Vehicle surrounding environment sensing method based on sensor fusion

Abstract

The invention relates to a method for sensing the surrounding environment of a vehicle, and the vehicle comprises an ultrasonic sensor and a millimeter wave radar which are used for detecting the same environment zone. The method comprises at least the following steps: converting obstacle data detected based on an ultrasonic sensor and point cloud data detected based on a millimeter wave radar into a same coordinate system; and filtering the obstacle data by using the point cloud data to obtain first real obstacle data representing the relative position of the real obstacle and the vehicle. A computer program product and a related vehicle electronic control unit that may implement the method are also provided.

Description

Technical Field

[0001] This application relates generally to the field of intelligent assisted driving technology for vehicles, and more specifically to a vehicle surrounding environment perception method based on sensor fusion and related computer program products and vehicle electronic control units. Background Technology

[0002] Object detection is a crucial aspect of achieving assisted driving, intelligent driving, and even autonomous driving. It helps perceive the vehicle's surrounding environment and provides auxiliary and / or critical decision-making information for vehicle maneuvering in specific scenarios. Vehicles are typically equipped with various sensors to detect their surroundings and, accordingly, allow for the detection, identification, and tracking of specific targets (e.g., motor vehicles, pedestrians, obstacles, etc.) based on the detected information.

[0003] Typically, multiple ultrasonic sensors can be installed on the front and rear bumpers of a vehicle to detect obstacles close to the vehicle in scenarios such as parking, low-speed (e.g., 2-12 km / h) driving, or when the vehicle is stationary. (Ultrasonic sensors are typically designed to detect stationary obstacles or obstacles with low relative speeds to the vehicle.) This enables intelligent driver assistance control in these scenarios (e.g., assisted parking, emergency braking, target indication, etc.). Furthermore, multiple millimeter-wave radars can be arranged around the vehicle body (e.g., at the front and rear of the body (e.g., at the front and rear bumpers) and at the four corners) to detect targets around the vehicle while it is in motion, and to identify and track those targets accordingly, thereby enabling intelligent driver assistance control in these scenarios (e.g., cruise control, navigation-assisted driving, lane keeping assist, and lane change assist).

[0004] Based on the respective operating characteristics of ultrasonic sensors and millimeter-wave radar, vehicles can use different sensing devices to perceive environmental information according to different scenarios during use. For example, when vehicle speed detection instructs the vehicle to slow down at a speed below a certain threshold, or when the vehicle is stationary, and / or when a user command from the human-machine interface instructs the activation of the vehicle's parking assist or reversing assist functions, the millimeter-wave radar in the vehicle's perception system may not be activated (for example, since millimeter-wave radar cannot accurately identify static and quasi-static (low-speed moving) obstacles, in these scenarios, even if the millimeter-wave radar is activated, the information collected is often not used by the corresponding functional modules), and ultrasonic sensors can be activated to detect static or quasi-static targets around the vehicle and construct obstacles accordingly based on the detection information, thereby making corresponding decisions and controls. Alternatively, if vehicle speed detection indicates that the vehicle is traveling at a speed exceeding a certain threshold, and / or if a user command from the human-machine interface indicates that the vehicle's cruise control or navigation-assisted driving function is enabled, the ultrasonic sensors of the vehicle perception system may not be activated, and the millimeter-wave radar may be activated to provide detection of dynamic targets around the vehicle and, accordingly, generate a point cloud characterizing the location of the detected targets based on this detection information, thereby making corresponding decisions and controls. Thus, although various types of sensing devices are installed in the vehicle, the sensing methods used by the vehicle in static and low-speed dynamic scenarios are often fixed and singular, for example, often relying solely on ultrasonic sensors to detect targets.

[0005] Therefore, it is desirable to provide an improved target detection method that fully utilizes different sensing methods of vehicles, including ultrasonic sensors, to provide real-time, efficient, and accurate target detection in static and low-speed dynamic scenarios. Summary of the Invention

[0006] This application proposes an obstacle recognition method suitable for use when a vehicle is stationary or at low speed. By fusing detection information from ultrasonic sensors and millimeter-wave radar, it allows for the rapid, accurate, and efficient identification of real targets within close range around the vehicle.

[0007] According to one aspect of this application, a method for perceiving the surrounding environment of a vehicle is provided, the vehicle including an ultrasonic sensor and a millimeter-wave radar configured for detecting the same environmental zone, the method comprising at least the following steps: S1, converting obstacle data detected by the ultrasonic sensor and point cloud data detected by the millimeter-wave radar into the same coordinate system; and S2, using the point cloud data to perform data filtering on the obstacle data to obtain first real obstacle data characterizing the relative position of real obstacles and the vehicle.

[0008] According to another aspect of this application, a computer program product is provided, comprising a computer program / instructions, characterized in that the computer program / instructions, when executed by a processor, implement the steps of the method described above.

[0009] According to another aspect of this application, an electronic control unit for a vehicle is provided, the electronic control unit being configured to implement the method described above and further implement driving assistance functions based on first real obstacle data, optionally, the driving assistance functions including at least one of parking assistance, reversing assistance, front cross-target braking, front cross-target warning, blind spot detection, low-speed emergency braking, and rear automatic emergency braking. Attached Figure Description

[0010] The embodiments according to the principles of this application are described in detail below with reference to the accompanying drawings. The drawings are given by way of example to facilitate understanding of the specific embodiments described. However, the drawings are not intended to be limiting. Accordingly, any unclaimed features shown in the drawings should not be construed as essential features for implementing the principles of the invention, nor should any claimed features shown in the drawings be construed as the only way to implement the function associated with that feature.

[0011] Figure 1A This is a schematic block diagram illustrating an example vehicle configuration in which the sensing method according to the principles of this application can be employed; and Figure 1B This is a flowchart illustrating an embodiment of a vehicle surrounding environment perception method based on the principles of this application.

[0012] Figure 2A and Figure 2B This is a schematic diagram illustrating the relationship between obstacle points (represented by large dots) constructed from cross-echoes received by ultrasonic sensors and point clouds (represented by small dots) generated from electromagnetic echoes received by millimeter-wave radar, relative to the real target (in...). Figure 2A The middle is a traffic cone, and in Figure 2B The image shows the location of "No Parking" signs.

[0013] Figure 3 It is a flowchart illustrating an example of a driving assistance method suitable for use in low-speed and static scenarios. Detailed Implementation

[0014] The basic concept and principles of the present invention are described below in detail with reference to preferred embodiments known to the inventors. It is understood that the following description is provided to make this disclosure sufficient and complete and to convey the spirit of the invention to those skilled in the art. Accordingly, the description is merely illustrative. Those skilled in the art, upon reading the following description, will be able to modify, alter, and substitute the disclosed embodiments as appropriate, without departing from the spirit and teachings of the invention.

[0015] For ease of description and to facilitate understanding, terms such as “low speed,” “close range,” “vehicle coordinate system,” “vehicle lateral direction,” and “vehicle vertical direction” are used herein, and these terms have the same meanings as commonly understood by those skilled in vehicle engineering and vehicle driver assistance technology. For example, “low speed” may refer to a vehicle speed in the range of 2-12 km / h, and may primarily relate to scenarios such as reversing, parking, and / or slow-moving traffic. Furthermore, “close range” generally refers to a distance range of no more than 3 meters, but depending on the specific configuration of the vehicle's ultrasonic sensors, the close range sensing range of the vehicle may be appropriately farther or closer.

[0016] Furthermore, although only the use of ultrasonic sensors and millimeter-wave radar is described herein, the vehicle may also be equipped with other sensing devices (e.g., lidar, infrared sensors, cameras, etc., as additional parts of the vehicle's perception system) without affecting the implementation of the target detection method (also referred to as the "vehicle surrounding environment perception method" or "perception method") disclosed herein. Accordingly, as will be readily understood by those skilled in the art, the vehicle may be provided with an electronic control unit (ECU) communicatively connected to the perception system formed by these sensing devices, and the ECU is also electrically connected to the vehicle's drive system, steering system, braking system, and human-machine interface to make decisions based on real-time information received by the perception system and accordingly send instructions to one or more of the drive system, steering system, braking system, and human-machine interface to achieve corresponding operations. Further, the vehicle's ECU may be configured to include a memory and a processor, the memory being configured to store pre-programmed programs and / or instructions for assisting vehicle driving, and the programs and / or instructions, when executed by the processor, can determine immediate corresponding actions for assisting vehicle driving.

[0017] Figure 1A The illustration shows an example vehicle configuration that can employ the sensing method based on the principles of this application. Figure 1AAs shown, vehicle 10 is equipped with an ultrasonic sensor 12 and a millimeter-wave radar 14 for detecting the same environmental zone, and an automatic parking controller 16 communicatively connected to both the ultrasonic sensor 12 and the millimeter-wave radar 14. The automatic parking controller 16 is configured to implement parking assistance functions based on detection information and / or intermediate processed data received from the ultrasonic sensor 12 and the millimeter-wave radar 14. Accordingly, the automatic parking controller 16 may also communicatively connect to one or more other vehicle components to implement corresponding actions of the parking assistance functions. Obstacle location information available for use by the parking assistance functions can be obtained through sensor fusion of the ultrasonic sensor 12 and the millimeter-wave radar 14.

[0018] Figure 1B This is a flowchart illustrating an example of a sensor fusion-based method 100 for perceiving the vehicle's surrounding environment according to the principles of this application. Accordingly, method 100 may be implemented electronically by the automatic parking controller 16 shown in Figure 1 to provide real-time obstacle location information. Alternatively, method 100 may also be implemented electronically by other domain controllers during vehicle operation to provide corresponding autonomous driving or driver assistance functions. For convenience, the collection of one or more electronic circuits configured to perform detection signal processing and corresponding logical operations will be collectively referred to as an electronic control unit (ECU). Accordingly, the ECU may be integrated on a single printed circuit board and presented as, for example, an automatic parking controller or a (area) controller; or, the ECU may be distributed and selectively linked together in an end-to-end manner, thus having different physical link arrangements for different usage functions.

[0019] At step 120, real-time ultrasonic detection information of the vehicle's surrounding environment from the onboard ultrasonic sensor is acquired, and simultaneously, real-time radar detection information of the vehicle's surrounding environment from the onboard millimeter-wave radar is acquired. In one example, the communication connection between the vehicle's electronic control unit and the sensing system can be used to: i) send excitation signals to the onboard ultrasonic sensor and millimeter-wave radar to cause them to generate and transmit corresponding ultrasonic and electromagnetic signals; and ii) acquire the ultrasonic echo signals and electromagnetic echo signals received by the ultrasonic sensor and millimeter-wave radar. Thus, through processing and analysis of the transmitted and returned ultrasonic and electromagnetic signals, real-time ultrasonic detection information and real-time radar detection information are determined. In another example, real-time ultrasonic detection information can be directly acquired from the ultrasonic sensor and real-time radar detection information from the millimeter-wave radar. The real-time ultrasonic detection information is, for example, the real-time distance between the ultrasonic sensor and the obstacle calculated by the built-in microcontroller (MCU) of the ultrasonic sensor based on the transmitted and received signals of the ultrasonic sensor. Similarly, the real-time radar detection information is, for example, the position information of the detected object relative to the radar obtained by analyzing the transmitted and received signals of the millimeter-wave radar using the built-in chip of the millimeter-wave radar. Therefore, the detection information obtained in step 120 can be either raw detection information including, for example, transmitted signals, received signals and clock pulse signals, or intermediate detection information obtained after preliminary processing of the raw detection information.

[0020] Optionally, step 120 may further include acquiring the position information (e.g., three-dimensional coordinates in the vehicle coordinate system) of the corresponding ultrasonic sensors and millimeter-wave radar during the electronic control unit initialization phase. Accordingly, the real-time ultrasonic detection information and real-time radar detection information acquired at step 120, as well as the optional position information of the ultrasonic sensors and millimeter-wave radar, are then further processed in step 140.

[0021] In step 140, obstacle coordinates (schematically represented by large dots in Figure 2) are constructed in the vehicle coordinate system using the real-time ultrasonic detection information obtained in step 120, and point cloud coordinates (schematically represented by small dots in Figure 2) are generated in the vehicle coordinate system using the real-time radar detection information obtained in step 120. The obstacle coordinates represent the position information of possible targets detected by ultrasonic sensors relative to the vehicle, and the point cloud coordinates represent the position information of potential targets detected by millimeter-wave radar relative to the vehicle.

[0022] In some cases, due to the natural divergence of the ultrasonic detection beam, the irregularity of the target shape, the absorption of sound waves by the target material, and ground reflection, the reflected signals received by the ultrasonic sensor may include echo signals from multiple different reflection paths. These include direct reflection signals reflected directly from the target to the ultrasonic sensor, and multipath reflection signals reflected or scattered from the target to other surfaces or objects and then reflected back to the ultrasonic sensor. Accordingly, the obstacle coordinates constructed based on the echo signals received by the ultrasonic sensor may include a first real obstacle coordinate OB1 representing the real obstacle and a second noisy obstacle coordinate OB2, OB21, OB22 representing a virtual mirror image of the real obstacle, as shown in... Figure 2A and Figure 2B As illustrated in the diagram.

[0023] Furthermore, in the detection of static or quasi-static objects, due to the discontinuities of the target surface, the radar detection orientation, and the similar electromagnetic wave intensities of nearby objects, the electromagnetic wave signal received by millimeter-wave radar may contain reflected signals from the actual target measurement points and other clutter signals. Accordingly, the point cloud coordinates generated based on the electromagnetic wave signals received by the millimeter-wave radar may include a series of point cloud coordinates distributed within the actual target area, including, for example, one or more first target point cloud coordinates PC1 located on one or more reflective surfaces of the actual target, and one or more second peripheral point cloud coordinates PC2 dispersed around the one or more reflective surfaces, such as in... Figure 2B The best example shown is...

[0024] Accordingly, the constructed obstacle coordinates, which include both the first real obstacle coordinates and the second noisy obstacle coordinates, and the generated point cloud coordinates, which include the first target point cloud coordinates and the second surrounding point cloud coordinates, are then fused at step 160.

[0025] At step 160, the point cloud coordinates obtained at step 140 are used to filter the obstacle coordinates obtained at step 140 to determine the coordinate information of the real obstacle relative to the vehicle.

[0026] Although millimeter-wave radar has poor resolution and correspondingly low detection accuracy for obstacles in static or quasi-static scenarios, its point cloud coordinates essentially reflect the actual obstacle's location. In contrast, while ultrasonic sensors are suitable for obstacle distance detection in static or quasi-static scenarios, they often produce secondary noisy obstacle coordinates, making target detection algorithms based solely on ultrasonic sensors prone to misjudgments. In this situation, by combining the operating characteristics of millimeter-wave radar and ultrasonic sensors, point cloud coordinates can be advantageously used to distinguish between real obstacles and noisy obstacles. This allows for the identification of the first real obstacle coordinates (representing the first real obstacle perceived by the ultrasonic sensor) and the second noisy obstacle coordinates (representing noisy obstacles constructed from signals perceived by the ultrasonic sensor), enabling the determination and filtering of the second noisy obstacle coordinates.

[0027] Accordingly, in one instance, step 160 can be implemented by labeling the first true obstacle coordinates and the second noisy obstacle coordinates by identifying obstacle coordinates that are within the envelope space defined by the point cloud coordinates and obstacle coordinates that are outside the envelope space defined by the point cloud coordinates. When it is determined that an obstacle coordinate is within the envelope space defined by the point cloud coordinates, the obstacle coordinate is identified as the first true obstacle coordinate and is retained for later use; and when it is determined that an obstacle coordinate is outside the envelope space defined by the point cloud coordinates, the obstacle coordinate is identified as the second noisy obstacle coordinate and is filtered out.

[0028] Whether an obstacle's coordinates lie within the envelope space defined by the point cloud coordinates can be determined in several ways. In one example, the coordinates of a first real obstacle and a second noisy obstacle are distinguished based on the range of point cloud coordinates along the vehicle's lateral and / or longitudinal directions. For instance, if the lateral coordinates contained in the obstacle's coordinates fall within the lateral coordinate range of the point cloud coordinates, then the obstacle's coordinates can be determined as the first real obstacle representing a real obstacle; conversely, if the lateral coordinates contained in the obstacle's coordinates fall outside the lateral coordinate range of the point cloud coordinates, then the obstacle's coordinates can be determined as the second noisy obstacle representing a virtual mirror image of a real obstacle. Furthermore, whether an obstacle's coordinates map a real obstacle can also be determined based on the coordinate range information of the point cloud coordinates along other directions of the two-dimensional vertical plane.

[0029] In another instance, such as Figure 2B As shown, an extended diameter range DR corresponding to a point cloud coordinate is defined with the point cloud coordinate as the center, and it is detected whether the obstacle coordinate intersects with or falls within the extended diameter range of at least one point cloud coordinate. If the obstacle coordinate intersects with the extended diameter range of one of the point cloud coordinates (e.g., in...), the process is complete. Figure 2BIf an obstacle coordinate falls within the extended diameter of at least one point cloud coordinate (as shown in the diagram), then the obstacle coordinate is determined to be the first true obstacle coordinate. Conversely, if the obstacle coordinate neither intersects with nor falls within the extended diameter of any point cloud coordinate, then the obstacle coordinate is determined to be the second noise obstacle coordinate. Preferably, the extended diameter has a radius of less than 20 cm, and more preferably, the extended diameter has a radius of greater than 10 cm. It is understood that the radius of the extended diameter can also be adjusted according to the actual application scenario (e.g., the functional module to which it is applied) and / or the required performance.

[0030] In another instance, step 160 may include determining whether an obstacle coordinate is a first true obstacle coordinate or a second noisy obstacle coordinate by searching for the presence of point cloud coordinates in the neighborhood of the obstacle coordinate. When point cloud coordinates are found in the neighborhood of the obstacle coordinate, the obstacle coordinate is determined to be the first true obstacle coordinate and is retained for later use. When it is confirmed after the search that no point cloud coordinates are found in the neighborhood of the obstacle coordinate, the obstacle coordinate is determined to be the second noisy obstacle coordinate and is filtered out from the obstacle coordinates.

[0031] exist Figure 2A In the example, the scan window SW is used to perform a neighborhood search for obstacle coordinates. For example... Figure 2A As shown, the scanning window SW is defined as a rectangular cell frame, including two long sides at a specified vertical distance from the obstacle coordinates and two short sides at a specified horizontal distance from the obstacle coordinates. During the search, the scanning window SW changes its positioning by rotating about a vertical axis (not shown) passing through the obstacle coordinates to achieve radial scanning and allows searching at various angles along the horizontal direction. In this way, it can be determined whether point cloud coordinates exist within a cylindrical space centered on the obstacle coordinates. Preferably, the scanning window SW has a length of no more than 20 cm, more preferably, the length is greater than 10 cm. However, it is understood that the length of the scanning window can also be adjusted according to the actual application scenario (e.g., the functional module to which it is applied) and / or the required performance.

[0032] After the coordinates of the second noisy obstacle are filtered out, the retained coordinates of the first real obstacle provide filtered detection information characterizing the coordinate position of the real obstacle. Accordingly, this real-time filtered detection information obtained by fusing the ultrasonic sensor and millimeter-wave radar is then output at step 180.

[0033] At step 180, based on the real-time filtered detection information obtained in step 140, it is determined whether a target exists within the area detected by the ultrasonic sensor and millimeter-wave radar, and a report on the detected target is output, the report containing two-dimensional or three-dimensional coordinate information of the target relative to the vehicle. In one example, step 180 may further include using multiple coordinates contained in the real-time filtered detection information to determine the outline of the target and making further decision analysis based on the determined target outline. Using the report on the detected target output by step 180, other modules loaded in the driver assistance system control unit can make further decisions based on a predetermined algorithm. For example, the report on the detected target output by step 180 may also include other analytical data, such as the minimum distance between the detected target and the vehicle, the moving speed of the detected target, the lateral span of the detected target, whether the detected target is in the vehicle's blind spot, etc., so that the control unit can respond to the report in a timely manner to formulate a parking plan, evaluate a reversing strategy, and / or trigger emergency braking and / or emergency alerts.

[0034] Figure 3 An example of a method suitable for assisting vehicle driving in low-speed and static scenarios is illustrated. The vehicle may include multiple perception zones associated with a perception system, such as the front, rear, sides, and blind spots of the vehicle, and at least one ultrasonic sensor and at least one millimeter-wave radar may be configured in the same perception zone. Method 200 may be used in conjunction with the triggering of vehicle driving assistance functions such as assisted parking, assisted reversing, FCTB (Forward Cross-Target Braking), FCTA (Forward Cross-Target Alert), MEB (Low-Speed ​​Emergency Braking), and RAEB (Rear Automatic Emergency Braking) in response to a detection of vehicle speed indicating that the current vehicle speed is below a preset threshold. Accordingly, a pre-programmed driver assistance system for implementing the vehicle driving assistance functions may be loaded into the vehicle's electronic control unit. This driver assistance system is stored in the memory of the electronic control unit in machine-readable language and implements the steps of method 200 when executed by the processor of the electronic control unit.

[0035] At step 210, the vehicle's electronic control unit is instructed to send commands to the ultrasonic sensors and millimeter-wave radar in the designated sensing zone(s) to send ultrasonic and electromagnetic waves to detect the presence of objects (also referred to as "targets") in the designated sensing zone(s).

[0036] Subsequently, at step 220, the vehicle's electronic control unit (ECU) is prompted to receive real-time detection information of objects in a designated sensing zone from the ultrasonic sensor and the millimeter-wave radar via a communication connection with them. In one example, the ultrasonic sensor may be configured with a signal processing chip, and this signal processing chip is configured to perform preliminary processing on the ultrasonic signals emitted and received by the ultrasonic sensor to obtain the distance between the ultrasonic sensor and the detected obstacle. Accordingly, the real-time ultrasonic detection information received by the vehicle's ECU from the ultrasonic sensor may include data information characterizing the distance between the ultrasonic sensor and the obstacle. Similarly, in another example, the millimeter-wave radar may be configured with a signal processing chip, and this signal processing chip is configured to perform preliminary processing on the electromagnetic wave signals emitted and received by the millimeter-wave radar (including, for example, first-dimensional FFT, second-dimensional FFT, CFAR filtering, and point target detection, etc.) to obtain the position information of the detected point target relative to the millimeter-wave radar. Accordingly, the real-time radar detection information received by the vehicle's ECU from the millimeter-wave radar may include data information characterizing the positioning of the detected point target relative to the millimeter-wave radar.

[0037] Subsequently, at step 230, the vehicle's electronic control unit is prompted to use pre-stored ultrasonic sensor position information and millimeter-wave radar position information to convert the received real-time ultrasonic detection information and real-time radar detection information into obstacle data and point cloud data represented in the same coordinate system (e.g., the vehicle coordinate system). The obstacle data preferably includes three-dimensional coordinate data of the obstacles, and the point cloud data may optionally not include velocity information of the detected point targets.

[0038] Then, at step 240, the vehicle's electronic control unit (ECU) is prompted to perform data filtering on the obstacle data using point cloud data to obtain first real obstacle data representing the relative positional relationship between the real obstacle and the vehicle. Generally, the ECU is configured to use the positional information of the group of detection target points provided by the point cloud data to determine whether the corresponding obstacle data is noise. Optionally, data filtering of obstacle data using point cloud data can be implemented by: i) comparing the distribution range of detection target points defined by the coordinate points of the point cloud data along the vehicle's lateral and / or vertical directions with the coordinate positions of obstacles mapped by the corresponding obstacle data along the vehicle's lateral and / or vertical directions; detecting whether the coordinate points corresponding to the obstacle data fall within or intersect with the extended diameter range of the coordinate points corresponding to the point cloud data; and searching the neighborhood of the coordinate points corresponding to the obstacle data to determine whether there are coordinate points of the point cloud data near the coordinate points corresponding to the obstacle data. Accordingly, obstacle data that is inconsistent with the positional range information of the detection target points reflected by the point cloud data will be identified as noise and filtered out.

[0039] Subsequently, at step 250, the vehicle's electronic control unit (ECU) makes a decision based on the first real obstacle data acquired at step 240. For example, the ECU may, according to a pre-programmed parking assist algorithm, reversing assist algorithm, FCTB algorithm, MEB algorithm, or RAEB algorithm, send a command to the vehicle's braking system to induce the braking system to perform a corresponding braking operation in response to a detected obstacle distance being less than a predetermined distance threshold. Alternatively, the ECU may, according to a pre-programmed parking assist algorithm, reversing assist algorithm, or FCTA algorithm, send a command to the vehicle's human-machine interface (HMI) to induce the HMI to provide a target cues to the driver in a perceptible manner in response to a detected obstacle distance being less than a predetermined distance threshold. Or, the ECU may, according to a pre-programmed BSD (Blind Spot Detection) algorithm, send a command to the HMI to induce the HMI to provide a target cues to the driver in a perceptible manner in response to a detected obstacle being located within a predetermined blind spot. Optionally, the target cues may include both audible alarms / tips and visual target locations.

[0040] Therefore, by using sensor fusion of ultrasonic sensors and millimeter-wave radar, accurate target detection and noise suppression can be provided in static and low-speed dynamic scenarios without the need for additional physical sensing equipment, thereby improving the reliability and effectiveness of vehicle driving assistance systems.

[0041] Therefore, the preferred practices of sensor fusion-based target detection known to the inventors have been described in detail from both process method and system architecture perspectives. For brevity, certain elements or features of the principles of this disclosure may be described in conjunction with the process method, and other elements or features of the principles of this disclosure may be described in conjunction with the system architecture. It is understood that elements or features stated in different aspects may be combined to form other embodiments not specifically described in this application. Such embodiments are also included within the scope of this invention.

Claims

1. A method for sensing the surrounding environment of a vehicle, the vehicle comprising an ultrasonic sensor and a millimeter-wave radar configured for detecting the same environmental zone, the method comprising at least the following steps: S1. Convert obstacle data detected by ultrasonic sensors and point cloud data detected by millimeter-wave radar into the same coordinate system; and S2. Use point cloud data to filter obstacle data to obtain first real obstacle data that represents the relative position of real obstacles and vehicles.

2. The method according to claim 1, wherein, Step S2 includes the following sub-steps: S21, Identify the first real obstacle data from the obstacle data based on the location information of the detection points mapped by the point cloud data.

3. The method according to claim 2, wherein, Step S2 includes the following sub-steps: S22, when the location information of the obstacle mapped by the obstacle data matches the location information of the detection point mapped by the point cloud data, the obstacle data is determined to be the first real obstacle data.

4. The method according to claim 3, wherein, Determine whether the location information of the obstacle mapped by the obstacle data matches the location information of the detection points mapped by the point cloud data using any of the following methods: i) Compare the coordinate information of the obstacles mapped by the obstacle data along the vehicle's lateral and / or vertical directions with the coordinate range of the detection points mapped by the point cloud data along the vehicle's lateral and / or vertical directions. ii) Detect whether an extended diameter range centered on the detection point mapped by the point cloud data includes or intersects with an obstacle mapped by the obstacle data; optionally, the extended diameter range has a radius greater than 10 cm and less than 20 cm; and iii) Search the neighborhood of the obstacle mapped by the obstacle data to determine whether there is a detection point mapped by the point cloud data. Optionally, the neighborhood is defined as a range that is greater than 10 cm and less than 20 cm away from the obstacle mapped by the obstacle data.

5. The method according to claim 4, wherein, If at least one of the following conditions exists, the location information of the obstacle mapped by the obstacle data is determined to match the location information of the detection point mapped by the point cloud data: i) The coordinates of the obstacles mapped by the obstacle data along the vehicle's lateral and / or vertical directions fall within the coordinate range of the detection points mapped by the point cloud data along the vehicle's lateral and / or vertical directions. ii) The obstacles mapped by the obstacle data are contained within or intersect with the extended diameter of at least one point cloud data; as well as iii) There exists at least one detection point mapped by point cloud data in the neighborhood of the obstacle mapped by the obstacle data.

6. The method according to any one of claims 1-5, wherein, Step S2 includes the following sub-steps: S23, using point cloud data, the first real obstacle data is obtained by filtering out obstacle data that does not match the position information of the detection points mapped by the point cloud data.

7. The method according to claim 6, further comprising the following step: S3. Provide target detection results based on first real obstacle data; and / or S0. Predetermine the installation location information of the ultrasonic sensor and millimeter-wave radar in the vehicle to allow the creation of the obstacle data and the point cloud data in the vehicle coordinate system.

8. The method according to claim 7, wherein, The method is permitted to be used when the vehicle speed is below a predetermined speed threshold, and / or the method is enabled in scenarios where ultrasonic sensors are applicable.

9. A computer program product, comprising a computer program / instructions, characterized in that... When the computer program / instructions are executed by the processor, they implement the steps of the method as described in any one of claims 1-8.

10. An electronic control unit for a vehicle, configured to implement the method as described in any one of claims 1-8 and obtain first real obstacle data to achieve any of the following driving assistance functions: parking assist, reversing assist, front cross-target braking, front cross-target warning, blind spot detection, low-speed emergency braking, and rear automatic emergency braking.

11. The electronic control unit according to claim 10, wherein, The electronic control unit is a domain controller.

Images