A fusion perception method and device of corner radar and computer readable storage medium

Abstract

The application relates to a fusion perception method of corner radars, a fusion perception device and a computer readable storage medium. The fusion perception method comprises the following steps: controlling each corner radar to process a return signal to obtain original point cloud data, and obtaining original point cloud data of left and right corresponding side corner radars; a left front corner radar and a left rear corner radar execute a moving target tracking algorithm according to the original point cloud data to obtain first target point information, and a right front corner radar and a right rear corner radar synchronously execute a static target passable area algorithm to obtain point trail information; the left front corner radar and the left rear corner radar mutually transmit the first target point information, and execute an algorithm again to obtain second target point information; finally, the three types of information are sent to a vehicle ADAS domain controller. The application can improve the real-time performance of perception and effectively reduce the system cost.

Description

Technical Field

[0001] This invention relates to the field of image processing technology for autonomous driving, and in particular to a fusion sensing method, fusion sensing device, and computer-readable storage medium for corner radar. Background Technology

[0002] With the rapid development of Advanced Driver Assistance Systems (ADAS) and autonomous driving technologies, vehicles face extremely stringent requirements regarding the comprehensiveness, real-time performance, and reliability of their environmental perception. Vehicle perception systems not only need to accurately detect and track dynamic targets on the road (such as vehicles and pedestrians), but also need to accurately identify static drivable areas (freespace) to provide core data for path planning and safety control.

[0003] In existing technologies, vehicle-mounted corner radars are typically used only for single functions such as blind spot detection (BSD), lane change assist (LCA), forward / rear cross-traffic alert and assist (FCTA / FCTB, RCTA / RCTB), and door opening warning (DOW), and the value of their perception data has not been fully explored. To achieve comprehensive environmental perception that combines dynamic target tracking and static drivable area detection, the industry mainly adopts two types of technical solutions, but both have significant drawbacks: Firstly, there is the full-function processing scheme based on a single radar. This scheme requires each corner radar in the system to independently complete all perception algorithms, that is, running moving target tracking and drivable area detection algorithms simultaneously on the same radar processing chip. However, both of these algorithms are computationally intensive tasks, and the embedded platform hardware resources of corner radars (such as CPU computing power and memory bandwidth) are limited. Executing two tasks simultaneously will lead to intense competition for computing resources. This directly results in a significant increase in the data processing frame cycle and a decrease in the output frequency of perception results, which cannot meet the stringent real-time requirements of perception in high-speed dynamic driving scenarios. Processing latency has become the core bottleneck restricting the improvement of system performance.

[0004] Secondly, there is the heterogeneous multi-sensor fusion solution. This solution introduces different types of sensors, such as forward-facing cameras and LiDAR, and fuses them with corner radar to leverage the complementary advantages of different sensors to obtain richer environmental attributes. For example, cameras provide texture and classification information, while LiDAR provides accurate 3D point clouds. However, this solution has significant drawbacks: on the one hand, the introduction of devices such as LiDAR significantly increases the system hardware cost; on the other hand, the time synchronization, spatial calibration, and data association algorithms between heterogeneous sensors are extremely complex, placing higher demands on the computing power of the central computing platform and introducing additional fusion error sources, making it difficult to widely deploy this solution in mass-produced vehicles where cost control is a priority.

[0005] It should be further explained that the existing hardware architecture of corner radars is not adapted to the requirements of distributed collaborative sensing. Traditional corner radars lack clear module division for functions such as signal transmission and reception, data processing, and algorithm execution. Even if some radars have basic signal transmission and reception and data processing capabilities, they have not formed the hardware foundation (such as multi-core processing units adapted to multiple tasks, efficient bus communication interfaces, etc.) to support "distributed algorithm division of labor + data exchange". This makes it impossible for corner radars to achieve computing load balancing through functional reconfiguration, further exacerbating the shortcomings of the two types of solutions mentioned above. Summary of the Invention

[0006] To address the aforementioned problems in the prior art, this invention proposes a fusion sensing method, fusion sensing device, and computer-readable storage medium for corner radar, which can improve real-time sensing performance and effectively reduce system costs.

[0007] Specifically, this invention proposes a fusion perception method for corner radars, applied to a vehicle, wherein the vehicle is equipped with at least four corner radars, namely, a left front corner radar and a right front corner radar located on the front side of the vehicle, and a left rear corner radar and a right rear corner radar located on the rear side of the vehicle. The fusion perception method includes the following steps: Each of the aforementioned corner radars processes the echo signals it receives to obtain raw point cloud data; Control each of the aforementioned corner radars to acquire the raw point cloud data of the corresponding left and right side corner radars; The left front corner radar and the left rear corner radar are controlled to execute a moving target tracking algorithm based on their own original point cloud data and the acquired original point cloud data, respectively, to obtain their respective first target point information; at the same time, the right front corner radar and the right rear corner radar are controlled to execute a stationary target passable area algorithm based on their own original point cloud data and the acquired original point cloud data, respectively, to obtain their respective point trace information. The left front corner radar and the left rear corner radar acquire each other's first target point information, and based on their own first target point information and the acquired first target point information, they execute the moving target tracking algorithm again to obtain their respective second target point information. The first target point information, the second target point information, and the dot information are sent to the vehicle's ADAS domain controller.

[0008] According to one embodiment of the present invention, the vehicle is equipped with a first CAN bus, which controls each of the corner radars to acquire the raw point cloud data of the corresponding left and right side corner radars through the first CAN bus.

[0009] According to one embodiment of the present invention, the vehicle is equipped with a second CAN bus, which controls the left front corner radar and the left rear corner radar to acquire each other's first target point information through the second bus.

[0010] According to one embodiment of the present invention, the automobile is equipped with a front public CAN bus disposed on the front side of the vehicle body and a rear public CAN bus disposed on the rear side of the vehicle body; The left front corner radar is controlled to send the first target point information and the second target point information to the ADAS domain controller via the front public CAN bus; The left rear corner radar is controlled to send the first target point information and the second target point information to the ADAS domain controller via the rear public CAN bus. The right front corner radar and the right rear corner radar are controlled to send their respective spot information to the ADAS domain controller via the front public CAN bus and the rear public CAN bus, respectively.

[0011] According to one embodiment of the present invention, both the first target point information and the second target point information include the target's distance, speed, position, RCS, and target type; the target type includes vehicles and pedestrians.

[0012] According to one embodiment of the present invention, the static target traversable area algorithm is a boundary screening algorithm, and the point information is a traversable area boundary line or obstacle contour point based on a radar coordinate system. The boundary screening algorithm includes static obstacle detection and fence point clustering steps.

[0013] According to one embodiment of the present invention, the fusion perception method further includes controlling the ADAS domain controller to fuse the first target point information and the second target point information to generate a global dynamic target tracking list; and / or controlling the ADAS domain controller to fuse the point information to generate a global drivable area model.

[0014] The present invention also provides an angle radar applicable to the aforementioned fusion sensing method, the angle radar comprising a transmitting antenna, a receiving antenna, and a single-chip processing system, the single-chip processing system comprising: The radio frequency unit transmits millimeter-wave signals through the transmitting antenna and receives echo signals through the receiving antenna. A signal processing unit is used to process the echo signal to obtain raw point cloud data; The multi-core isomorphic MCU includes a scheduling core and a computing core. The scheduling core is used to execute the basic tasks of the corner radar, including wave transmission task processing, signal diagnostic processing, and communication task management. The computing core is used to execute the moving target tracking algorithm or the stationary target passable area algorithm.

[0015] According to one embodiment of the present invention, the corner radar further includes a power management chip, which supplies power to the monolithic processing system via the SPI / I2C protocol.

[0016] According to one embodiment of the present invention, the single-chip processing system is externally connected to two CAN buses. One CAN bus is used to transmit raw point cloud data, and the other CAN bus is used to transmit the running results of the moving target tracking algorithm or the static target passable area algorithm to the ADAS domain controller.

[0017] The present invention also provides a fusion sensing device for an angle radar, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the fusion sensing method described in any of the preceding claims.

[0018] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the fusion sensing method as described in any of the preceding claims.

[0019] This invention provides a fusion perception method, fusion perception device, and computer-readable storage medium for corner radar. By using corner radar configured on a vehicle body for modular division of labor and data interaction, it enables distributed algorithm execution, improves real-time perception, reduces costs, and outputs comprehensive environmental perception information.

[0020] It should be understood that the above general description and the following detailed description of the present invention are exemplary and illustrative, and are intended to provide further explanation of the present invention. Attached Figure Description

[0021] The accompanying drawings are included to provide further explanation of the invention; they are incorporated into and constitute a part of this application. The drawings illustrate embodiments of the invention and, together with this specification, serve to explain the principles of the invention. In the drawings: Figure 1 A flowchart of a fusion sensing method for an angle radar according to an embodiment of the present invention is shown.

[0022] Figure 2 A schematic diagram of the structure of a corner radar fusion sensing system according to an embodiment of the present invention is shown.

[0023] Figure 3 A timing control diagram for the transmission of a dual-sided corresponding angle radar according to an embodiment of the present invention is shown.

[0024] Figure 4 A schematic diagram of the structure of an angle radar according to an embodiment of the present invention is shown.

[0025] Figure 5 A system block diagram of a fusion sensing device based on corner radar according to an embodiment of the present invention is shown.

[0026] The above figures include the following reference numerals: Car 200 Corner radar 210, 400 Left front corner radar 211 Right front corner radar 212 Left rear corner radar 213 Right rear corner radar 214 ADAS Domain Controller 220 First CAN bus 230 Second CAN bus 240 CAN bus 250 CAN bus 260 Transmitting antenna 410 Receiving antenna 420 430 single-chip processing system RF Unit 431 Signal processing unit 432 Multi-core homogeneous MCU 433 Scheduling core 434 Computing core 435 Power management chip 440 CAN bus 450 Detailed Implementation

[0027] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0028] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this application or its application or use. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0029] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0030] Unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps described in these embodiments do not limit the scope of this application. It should also be understood that, for ease of description, the dimensions of the various parts shown in the drawings are not drawn to actual scale. Techniques, methods, and devices known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and devices should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values. It should be noted that similar reference numerals and letters in the following drawings denote similar items; therefore, once an item is defined in one drawing, it need not be further discussed in subsequent drawings.

[0031] In the description of this application, it should be understood that the orientation or positional relationship indicated by directional terms such as "front, back, up, down, left, right", "horizontal, vertical, horizontal" and "top, bottom" is usually based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing this application and simplifying the description. Unless otherwise stated, these directional terms do not indicate or imply that the device or element referred to must have a specific orientation or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on the scope of protection of this application; the directional terms "inner" and "outer" refer to the inner and outer contours relative to the outline of each component itself.

[0032] Furthermore, it should be noted that the use of terms such as "first" and "second" to define components is merely for the purpose of distinguishing the corresponding components. Unless otherwise stated, these terms have no special meaning and therefore should not be construed as limiting the scope of protection of this application. In addition, although the terminology used in this application is selected from commonly known and used terms, some terms mentioned in this application's specification may have been chosen by the applicant according to his or her judgment, and their detailed meanings are explained in the relevant sections of this description. Moreover, this application should be understood not only through the actual terms used, but also through the meaning implied by each term.

[0033] Figure 1 A flowchart of a fusion sensing method for an angle radar according to an embodiment of the present invention is shown. Figure 2 A schematic diagram of a corner radar fusion perception system according to an embodiment of the present invention is shown. As shown in the figure, the corner radar fusion perception method provided by the present invention addresses the problems of insufficient computing power of existing single corner radars and excessively high cost of fusion of heterogeneous sensors. Based on the four corner radars 210 mounted on the vehicle 200, configured as a left front corner radar 211, a right front corner radar 212, a left rear corner radar 213, and a right rear corner radar 214, a distributed cooperative perception scheme is achieved. The fusion perception method includes the following steps: First, each corner radar 210 is controlled to process the echo signals it receives to obtain raw point cloud data. This step relies on the corner radar 210 to complete the echo calculation and extract the target's basic attributes such as distance, angle, and speed. This data is the source for subsequent perception algorithms, ensuring that each radar first acquires environmental information within its own detection range.

[0034] To overcome the limited field of view of a single corner radar 210, it is necessary to control each corner radar 210 to acquire the raw point cloud data of the corresponding left and right side corner radars 210. After the corresponding left and right side corner radars 210 exchange raw point cloud data, each corner radar 210 can obtain more comprehensive basic data of the entire vehicle's perimeter, avoiding the loss of information from a single perspective.

[0035] Next, the left front corner radar 211 and left rear corner radar 213 execute moving target tracking algorithms based on their own and corresponding side's original point cloud data, while the right front corner radar 212 and right rear corner radar 214 execute stationary target passable area algorithms based on their own and corresponding side's original point cloud data. This left-right division of labor is because both moving target tracking and stationary target area detection are computationally intensive tasks. If a single radar operates simultaneously, it will cause resource contention and increase processing latency. By splitting the tasks according to the vehicle's position, the left front corner radar 211 and left rear corner radar 213 are responsible for executing the moving target tracking algorithm, while the right front corner radar 212 and right rear corner radar 214 are responsible for executing the stationary target passable area algorithm. This allows each radar to focus on a single algorithm, with the left corner radars 211 and 213 obtaining the first target point information and the right corner radars 212 and 214 obtaining the point trace information.

[0036] Subsequently, the left front corner radar 211 and the left rear corner radar 213 exchange information on the first target point, and based on the fused information, they execute the moving target tracking algorithm again to obtain their respective second target point information. This step is an important design for the left-side radar cluster to achieve collaborative perception. The purpose is to compensate for the detection blind spots of a single front or rear corner radar by complementing the information from the front and rear radars, thereby enhancing the integrity and reliability of full-domain dynamic target tracking. This is because the detection range of a single corner radar 210 is limited by its installation position and angle. Specifically, the detection field of view of the left front corner radar 211 focuses more on the left front side of the vehicle and the extended area in front. Although it integrates the detection data of the right front corner radar 212, it is difficult to completely cover the close-range area at the rear. The left rear corner radar 213 is the opposite. When the two work alone, target tracking breaks are likely to occur. The moving target tracking algorithm is executed again, enabling the left front corner radar 211 to more completely capture the entire trajectory of a target moving from rear to front in the same direction as the vehicle, avoiding tracking loss due to blind spots when the target enters the front area from the rear area of ​​the vehicle. Similarly, the left rear corner radar 213 can accurately capture the target information of a target moving from front to rear in the opposite direction, making up for the insufficient detection of distant targets at the rear by the left front corner radar 211. Finally, the two algorithms are fused to form a second target point information covering the entire front and rear areas of the vehicle, providing more comprehensive and stable basic data for the ADAS domain controller 220 to generate a global dynamic target tracking list.

[0037] Finally, the information on the first target point, the second target point, and the tracking information are sent to the ADAS domain controller 220. With its enhanced computing power, the ADAS domain controller 220 can fuse this information into a global dynamic target tracking list and a drivable area model, ultimately outputting a more comprehensive and reliable environmental perception result, providing support for the vehicle's assisted driving or autonomous driving decisions. It should be noted that the functions of the left corner radars 211 and 213 and the right corner radars 212 and 214 are interchangeable to ultimately obtain the same environmental perception result.

[0038] In automotive environments, the CAN bus, due to its stability, reliability, adaptability to complex electromagnetic environments, and low communication latency, becomes the core carrier for data transmission between the corner radars 210 and between the corner radars 210 and the ADAS domain controller 220 in this invention. In some examples, refer to... Figure 2 The first CAN bus 230 configured in the vehicle 200 is specifically used to realize the mutual transmission of raw point cloud data between each corner radar 210 and the corresponding left and right side corner radars 210. The reason for using an independent first CAN bus 230 to complete this transmission task is to ensure the timeliness and stability of the transmission of the basic perception data of raw point cloud. After each corner radar 210 acquires the point cloud data of its own detection range, it can quickly supplement the point cloud information of the corresponding side radar, break through the perspective limitation of a single radar, and provide comprehensive basic data support for the subsequent precise execution of algorithms.

[0039] In some examples, the second CAN bus 240 configured in the vehicle 200 is used for the mutual transmission of first target point information between the left front corner radar 211 and the left rear corner radar 213. The independent configuration of this bus is to distinguish the transmission links of raw data and preliminary results of the algorithm, avoid transmission conflicts of different types of data, and ensure that the left front and left rear corner radars 211 and 213 can efficiently synchronize their preliminary dynamic target perception results, providing reliable communication guarantee for the secondary execution of the dynamic target tracking algorithm, making up for the blind spots of the front and rear radars, and generating more complete left-side global target information.

[0040] In some examples, to balance the bus communication load, avoid congestion on a single bus due to excessive data volume, and ensure smooth transmission of perception results to the ADAS domain controller 220, the vehicle 200 is also specifically configured with a front public CAN bus 250 located at the front of the vehicle and a rear public CAN bus 260 located at the rear of the vehicle. Based on this, the left front corner radar 211 sends the first target point information and the second target point information to the ADAS domain controller 220 via the front public CAN bus 250, while the left rear corner radar 213 completes the corresponding information transmission via the rear public CAN bus 260. The right front corner radar 212 and the right rear corner radar 214 also upload their respective point information to the ADAS domain controller 220 via the front public CAN bus 250 and the rear public CAN bus 260, respectively. This design, with public CAN buses configured according to front and rear vehicle zones, allows for parallel transmission of perception data from the front and rear radars, significantly improving data upload efficiency, and also provides convenience for the ADAS domain controller 220 to receive and integrate data by region.

[0041] In some examples, both the first target point information and the second target point information include the target's distance, speed, position, RCS (Radar Cross Section), and target type. Among these, distance, speed, and position are the core parameters for achieving dynamic target tracking, which can accurately depict the target's motion state. RCS can help distinguish the target's size and material. The target type is clearly defined as vehicles, pedestrians, etc., because these types of targets are the most critical dynamic obstacles during driving. Accurately identifying their type can provide key basis for the ADAS system's risk prediction and decision control. The combination and configuration of this information ensures the comprehensiveness and accuracy of dynamic target perception, laying the foundation for the subsequent generation of a global dynamic target tracking list by the domain controller.

[0042] In some examples, for the perception requirements of passable areas for static targets, the passable area algorithm for static targets adopts a lightweight boundary screening algorithm. This choice is made to adapt to the limited computing power of the Radar210 embedded chip, avoid resource contention and processing delays caused by algorithm complexity, and conform to the design concept of distributed division of labor to improve real-time performance in this invention. The algorithm specifically includes static obstacle detection and fence point clustering steps. Static obstacle detection can identify static objects that obstruct driving, such as curbs and stone blocks, while fence point clustering can delineate continuous static boundaries such as guardrails. Finally, it outputs the boundary line of the drivable area or the outline of obstacle points based on the radar coordinate system.

[0043] In some examples, the ADAS domain controller 220 integrates first and second target point information to generate a global dynamic target tracking list. This aims to address the spatiotemporal misalignment and complementary perspective issues of perception data from different angle radars 210. Because the front left and rear left radars 211 and 213 are installed in different locations, their acquisition times and spatial perspectives of the same dynamic target differ. The ADAS domain controller 220 first performs spatiotemporal synchronization processing on this target information, aligning the data of the same target collected by different radars at different times to a unified time reference and spatial coordinate system. Then, through a data association algorithm, it matches the first and second target point information of the same target from the front left and rear left radars 211 and 213, forming a complete target motion trajectory and avoiding duplicate target marking or tracking breakpoints. The resulting global dynamic target tracking list integrates complete status information such as distance, speed, position, and RCS of all dynamic targets (vehicles, pedestrians, etc.) around the vehicle, providing accurate global dynamic risk assessment for ADAS system functions such as lane change assist, collision warning, and cruise control.

[0044] Meanwhile, controlling the ADAS domain controller 220 to fuse point information to generate a global drivable area model is to overcome the limited field of view of a single-corner radar 210 for detecting static drivable areas. The point information output by the right front corner radar 212 and the right rear corner radar 214 can only cover the drivable area boundary or obstacle outline within their respective detection ranges, and there is a certain degree of discrete error. Based on the spatial consistency of the radar coordinate system, the ADAS domain controller 220 performs multi-view geometric fusion processing on this point information, smoothly stitching together the static boundary of the front side of the vehicle captured by the right front corner radar 212 and the static boundary of the rear side captured by the right rear corner radar 214 to form a continuous global static environment boundary; then, a lightweight optimization algorithm is used to eliminate the discrete error of the single radar point information, ultimately generating a complete and accurate global drivable area model. This model clearly defines the area where the vehicle can safely pass, and is the core static environment basis for the ADAS system to realize functions such as automatic parking, narrow road passage, and path planning. In conjunction with the global dynamic target tracking list, it jointly constructs a global environmental perception system to support the safety decisions of vehicle assisted driving or autonomous driving.

[0045] Figure 3 A timing control diagram for the transmission of dual-sided corresponding angle radars according to an embodiment of the present invention is shown. As shown in the figure, M and S represent the left and right corresponding angle radars 210. For example, the left front angle radar 211 is represented by S, and the right front angle radar 212 by M. M0-M3 are the transmission times of the right front angle radar 212, and S0-S4 are the transmission times of the left front angle radar 211. The staggered arrangement of the transmission times of radars M and S in the figure controls the transmission interval. That is, by staggering the transmission times, combined with the alternating transmission and reception design, when the left front angle radar 211 is transmitting point cloud data, the right front angle radar 212 is in receiving mode. Similarly, when the right front angle radar 212 is transmitting point cloud data, the left front angle radar 211 switches to receiving mode. This timing control directly achieves load balancing of the CAN bus 230, avoiding bus congestion caused by the simultaneous data transmission and reception of the left and right corresponding side radars 210. It ensures smoother and more timely transmission of raw point cloud data on the CAN bus 230, providing timing-level support for the process of exchanging raw point cloud data, and guaranteeing efficient and stable interaction of basic data in the distributed fusion sensing method.

[0046] Figure 4A schematic diagram of an angle radar according to an embodiment of the present invention is shown. As shown in the figure, the present invention also provides an angle radar 400 suitable for the aforementioned fusion sensing method. The angle radar 400 mainly consists of a transmitting antenna 410 (TX1~TX4), a receiving antenna 420 (RX1~RX4), and a single-chip processing system 430 (System on a Chip, SOC). This highly integrated architecture design meets the requirements of vehicle-mounted scenarios for hardware miniaturization and low power consumption. Because the space in the vehicle environment is limited, the integrated design can reduce the hardware footprint, while reducing signal interference between components and improving operational stability.

[0047] The single-chip processing system 430 includes a radio frequency unit 431, a signal processing unit 432, and a multi-core homogeneous MCU 433. Among them, the radio frequency unit 431, as the core module for signal transceiver, transmits millimeter-wave signals through a transmitting antenna 410 and receives echo signals through a receiving antenna 420. Millimeter-wave signals are chosen because of their short wavelength and high spatial resolution, which can accurately capture the detailed features of targets in the near-field environment of a vehicle, and meet the accuracy requirements of dynamic target tracking and static drivable area detection.

[0048] The function of the signal processing unit 432 is to process the echo signal to obtain raw point cloud data. The signal processing unit 432 can accurately extract key attributes of the target such as distance, angle, velocity, and RCS from the echo signal to form complete raw point cloud data. This data is the basic data source for subsequent moving target tracking algorithms and stationary target passable area algorithms, and its processing accuracy directly determines the reliability of subsequent perception results.

[0049] The multi-core homogeneous MCU433 employs a dual-core design with a scheduling core 434 and a computing core 435 to meet the task processing requirements of the corner radar 210 in the aforementioned fusion sensing method. The reason for using a multi-core architecture instead of a traditional single-core design is that when executing the fusion sensing method, the corner radar 210 needs to simultaneously complete basic tasks such as beam transmission control, signal diagnosis, and communication management, as well as algorithmic tasks such as moving target tracking or stationary target passability area tracking. A single-core architecture is prone to task conflicts and insufficient computing power, leading to increased processing latency. Specifically, the scheduling core 434 focuses on executing basic tasks such as beam transmission processing, signal diagnosis processing, and communication task management, ensuring the stable operation of the radar itself while providing timing coordination for data transmission and algorithm execution. The computing core 435 is dedicated to executing the adapted algorithms. When the corner radar 400 is the left front corner radar 211 or the left rear corner radar 213, the computing core 435 executes the moving target tracking algorithm; if it is the right front corner radar 212 or the right rear corner radar 214, it executes the stationary target passability area algorithm. This division of labor design allows algorithms to execute more efficiently, avoiding competition for computing power between basic tasks and algorithmic tasks, which aligns with the design concept of improving real-time perception through distributed division of labor.

[0050] In some examples, the corner radar 400 is also equipped with a power management chip 440, which supplies power to the monolithic processing system 430 via the SPI / I2C protocol. The SPI / I2C protocol offers advantages such as adaptable transmission rates, minimal hardware pin usage, and strong communication stability, making it ideal for the compact layout requirements of automotive hardware. The core function of the power management chip 440 is to provide a stable and compatible power supply to the monolithic processing system 430, preventing voltage fluctuations from interfering with the operation of the RF unit 431, signal processing unit 432, and multi-core homogeneous MCU 433. A stable power supply is essential for the radar to continuously and accurately output sensing data, directly impacting the stability of data transmission and algorithm execution in the fusion sensing method.

[0051] In some examples, the single-chip processing system 430 is also connected to two CAN buses 450 via a CAN transceiver. This design is also to adapt to the data transmission requirements of the aforementioned fusion perception method. One CAN bus 450 is dedicated to transmitting raw point cloud data, corresponding to the process in the method where "each corner radar acquires the corresponding left and right side point cloud data through the first CAN bus"; the other is dedicated to transmitting the running results of the moving target tracking algorithm or the stationary target passable area algorithm, corresponding to the requirement in the method to "send the first target point information, the second target point information, and the point trace information to the ADAS domain controller". This split transmission design can avoid conflicts between basic data such as raw point clouds and core perception data such as algorithm results during transmission, ensuring the timeliness and integrity of the transmission of both types of key data. At the same time, the CAN bus 450 itself has strong anti-electromagnetic interference capabilities and adaptability to the complex vehicle environment, which can further improve the reliability of data transmission.

[0052] In some examples, the monolithic processing system 430 connects to Ethernet for radar debugging or extended data transmission. The monolithic processing system 430 also connects to QSPI FLASH for storing radar programs and configuration data.

[0053] Figure 5A system block diagram of a corner radar-based fusion sensing device according to an embodiment of the present invention is shown. As shown, the fusion sensing device 500 may include an internal communication bus 501, a processor 502, a read-only memory (ROM) 503, a random access memory (RAM) 504, and a communication port 505. When applied to a personal computer, the fusion sensing device 500 may also include a hard disk 506. The internal communication bus 501 enables data communication between the components of the fusion sensing device 500. The processor 502 can perform judgments and issue prompts. In some embodiments, the processor 502 may consist of one or more processors. The communication port 505 enables data communication between the fusion sensing device 500 and external devices. In some embodiments, the fusion sensing device 500 can send and receive information and data from a network through the communication port 505. The fusion sensing device 500 may also include different types of program storage units and data storage units, such as a hard disk 506, a read-only memory (ROM) 503, and a random access memory (RAM) 504, capable of storing various data files used for computer processing and / or communication, as well as possible program instructions executed by the processor 502. The processor 502 executes these instructions to implement the main part of the method. The results processed by the processor 502 are transmitted to the user equipment via the communication port 505 and displayed on the user interface.

[0054] The above-described signal processing method can be implemented as a computer program, stored in the hard disk 506, and loaded into the processor 502 for execution to implement the fusion sensing processing method of this application.

[0055] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the aforementioned fusion sensing methods.

[0056] The specific implementation methods and technical effects of the fusion sensing device and computer-readable storage medium of the corner radar can be found in the embodiments of the detection method provided by the present invention, and will not be repeated here.

[0057] The present invention provides a fusion sensing method, fusion sensing device, and computer-readable storage medium for corner radar, which have the following beneficial effects: 1. Improved real-time perception: By using distributed algorithms for corner radar and task splitting across multi-core MCUs, the computational power competition of a single chip is avoided, reducing perception processing latency; 2. Controllable cost: Reuse existing corner radars without adding new heterogeneous sensors, adapting to the cost requirements of mass-produced vehicles; 3. Enhanced transmission stability: The split CAN bus design (first channel, second channel, front / rear public CAN) achieves load balancing, avoids data congestion, and ensures timely transmission; 4. Excellent hardware adaptability: The integrated architecture and multi-core MCU design of the corner radar accurately match the functional requirements of the distributed fusion sensing method, improving the reliability of system operation; 5. Comprehensive perception dimensions: Simultaneously outputs dynamic target information and static drivable area information, providing full-domain environmental support for ADAS.

[0058] Those skilled in the art will further appreciate that the various illustrative logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or a combination of both. To clearly illustrate this interchangeability between hardware and software, the various illustrative components, blocks, modules, circuits, and steps are described above in a generalized manner in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system. Those skilled in the art may implement the described functionality in different ways for each specific application, but such implementation decisions should not be construed as departing from the scope of the invention.

[0059] The various illustrative logic modules and circuits described in conjunction with the embodiments disclosed herein may be implemented or performed using a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. The general-purpose processor may be a microprocessor, but in alternatives, it may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors cooperating with a DSP core, or any other such configuration.

[0060] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of both. The software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor such that the processor can read and write information to / from the storage medium. In an alternative, the storage medium may be integrated into the processor. The processor and storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In an alternative, the processor and storage medium may reside as discrete components in the user terminal.

[0061] In one or more exemplary embodiments, the described functionality may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software as a computer program product, the functionality may be stored or transmitted as one or more instructions or code on or through a computer-readable medium. A computer-readable medium includes both computer storage media and communication media, encompassing any medium that facilitates the transfer of a computer program from one location to another. A storage medium may be any available medium accessible to a computer. By way of example and not limitation, such a computer-readable medium may include RAM, ROM, EEPROM, CD-ROM or other optical disc storage, disk storage or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and is accessible to a computer. Any connection is also legitimately referred to as a computer-readable medium. For example, if the software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of a medium. As used in this article, disk and disc include compact discs (CDs), laser discs, optical discs, digital multi-purpose discs (DVDs), floppy disks, and Blu-ray discs. Disks typically reproduce data magnetically, while discs reproduce data optically using lasers. Combinations of these should also be included within the scope of computer-readable media.

[0062] It will be apparent to those skilled in the art that various modifications and variations can be made to the exemplary embodiments described above without departing from the spirit and scope of the invention. Therefore, it is intended that this invention cover modifications and variations falling within the scope of the appended claims and their equivalents.

Claims

1. A fusion perception method for corner radars, applied to a vehicle, wherein the vehicle is equipped with at least four corner radars, the four corner radars being a left front corner radar and a right front corner radar located on the front side of the vehicle, and a left rear corner radar and a right rear corner radar located on the rear side of the vehicle, the fusion perception method comprising the steps of: Each of the aforementioned corner radars processes the echo signals it receives to obtain raw point cloud data; Control each of the aforementioned corner radars to acquire the raw point cloud data of the corresponding left and right side corner radars; The left front corner radar and the left rear corner radar are controlled to execute a moving target tracking algorithm based on their own original point cloud data and the acquired original point cloud data, respectively, to obtain their respective first target point information; at the same time, the right front corner radar and the right rear corner radar are controlled to execute a stationary target passable area algorithm based on their own original point cloud data and the acquired original point cloud data, respectively, to obtain their respective point trace information. The left front corner radar and the left rear corner radar acquire each other's first target point information, and based on their own first target point information and the acquired first target point information, they execute the moving target tracking algorithm again to obtain their respective second target point information. The first target point information, the second target point information, and the dot information are sent to the vehicle's ADAS domain controller.

2. The fusion sensing method as described in claim 1, characterized in that, The vehicle is equipped with a first CAN bus, which controls each of the corner radars to acquire the raw point cloud data of the corresponding left and right side corner radars through the first CAN bus.

3. The fusion sensing method as described in claim 1, characterized in that, The vehicle is equipped with a second CAN bus, which controls the left front corner radar and the left rear corner radar to obtain each other's first target point information through the second bus.

4. The fusion sensing method as described in claim 1, characterized in that, The vehicle is equipped with a front CAN bus located on the front side of the vehicle body and a rear CAN bus located on the rear side of the vehicle body. The left front corner radar is controlled to send the first target point information and the second target point information to the ADAS domain controller via the front public CAN bus; The left rear corner radar is controlled to send the first target point information and the second target point information to the ADAS domain controller via the rear public CAN bus. The right front corner radar and the right rear corner radar are controlled to send their respective spot information to the ADAS domain controller via the front public CAN bus and the rear public CAN bus, respectively.

5. The fusion sensing method as described in claim 1, characterized in that, The first target point information and the second target point information both include the target's distance, speed, position, RCS, and target type; the target type includes vehicles and pedestrians.

6. The fusion sensing method as described in claim 1, characterized in that, The static target passable area algorithm is a boundary screening algorithm. The point information is the boundary line of the drivable area or the outline of the obstacle based on the radar coordinate system. The boundary screening algorithm includes static obstacle detection and fence point clustering steps.

7. The fusion sensing method as described in claim 1, characterized in that, It also includes controlling the ADAS domain controller to fuse the first target point information and the second target point information to generate a global dynamic target tracking list; and / or controlling the ADAS domain controller to fuse the point information to generate a global drivable area model.

8. An angle radar, applicable to the fusion sensing method as described in any one of claims 1 to 7, characterized in that, The corner radar includes a transmitting antenna, a receiving antenna, and a single-chip processing system, wherein the single-chip processing system includes: The radio frequency unit transmits millimeter-wave signals through the transmitting antenna and receives echo signals through the receiving antenna. A signal processing unit is used to process the echo signal to obtain raw point cloud data; The multi-core isomorphic MCU includes a scheduling core and a computing core. The scheduling core is used to execute the basic tasks of the corner radar, including wave transmission task processing, signal diagnostic processing, and communication task management. The computing core is used to execute the moving target tracking algorithm or the stationary target passable area algorithm.

9. The angle radar as described in claim 8, characterized in that, The corner radar also includes a power management chip, which supplies power to the monolithic processing system via the SPI / I2C protocol.

10. The angle radar as described in claim 8, characterized in that, The single-chip processing system is connected to two CAN buses. One CAN bus is used to transmit the raw point cloud data, and the other CAN bus is used to transmit the running results of the moving target tracking algorithm or the static target passable area algorithm to the ADAS domain controller.

11. A fusion sensing device for angle radar, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the fusion sensing method as described in any one of claims 1-7.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the fusion sensing method as described in any one of claims 1-7.

Images