used to provide a comprehensive identifier of the signal source and processing efficiency; the second identity feature: combines the pseudo-random code and the separation mechanism, mainly emphasizing the signal processing method. These identity features exist in the form of strings, codes, or hash values, aiming to ensure that each radar signal can be uniquely and accurately identified.

[0116] Optionally, for radar A (pseudo-random code PA): the first identity feature is "FPGA1 - Advanced - PA", which indicates that the signal comes from radar A and is processed by FPGA1 with a high level of signal separation ability; the second identity feature is "Beamforming - PA", which emphasizes that the signal is processed through the beamforming algorithm.

[0117] For radar B (pseudo-random code PB): The first identity feature is "DSP2 - Intermediate - PB", indicating that the signal comes from radar B and is processed by DSP2 with an intermediate-level signal separation ability; the second identity feature is "Direction Estimation - PB", indicating that the signal processing method is direction estimation.

[0118] Therefore, based on the first identity feature, the second identity feature, and each pseudo-random code, the identity markers of each radar signal are determined; according to the identity markers of each radar signal, the separation of each radar signal is triggered, achieving the precise separation of each radar signal.

[0119] At this time, the first identity feature, the second identity feature, and the original pseudo-random code generated previously are integrated to jointly form a comprehensive identity marker for uniquely identifying each radar signal. The design of this identity marker aims to ensure that the signal can be accurately tracked and identified during subsequent processing and analysis.

[0120] The system first merges the first identity feature, the second identity feature, and the pseudo-random code, which is achieved through simple string concatenation, encoding combination, or more complex data structures (such as hash tables or database records); after merging the features, the system performs uniqueness verification to ensure that each identity marker is unique, which is completed by checking whether the identity marker already exists in the system's identity marker database; if the identity marker passes the uniqueness verification, the system will formally generate and store it for subsequent use.

[0121] The system first checks the identity marker of each received signal and matches it with the markers stored in the identity marker database; once a matching identity marker is found, the system will allocate appropriate processing resources according to the information contained in the marker (such as the type of hardware accelerator, the signal separation level, and the separation mechanism); finally, the system will perform the signal separation task according to the allocated resources and the applied separation mechanism.

[0122] Specifically, consider a complex environment containing multiple radar systems, where each radar system uses a unique pseudo-random code to identify the signals it emits; the system has generated the first identity feature and the second identity feature for each signal.

[0123] For radar A (pseudo-random code PA): The first identity feature: FPGA1 - Advanced - PA; The second identity feature: Beamforming - PA.

[0124] For radar B (pseudo-random code PB): The first identity feature: DSP2 - Intermediate - PB; The second identity feature: Direction Estimation - PB.

[0125] The system now receives two radar signals and needs to trigger the separation process based on their identity tags; for the first signal, its identity tag is "ID-FPGA1-High-Level-Beamforming-PA", and the system recognizes that this is a signal from Radar A, which should be processed by FPGA1 with high-level signal separation capabilities and beamforming algorithms; for the second signal, its identity tag is "ID-DSP2-Medium-Direction Estimation-PB", and the system determines that this is a signal from Radar B, which should be processed by DSP2 with medium-level signal separation capabilities and direction estimation algorithms.

[0126] In step S15, based on the sensing range of the millimeter-wave radar, the imaging range of the intelligent vehicle, and each signal reception path, determine the early warning range of the intelligent vehicle during autonomous driving;

[0127] In the specific implementation process of the present invention, the specific steps are as follows:

[0128] S151: Obtain the sensing range of the millimeter-wave radar;

[0129] S152: Based on the model of the camera on the intelligent vehicle, the environment where the millimeter-wave radar is located, and the distribution positions of each camera, determine the imaging range of the intelligent vehicle;

[0130] S153: Collect each signal reception path;

[0131] S154: Interact with the sensing range of the millimeter-wave radar, the imaging range of the intelligent vehicle, and each signal reception path;

[0132] S155: Determine the early warning range of the intelligent vehicle according to the interaction of the sensing range of the millimeter-wave radar, the imaging range of the intelligent vehicle, and each signal reception path.

[0133] In the embodiment of the present application, obtain the sensing range of the millimeter-wave radar; based on the model of the camera on the intelligent vehicle, the environment where the millimeter-wave radar is located, and the distribution positions of each camera, determine the imaging range of the intelligent vehicle, introduce the sensing range of the millimeter-wave radar and the imaging range of the intelligent vehicle, and control the sensing range of the millimeter-wave radar and the imaging range of the intelligent vehicle.

[0134] At this time, obtain the accurate sensing range of the millimeter-wave radar, and comprehensively consider the model of the camera on the intelligent vehicle, the specific environment where the millimeter-wave radar is located, and the distribution position of the camera to determine the imaging range; the imaging range generally refers to the area where the camera can clearly capture images, and this area is affected by factors such as the camera focal length, field of view, resolution, and environmental factors (such as lighting conditions, obstacles).

[0135] To determine the camera range, the system takes the following measures: The system analyzes the model of the camera installed on the intelligent vehicle to understand its key parameters such as focal length, field of view, and resolution, which are usually found in the camera's specification sheet; The system evaluates the environment where the millimeter-wave radar and the camera are located, including lighting conditions (such as day, night, backlight), obstacles (such as trees, buildings), weather conditions (such as rain, snow, fog), etc. These factors all affect the camera's range and image quality; Finally, the system considers the specific positions of each camera on the vehicle and their relative relationships; The positions of the cameras should ensure that they can cover key areas around the vehicle, such as the front road, side blind spots, rear view, etc.; At the same time, the overlapping area between the cameras should be minimized to avoid redundant information.

[0136] Specifically, assume that four cameras are installed on the intelligent vehicle, located at the front, both sides, and rear of the vehicle respectively; Among them, the front camera is a wide-angle lens with a field of view of 180 degrees and a resolution of 1080P; The cameras on both sides are standard lenses with a field of view of 90 degrees and a resolution of 720P; The rear camera is a narrow-angle lens with a field of view of 45 degrees, but has higher resolution and zoom capabilities; Considering that the vehicle will travel under various environmental conditions, the system dynamically adjusts the camera range; For example, at night or in low-light conditions, the system will enable the camera's night vision mode or increase the exposure time to improve image brightness; In rainy or snowy weather, the system will apply image de-fogging or de-raining algorithms to reduce the impact of weather on image quality; At the same time, the system will also optimize the overlapping area of the camera range according to the distribution positions and relative relationships of the cameras to ensure that key areas around the vehicle can be clearly captured.

[0137] Therefore, each signal reception path is collected; The sensing range of the millimeter-wave radar, the camera range of the intelligent vehicle, and each signal reception path are interacted; The warning range of the intelligent vehicle is determined based on the interaction of the sensing range of the millimeter-wave radar, the camera range of the intelligent vehicle, and each signal reception path, realizing the interaction of the sensing range of the millimeter-wave radar, the camera range of the intelligent vehicle, and each signal reception path, and ensuring the accuracy of the warning range of the intelligent vehicle;

[0138] At this time, the sensing range of the millimeter-wave radar, the camera range of the intelligent vehicle, and the reception paths of each signal (including radar signals and camera image signals) are interacted and integrated, which usually involves data communication and synchronization between multiple sensor systems to ensure that the information they provide can complement and verify each other.

[0139] To achieve this goal, the system will take the following measures: The system will ensure the temporal synchronization of the data from the millimeter-wave radar and the camera, which is achieved by using a common clock signal or timestamps to ensure that the objects detected by the radar are consistent in time with the images captured by the camera; The system will perform spatial calibration to determine the relative positions of the millimeter-wave radar and the camera in space, which typically involves using known physical reference objects or calibration tools to measure the relative angles and distances between the radar and the camera; Based on these measurement values, the system establishes a unified spatial coordinate system to convert the object positions detected by the radar and the pixel positions of the images captured by the camera with each other.

[0140] Based on data synchronization and spatial calibration, the system will perform data fusion, which typically involves integrating data from different sensors to generate a more comprehensive and accurate environmental perception model; For example, the system will match the object positions detected by the radar with the images captured by the camera to generate a comprehensive environmental perception map containing information such as object positions, speeds, types, etc.; In addition to the direct data from the radar and the camera, the system will also analyze the reception paths of each signal, which includes considering factors such as obstacles, reflections, and attenuations encountered by the signal during propagation; By analyzing these paths, the system further optimizes the accuracy of the environmental perception model.

[0141] The system will determine the warning range of the intelligent vehicle based on the environmental perception model integrated in the previous steps; The warning range generally refers to the area where the intelligent vehicle can perceive in advance and make responses to ensure driving safety; To determine the warning range, the system will consider the following factors:.

[0142] The system will determine which objects pose potential threats to the intelligent vehicle based on the distance and speed information of the obstacles detected by the radar; For example, if there is a vehicle approaching rapidly ahead, the system will include it in the warning range.

[0143] The system will also consider the type and size of the obstacles; Different types of objects (such as pedestrians, bicycles, trucks, etc.) pose different threats to the intelligent vehicle; At the same time, the size of the object will also affect its visibility and detectability in the environment.

[0144] The system will also analyze the reception path quality of the radar signal and the camera image signal; If the signal encounters severe attenuation or occlusion during propagation, the system will adjust the warning range accordingly to ensure sufficient perception ability under adverse conditions.

[0145] Finally, the system will also consider a certain safety margin, which means that the warning range will be slightly larger than the actual required range to ensure that the intelligent vehicle has sufficient time and space to make responses in case of emergencies.

[0146] Specifically, assume that an intelligent vehicle is driving on a highway and there is a large truck ahead; the millimeter-wave radar first detects the presence of an object ahead and provides information on the distance and speed of the object; at the same time, the front camera also captures an image of the truck and provides more detailed information on the object type, size, and shape; the system integrates this information into a unified environmental perception model through data synchronization and spatial calibration; in addition, the system also analyzes the reception paths of the radar signals and camera image signals, considering factors such as ground reflection and atmospheric attenuation that affect the signal quality; finally, the system generates a comprehensive environmental perception map containing information such as the position, speed, type, and size of the truck, providing strong support for subsequent warnings and decision-making.

[0147] The intelligent vehicle is driving on a highway and there is a large truck ahead; based on the sensing range of the millimeter-wave radar, the imaging range of the intelligent vehicle, and the reception path quality of each signal, the system determines a warning range that includes a certain area in front of the truck to ensure that the intelligent vehicle can respond in a timely manner when the truck suddenly decelerates or stops; at the same time, considering the situation where the truck blocks the vehicles behind, the system also expands the warning range at the side and rear to detect vehicles approaching from the side or rear of the truck; through these measures, the system ensures that the intelligent vehicle can maintain a high level of perception ability and driving safety in a complex road environment.

[0148] In the intelligent vehicle system, to determine the warning range, a warning range matching table is used to integrate the information of the millimeter-wave radar, camera, and signal reception path; the matching table is a two-dimensional or multi-dimensional table used to correlate the data of different sensors with each other and determine the warning range based on this correlated information.

[0149] Example of a warning range matching table:

[0150] Radar induction range (meters) Camera range (degrees) Signal reception path quality Warning range (meters / degrees) 200 180° (front) Excellent 180 meters in front, ±30° on the side 150 90° (side) Good 120 meters on the side, ±15° deflection angle 80 45° (rear) Medium 60 meters in the rear, ±20° deflection angle

[0151] In this example, the warning range matching table determines the warning range based on the sensing range of the radar, the field of view of the camera, and the quality of the signal reception path; for example, if the radar senses an object within 200 meters ahead and the field of view of the front camera is 180 degrees and the signal reception path quality is "excellent", the system will set the warning range to 180 meters ahead and consider a ±30-degree side range to ensure early perception when the object approaches.

[0152] In step S16, according to the warning range of the intelligent vehicle, the number of radar signals, and the service life of the intelligent vehicle, the hardware accelerator is triggered to autonomously optimize the signal separation level of each radar signal.

[0153] In the specific implementation process of the present invention, the specific steps are as follows:

[0154] S161: Obtain the warning range of the intelligent vehicle;

[0155] S162: Determine multiple warning features based on the warning range of the intelligent vehicle and the position where the intelligent vehicle is located;

[0156] S163: Interact with multiple warning features, the number of radar signals, and the service life of the intelligent vehicle; Determine the first signal separation parameter according to multiple warning features and the number of radar signals; Determine the second signal separation parameter according to multiple warning features and the service life of the intelligent vehicle;

[0157] S164: Determine the optimized separation level of the hardware accelerator based on the first separation parameter and the second separation parameter, and trigger the autonomous optimization of the signal separation level of each radar signal by the hardware accelerator according to the optimized separation level of the hardware accelerator.

[0158] In the embodiment of the present application, obtain the warning range of the intelligent vehicle; Determine multiple warning features according to the warning range of the intelligent vehicle and the position where the intelligent vehicle is located, and accurately control multiple warning features.

[0159] At this time, obtain the warning range of the intelligent vehicle, and determine multiple warning features based on the warning range of the intelligent vehicle and the current position of the vehicle. These warning features are crucial for the safe driving of the intelligent vehicle because they can help the vehicle perceive and respond to potential dangerous situations in advance; The following is a detailed explanation of this step.

[0160] The system first obtains the warning range of the intelligent vehicle, which usually includes the distance and angle information in the front, side, and rear; The warning range is comprehensively obtained based on the data of sensors such as millimeter-wave radars and cameras, and is used to indicate the area where potential dangers exist around the intelligent vehicle; Then, the system determines the current position of the intelligent vehicle, including its specific position on the road, driving direction, speed, etc. These information are usually obtained through GPS, map matching, and the vehicle's own sensors; The system analyzes the objects within the warning range, including their types (such as pedestrians, bicycles, cars, etc.), quantity, distance, speed, and movement direction, etc. These information are mainly obtained through camera image recognition and millimeter-wave radar detection.

[0161] The system determines multiple warning features, which include: distance feature: the relative distance between the intelligent vehicle and the object within the warning range; angle feature: the direction or angle of the object relative to the intelligent vehicle; speed feature: the moving speed of the object within the warning range; acceleration feature: the acceleration of the object, used to predict its future motion trajectory; type feature: the type of the object, which is usually obtained through camera image recognition; path feature: the driving path or motion trajectory of the object, predicted based on its current speed, acceleration and other information; priority feature: according to factors such as the type, distance, and speed of the object, determine the priority of different objects within the warning range, so that the intelligent vehicle can give priority to paying attention to and dealing with more dangerous objects.

[0162] Specifically, assume that the intelligent vehicle is driving at a complex urban intersection. The current position of the vehicle is in the left lane near the intersection and is preparing to turn left; at this time, the warning range obtained by the system includes the area 50 meters ahead, within 30 degrees to the side, and 20 meters behind.

[0163] Obtain the warning range: 50 meters ahead, 30 degrees to the side, 20 meters behind.

[0164] Determine the vehicle position: The vehicle is in the left lane of the intersection, preparing to turn left, and the current speed is 30 km / h; Analyze the objects within the warning range: Ahead: A straight-going car, 30 meters away, with a speed of 40 km / h; To the side: A pedestrian, 10 meters away, is crossing the road; Behind: A motorcycle following closely, 5 meters away, with a speed of 35 km / h.

[0165] Determine the warning features: Distance feature: The car ahead is 30 meters away, the pedestrian to the side is 10 meters away, and the motorcycle behind is 5 meters away; Angle feature: The pedestrian to the side is located about 15 degrees to the left of the vehicle; Speed feature: The car ahead is 40 km / h, the motorcycle behind is 35 km / h (the vehicle's own speed is 30 km / h); Acceleration feature: The current acceleration of all objects is close to 0, indicating that they are all moving or walking at a constant speed; Type feature: Ahead is a car, to the side is a pedestrian, and behind is a motorcycle; Path feature: The car ahead will go straight through the intersection, the pedestrian to the side will cross the road and enter the left lane, and the motorcycle behind will follow the vehicle; Priority feature: The pedestrian to the side has the highest priority because pedestrians are usually the most vulnerable traffic participants, and their behavior of crossing the road causes sudden danger; The car ahead and the motorcycle behind are next; Based on these warning features, the intelligent vehicle makes corresponding decisions, such as decelerating, stopping, or steering to avoid, to ensure safe passage through the intersection.

[0166] Furthermore, interactions are carried out among multiple warning features, the number of radar signals, and the service life of the intelligent vehicle; a first signal separation parameter is determined based on the multiple warning features and the number of radar signals; a second signal separation parameter is determined based on the multiple warning features and the service life of the intelligent vehicle, and the first signal separation parameter and the second signal separation parameter are introduced.

[0167] At this time, the system comprehensively considers multiple warning features, the number of radar signals, and the service life of the intelligent vehicle to determine two key signal separation parameters: the first signal separation parameter and the second signal separation parameter, which are crucial for the intelligent vehicle to effectively process and distinguish radar signals from different objects.

[0168] Interaction between multiple warning features and the number of radar signals: The system first analyzes multiple warning features, such as the distance, speed, type of objects, etc., and the number of currently received radar signals; the number of radar signals refers to the total number of radar signals received by the intelligent vehicle from different objects at the same time; by analyzing the warning features and the number of radar signals, the system evaluates the complexity of the current environment; for example, if there are many objects with diverse types within the warning range and the number of radar signals is large, then the environment is considered complex.

[0169] The first signal separation parameter is mainly determined based on warning features and the number of radar signals; its goal is to ensure that the system can accurately distinguish radar signals from different objects, especially in complex environments with signal overlap or interference; the first signal separation parameter includes the resolution of signal processing, the bandwidth and shape of the filter, and the complexity of the signal separation algorithm, etc.; in a complex environment, the system requires higher resolution and more complex algorithms to accurately separate signals.

[0170] Interaction between multiple warning features and the service life of the intelligent vehicle: In addition to warning features and the number of radar signals, the system also considers the service life of the intelligent vehicle; as the service life increases, the vehicle's sensors (including radars) will gradually age, resulting in performance degradation, such as reduced signal quality and increased noise, etc.; by analyzing the warning features and the vehicle's service life, the system evaluates the degree of sensor performance degradation and adjusts the signal separation strategy accordingly.

[0171] The second signal separation parameter is mainly determined based on warning features and the vehicle's service life; its goal is to compensate for the degradation of sensor performance and ensure that even in the case of sensor aging, the system can still accurately separate and process radar signals; the parameters include the adjustment of signal gain, the optimization of noise suppression strategies, and the adaptive adjustment of signal processing algorithms, etc.; as the service life increases, the system needs to increase signal gain, adopt more advanced noise suppression technologies, or adjust the algorithm to adapt to changes in sensor performance.

[0172] Specifically, assume that the intelligent vehicle is traveling on a busy highway, and at this time, the system receives radar signals from multiple objects.

[0173] The warning features include: there are two cars ahead, with distances of 50 meters and 100 meters respectively, and speeds of 120 km / h; there is a truck on the side, with a distance of 30 meters and a speed of 100 km / h; there is a motorcycle behind, with a distance of 10 meters and a speed of 130 km / h; the number of radar signals: the system simultaneously receives radar signals from these four objects; the complexity of the environment: since there are two cars with similar speeds ahead, and there are also objects on the side and behind, the environment is considered complex.

[0174] Determine the first signal separation parameter: The system decides to adopt a high-resolution signal processing algorithm to accurately distinguish radar signals from different objects; at the same time, the system also selects a filter with a narrow bandwidth to reduce interference between signals; the first signal separation parameter includes: high-resolution signal processing algorithm, narrow-bandwidth filter.

[0175] Interaction between multiple warning features and the service life of the intelligent vehicle: The intelligent vehicle has been used for 5 years, and the performance of the radar sensor has degraded, manifested as a slight decrease in signal quality and a slight increase in noise; the system evaluates the degree of degradation of the sensor performance by analyzing the warning features and the vehicle service life;

[0176] Determine the second signal separation parameter: The system decides to increase the signal gain to compensate for the degradation of the radar sensor performance; at the same time, the system also adopts a more advanced noise suppression technology to reduce the impact of noise on the signal; the second signal separation parameter includes: increased signal gain, advanced noise suppression technology; by comprehensively considering multiple warning features, the number of radar signals, and the service life of the intelligent vehicle, the system determines the first signal separation parameter and the second signal separation parameter, which ensure that the intelligent vehicle can still accurately separate and process radar signals even in complex environments and with sensor aging, thus making correct decisions and actions.

[0177] Therefore, based on the first separation parameter and the second separation parameter, the optimized separation level of the hardware accelerator is determined, and according to the optimized separation level of the hardware accelerator, the hardware accelerator is triggered to autonomously optimize the signal separation level of each radar signal, realizing the autonomous optimization of the signal separation level of each radar signal by the hardware accelerator, ensuring the control of the optimized separation level of the hardware accelerator, and improving the accurate separation of multiple radar signals in different environments.

[0178] At this time, based on the first signal separation parameter and the second signal separation parameter determined previously, the system further determines the optimized signal separation level of the hardware accelerator and triggers the hardware accelerator to autonomously adjust the separation level of each radar signal.

[0179] The system first determines the optimization level of the hardware accelerator based on the first signal separation parameter (mainly based on warning features and the number of radar signals) and the second signal separation parameter (mainly based on warning features and the vehicle usage years); the optimization level reflects the performance requirements of the hardware accelerator when processing radar signals; in complex environments, or when the sensor performance degrades, the system requires a higher optimization level to ensure the accuracy of signal separation.

[0180] Once the optimization level is determined, the system sends an instruction to the hardware accelerator to trigger its autonomous optimization process; the hardware accelerator is a hardware component specifically designed to accelerate specific computing tasks, such as FPGA (Field Programmable Gate Array) or ASIC (Application Specific Integrated Circuit); in this scenario, the hardware accelerator is used to accelerate the separation and processing of radar signals; the autonomous optimization process includes adjusting the internal configuration of the hardware accelerator, optimizing the signal processing algorithm, reallocating computing resources, etc., to adapt to the current signal separation requirements.

[0181] During the autonomous optimization process, the hardware accelerator adjusts the separation level of each radar signal according to the optimization level; the signal separation level determines the fineness and computing resource investment of the hardware accelerator when processing each radar signal; a higher separation level means finer signal processing and more computing resource investment, but also brings higher energy consumption and cost; therefore, the system needs to balance the computing resource investment and energy consumption cost while ensuring the accuracy of signal separation.

[0182] Specifically, assume that an intelligent vehicle is driving in a complex urban environment. At this time, the system receives radar signals from multiple objects, and the vehicle has been in use for several years, and the performance of the radar sensor has degraded; the system analyzes the warning features (such as the distance, speed, type, etc. of the object) and the number of radar signals, and finds that the environment is complex and the performance of the radar sensor has degraded due to the long vehicle usage years; based on this information, the system determines a higher optimization level to ensure that the radar signals can still be accurately separated and processed in complex environments and when the sensor performance degrades; the system sends an instruction to the hardware accelerator to trigger its autonomous optimization process; in this example, the hardware accelerator is an FPGA; the FPGA starts to adjust its internal configuration, optimize the signal processing algorithm, and reallocate computing resources to adapt to the current signal separation requirements.

[0183] During the autonomous optimization process, the FPGA adjusts the separation level of each radar signal according to the optimization level; for fast-moving objects (such as pedestrians or bicycles) at close range in the front, the FPGA assigns a higher separation level to ensure that the subtle motion characteristics of these objects can be accurately captured; for distant objects from the side or rear, since they have less direct impact on vehicle driving safety, the FPGA assigns a lower separation level to save computing resources and energy consumption.

[0184] In another embodiment of the present application, for the hardware accelerator, the hardware accelerator is used to achieve real-time separation of signals from different transmitters, avoiding the use of circuits with low precision and difficult integration such as analog phase shifters.

[0185] The radar architecture using a hardware accelerator for pseudo-random code reception phase demodulation is as Figure 1 shown. A pseudo-random code generator generates N groups of mutually orthogonal pseudo-random codes. Each group of pseudo-random codes controls the phase modulator in one transmitter according to a fixed timing sequence; N transmitters simultaneously transmit radar signals, and each group of radar signals is phase-modulated by a different pseudo-random code controlling the phase modulator; the N groups of mutually orthogonal pseudo-random codes generated by the pseudo-random code generator are simultaneously input into the hardware accelerator.

[0186] The millimeter-wave radar has M receivers. Each receiver simultaneously receives radar signals from N transmitters. In each receiver, the received radar signals are amplified, demodulated, filtered, and sampled, and then analog-to-digital conversion is performed to form sampled intermediate-frequency signals; the sampled intermediate-frequency signals are sent to the signal processing hardware accelerator, and signal processing is performed in this hardware accelerator to output the velocity-distance two-dimensional spectra of the signals of the N transmitters received by this receiver respectively; thus, N transmitters simultaneously transmit and M receivers simultaneously receive, totaling N*M radar transceiver signals.

[0187] The signal generated by the voltage-controlled oscillator is LO. The transmitted signal after the LO signal modulates the phase according to pseudo-random code 1 in transmitter 1 is TX1, the transmitted signal after the LO signal modulates the phase according to pseudo-random code 2 in transmitter 2 is TX2, and the transmitted signal after the LO signal modulates the phase according to pseudo-random code N in transmitter N is TXN. These transmitted signals TX1, TX2, ……, TXN are reflected back after hitting an obstacle. Each receiver (receiver 1, receiver 2, ……, receiver M) simultaneously receives N transmitted signals. Without loss of generality, taking receiver 1 as an example, receiver 1 receives the transmitted signals RX1(TX1, TX2, ……, TXN) from N transmitters. This signal is input into the signal processing hardware accelerator after being amplified, demodulated, filtered, and sampled.

[0188] The signal processing algorithm flow of the hardware accelerator is as Figure 4As shown in the figure, the intermediate frequency sampling signal of each receiving channel is windowed and then subjected to range dimension FFT to extract range information, obtaining the one-dimensional range spectrum of the signal. Then, the one-dimensional range spectrum is windowed N times and subjected to N times of Doppler dimension FFT to obtain the range-Doppler two-dimensional spectrum of the signals received by the receiver from N transmitters. Among them, the phases of the windowing coefficients W21...W2N are equal to or differ by a fixed value from the phases corresponding to the N pseudo-random codes of the N transmitters. Thus, the signals of the N transmitters are separated, enabling N transmitters to transmit simultaneously and M receivers to receive simultaneously, for a total of N*M radar transceiver signals.

[0189] The hardware accelerator structure required for the corresponding signal processing flow is as Figure 5 shown, and it consists of a digital multiplier, a windowing system memory, and an FFT hardware accelerator. Among them, the windowing coefficient memory stores the current signal windowing coefficients W1, or W21, W22,..., W2N.

[0190] Considering the parallel processing of the signal windowing W21 or W22 or... or W2N required for separating the signals from N transmit channels by M receiving channels and the Doppler dimension FFT operation, one or more hardware accelerators are integrated in the chip to improve the operation speed through parallel processing.

[0191] With the in-depth development of autonomous driving technology, the number of transceiver channels of millimeter-wave radars will increase. Multiple transmit channels transmitting simultaneously increases the detection range of the radar. However, the isolation requirements achieved through pseudo-random code phase modulation between multiple transmit channels will become higher and higher. The present invention provides a highly efficient, highly integrated, parallel, and pure digital low-cost solution, avoiding the use of complex analog circuits and inefficient and high-cost DSPs and CPUs.

[0192] Embodiment III

[0193] Please refer to Figure 3 , Figure 3 which is a schematic structural composition diagram of a system for separating multiple radar signals based on a hardware accelerator in an embodiment of the present invention.

[0194] As Figure 3 shown, a system for separating multiple radar signals based on a hardware accelerator, the system for separating multiple radar signals based on a hardware accelerator includes:

[0195] A radar signal module 21, configured to, when the millimeter-wave radar moves with the movement of the intelligent vehicle, receive multiple radar signals emitted by each transmitter based on the millimeter-wave radar.

[0196] A pseudo-random code module 22, configured to determine a signal reception path according to the millimeter-wave radar and each transmitter, and determine a corresponding pseudo-random code according to the multiple radar signals, the corresponding signal reception path, and a pseudo-random code generator.

[0197] A separation module 23, configured to determine a corresponding separation mechanism based on each pseudo-random code, the model of the millimeter-wave radar, and the models of multiple transmitters;

[0198] An identity marking module 24, configured to determine the identity markings of each radar signal according to each pseudo-random code, the hardware accelerator, and the separation mechanism, and separate each radar signal;

[0199] A warning range module 25, configured to determine the warning range of the intelligent vehicle during autonomous driving based on the sensing range of the millimeter-wave radar, the imaging range of the intelligent vehicle, and each signal reception path;

[0200] An optimization module 26, configured to trigger the hardware accelerator to autonomously optimize the signal separation level of each radar signal according to the warning range of the intelligent vehicle, the number of radar signals, and the service life of the intelligent vehicle.

[0201] For any combination of the technical features of the above embodiments, for the sake of brevity of description, not all combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

Claims

1. A method for separating multiple radar signals based on a hardware accelerator, characterized in that, Including: When the millimeter-wave radar moves with the intelligent vehicle, receiving a plurality of radar signals emitted by each transmitter based on the millimeter-wave radar; Determining a signal reception path according to the millimeter-wave radar and each transmitter, and determining a corresponding pseudo-random code according to the plurality of radar signals, the corresponding signal reception path, and a pseudo-random code generator; Determining a corresponding separation mechanism based on each pseudo-random code, the model of the millimeter-wave radar, and the models of the plurality of transmitters; Determining an identity tag for each radar signal according to each pseudo-random code, a hardware accelerator, and a separation mechanism, and separating each radar signal; Determining a warning range of the intelligent vehicle during autonomous driving based on the sensing range of the millimeter-wave radar, the imaging range of the intelligent vehicle, and each signal reception path; Triggering the hardware accelerator to autonomously optimize the signal separation level of each radar signal according to the warning range of the intelligent vehicle, the number of radar signals, and the service life of the intelligent vehicle.

2. The method for separating a plurality of radar signals based on a hardware accelerator according to claim 1, wherein The step of, when the millimeter-wave radar moves with the intelligent vehicle, receiving a plurality of radar signals emitted by each transmitter based on the millimeter-wave radar, includes: The millimeter-wave radar is installed on the intelligent vehicle, moves with the intelligent vehicle, and enters various environments; Determining the sensing range of the millimeter-wave radar according to the position of the millimeter-wave radar, the moving speed of the millimeter-wave radar, and the signal interference level of the environment; Responding to corresponding multiple transmitters based on the position of the millimeter-wave radar and the sensing range of the millimeter-wave radar, and the multiple transmitters emit multiple radar signals in the response state, and at this time, the multiple radar signals are received by the millimeter-wave radar.

3. The method for separating a plurality of radar signals based on a hardware accelerator according to claim 2, wherein The step of determining a signal reception path according to the millimeter-wave radar and each transmitter, and determining a corresponding pseudo-random code according to the plurality of radar signals, the corresponding signal reception path, and a pseudo-random code generator, includes: Interacting between the millimeter-wave radar and each transmitter; Determining a signal reception path according to the interaction between the millimeter-wave radar and each transmitter; Determining a pseudo-random code generator based on the matching of the millimeter-wave radar, each transmitter, and the intelligent vehicle; Interacting between the plurality of radar signals, the corresponding signal reception path, and the pseudo-random code generator; Determining a corresponding pseudo-random code according to the interaction between the plurality of radar signals, the corresponding signal reception path, and the pseudo-random code generator, and at this time, each pseudo-random code changes dynamically with the change of the usage time of the pseudo-random code generator.

4. The method for separating a plurality of radar signals based on a hardware accelerator according to claim 3, wherein The step of determining a corresponding separation mechanism based on each pseudo-random code, the model of the millimeter-wave radar, and the models of the plurality of transmitters, includes: Obtaining each pseudo-random code; Interacting between each pseudo-random code, the model of the millimeter-wave radar, and the models of the plurality of transmitters; Determining a first separation parameter based on each pseudo-random code and the model of the millimeter-wave radar, and determining a second separation parameter based on each pseudo-random code and the models of the plurality of transmitters; Determining a corresponding separation mechanism according to the first separation parameter, the second separation parameter, and the hardware accelerator.

5. The method for separating a plurality of radar signals based on a hardware accelerator according to any one of claims 1 to 4, characterized in that The step of determining an identity tag for each radar signal according to each pseudo-random code, a hardware accelerator, and a separation mechanism, and separating each radar signal, includes: Collect multiple working parameters of the hardware accelerator, and determine the signal separation level of the hardware accelerator according to the multiple working parameters of the hardware accelerator and the response model of the hardware accelerator; Associate each pseudo-random code, the signal separation level of the hardware accelerator, and the separation mechanism; Determine the first identity feature according to each pseudo-random code and the signal separation level of the hardware accelerator, and determine the second identity feature according to each pseudo-random code and the separation mechanism; Determine the identity mark of each radar signal based on the first identity feature, the second identity feature, and each pseudo-random code; Trigger the separation of each radar signal according to the identity mark of each radar signal.

6. The method for separating multiple radar signals based on a hardware accelerator according to claim 5, wherein The method for determining the warning range of the intelligent vehicle during autonomous driving based on the sensing range of the millimeter-wave radar, the imaging range of the intelligent vehicle, and each signal reception path includes: Obtain the sensing range of the millimeter-wave radar; Determine the imaging range of the intelligent vehicle based on the model of the camera of the intelligent vehicle, the environment where the millimeter-wave radar is located, and the distribution positions of each camera.

7. The method for separating multiple radar signals based on a hardware accelerator according to claim 6, wherein The method for determining the warning range of the intelligent vehicle during autonomous driving based on the sensing range of the millimeter-wave radar, the imaging range of the intelligent vehicle, and each signal reception path further includes: Collect each signal reception path; Interact the sensing range of the millimeter-wave radar, the imaging range of the intelligent vehicle, and each signal reception path; Determine the warning range of the intelligent vehicle according to the interaction of the sensing range of the millimeter-wave radar, the imaging range of the intelligent vehicle, and each signal reception path.

8. The method for separating a plurality of radar signals based on a hardware accelerator according to claim 7, wherein The method for triggering the autonomous optimization of the signal separation level of each radar signal by the hardware accelerator according to the warning range of the intelligent vehicle, the number of radar signals, and the service life of the intelligent vehicle includes: Obtain the warning range of the intelligent vehicle; Determine multiple warning features according to the warning range of the intelligent vehicle and the position where the intelligent vehicle is located.

9. The method for separating multiple radar signals based on a hardware accelerator according to claim 8, wherein The method for triggering the autonomous optimization of the signal separation level of each radar signal by the hardware accelerator according to the warning range of the intelligent vehicle, the number of radar signals, and the service life of the intelligent vehicle further includes: Interact multiple warning features, the number of radar signals, and the service life of the intelligent vehicle; determine the first signal separation parameter according to multiple warning features and the number of radar signals; determine the second signal separation parameter according to multiple warning features and the service life of the intelligent vehicle; Determine the optimized separation level of the hardware accelerator based on the first separation parameter and the second separation parameter, and trigger the autonomous optimization of the signal separation level of each radar signal by the hardware accelerator according to the optimized separation level of the hardware accelerator.

10. A separation system for multiple radar signals based on a hardware accelerator, characterized in that, The separation system for multiple radar signals based on the hardware accelerator is applied to the method for separating multiple radar signals based on the hardware accelerator as described in any one of claims 1-9. The separation system for multiple radar signals based on the hardware accelerator includes: A radar signal module, configured to receive multiple radar signals emitted by each transmitter based on the millimeter-wave radar when the millimeter-wave radar moves with the movement of the intelligent vehicle; A pseudo-random code module, which is used to determine the signal reception path according to the millimeter-wave radar and each transmitter, and determine the corresponding pseudo-random code according to multiple radar signals, the corresponding signal reception path, and a pseudo-random code generator; A separation module, which is used to determine the corresponding separation mechanism based on each pseudo-random code, the model of the millimeter-wave radar, and the models of multiple transmitters; An identity marking module, which is used to determine the identity marking of each radar signal according to each pseudo-random code, the hardware accelerator, and the separation mechanism, and separate each radar signal; A warning range module, which is used to determine the warning range of the intelligent vehicle during autonomous driving based on the sensing range of the millimeter-wave radar, the camera range of the intelligent vehicle, and each signal reception path; An optimization module, which is used to trigger the hardware accelerator to autonomously optimize the signal separation level of each radar signal according to the warning range of the intelligent vehicle, the number of radar signals, and the service life of the intelligent vehicle.

Images