A method and system for recognizing a limit block based on ultrasonic radar

Abstract

The application relates to the technical field of automatic parking, and provides a limiting block identification method and system based on ultrasonic radar, direct echo data and cross echo data are cooperatively collected by multiple ultrasonic radars, whether an obstacle meets preset shape characteristics is judged according to a measured cross echo path and a theoretical cross echo path, that is, the preset shape characteristics are verified in combination with a triangular positioning principle to perform preliminary screening; then, through limiting block feature matching, the distribution position characteristics of the obstacle are further analyzed to perform secondary positioning screening, so that accurate identification of low and small limiting blocks is realized; through pose detection of the limiting block, the safety and reliability of automatic parking into a position are further improved.

Description

Technical Field

[0001] This invention relates to the field of automatic parking technology, and in particular to a limit block recognition method and system based on ultrasonic radar. Background Technology

[0002] With the development of automotive technology and people's needs for car use, more and more cars are equipped with reversing assistance devices, including automatic parking systems and reversing radar alarm systems. Garage limit blocks (such as rubber parking barriers) serve as important reference points for parking spaces, ensuring accurate vehicle parking. However, the low (height ≤ 15cm) and small (width 5-10cm) structure of these limit blocks makes them susceptible to ground noise interference when detected by traditional ultrasonic radar, resulting in a high false alarm rate. Although existing technologies have applied the triangulation principle to the recognition of rod-shaped objects, they have not been specifically optimized for the characteristics of limit blocks in garage scenarios (fixed distance range, environmental noise interference), therefore the following technical problems still exist: 1. The inability to recognize parking space limit blocks and other obstacles leads to errors in the parking path calculation during automatic parking, resulting in the vehicle failing to park accurately or colliding with the limit blocks, thus affecting the parking experience. Furthermore, if the driver's estimation of the limit blocks is inaccurate when parking, there is a risk of collision, potentially damaging the vehicle and the parking space device.

[0003] Second, conventional single-radar detection is susceptible to interference from ground reflections, and traditional limit block recognition relies on preset thresholds, which cannot adapt to the differences in reflection characteristics of different materials (rubber / concrete).

[0004] Third, the system may misjudge the parking space boundary, causing the vehicle to lurch forward or roll backward excessively, especially in parking lots with a large slope, which may cause scratches or collisions.

[0005] Fourth, the obstruction problem: the limit block is located at the bottom of the parking space, which will block part of the ultrasonic radar's transmission and reception path, causing the radar to be unable to detect obstacles within a certain range behind the limit block, which may lead to false alarms and increase blind spots.

[0006] 5. Reflection Interference: The limiting block is made of hard materials such as metal or plastic, which will strongly reflect ultrasonic waves. In addition, if there are many recessed structures and irregular protrusions in the parking space, secondary echoes are easily generated, which may lead to radar misjudgment or false alarm, affecting the accuracy of distance and speed measurement. Summary of the Invention

[0007] This invention provides a limit block recognition method and system based on ultrasonic radar, which solves the technical problems of limited resolution, susceptibility to environmental interference, and inability to effectively recognize limit blocks in traditional ultrasonic radar technology, resulting in low reliability of automatic parking.

[0008] To address the above technical problems, this invention provides a method for identifying limit blocks based on ultrasonic radar, comprising the following steps: Obstacle data on obstacles in parking spaces are collected based on ultrasonic radar detection. The obstacle data includes measured cross echo paths and at least one set of direct echo distances. A direct echo measurement model is constructed based on the echo signal, and then an equivalent model of indirect echo is constructed based on the direct echo measurement model to calculate the theoretical cross echo path. Based on the measured cross-echo path and the theoretical cross-echo path, determine whether the obstacle meets the preset shape characteristics; if so, proceed to the next step of feature matching. The obstacle information of each frame is traversed, and feature matching is performed based on the standard geometric features of the limit block. If the match is consistent, the coordinate information is output, and the coordinate information is integrated to obtain the limit block information.

[0009] This basic solution uses multiple ultrasonic radars to collaboratively collect direct echo data and cross echo data. Based on the measured and theoretical cross echo paths, it determines whether obstacles meet preset shape characteristics, i.e., it verifies the preset shape characteristics using the triangulation principle for initial screening. Then, through limit block feature matching, it further analyzes the distribution and location characteristics of obstacles for secondary positioning and screening, thereby achieving accurate identification of low and small limit blocks. Through limit block pose detection, the safety and reliability of automatic parking are further improved.

[0010] In a further implementation scheme, based on the measured cross-echo path and the theoretical cross-echo path, it is determined whether the obstacle meets the preset shape characteristics. If so, the next step of feature matching is performed, including: Calculate the relative error value based on the measured cross-echo path and the theoretical cross-echo path; Determine whether the relative error value is greater than a preset error threshold. If so, determine that the obstacle meets the preset morphological characteristics and proceed to the next step of feature matching.

[0011] This scheme is based on the characteristic that the echo intensity of rod / line objects is unevenly distributed along the axial direction due to reflection along the length direction. By calculating the relative error value between the measured cross echo path and the theoretical cross echo path, the scheme calculates the difference in echo amplitude and combines it with a preset error threshold to determine whether the obstacle meets the preset shape characteristics. If so, it is identified as a rod / line object and a preliminary screening judgment is performed. Since it uses the inherent physical characteristics of rod-shaped objects for identification, it is less affected by environmental clutter, has strong anti-interference ability, and high identification accuracy.

[0012] In a further implementation, the obstacle information for each frame of the obstacle is traversed, including: Based on the obstacle information collected by the ultrasonic radar in each frame, and combined with the installation location of the ultrasonic radar, the coordinates of the two measurable endpoints of the line segment where the obstacle is located in each frame are determined. Coordinate system transformation is performed based on the coordinate system transformation matrix to convert the coordinates of the measurable endpoints into the vehicle coordinate system; Based on the coordinates of the two measurable endpoints after coordinate system transformation, a straight line fit is performed to calculate the slope-intercept equation of the line segment where the obstacle is located in the current frame; The obstacle information includes measurable endpoint coordinates and the slope-intercept equation of a line.

[0013] Based on the three-dimensional structural characteristics of the limiting block, this solution performs multi-frame data acquisition. By acquiring the coordinates of the two measurable endpoints of the line segment where the obstacle is located in different frames, the data is aggregated into a dense set of pixels. This significantly improves the reliability of the limiting block contour recognition and effectively overcomes occlusion and noise interference in single-frame measurement, achieving a more complete and stable three-dimensional perception of the structural region.

[0014] In a further implementation, feature matching is performed based on the standard geometric features of the limiting block. If a match is found, coordinate information is output. The coordinate information is then integrated to obtain the limiting block information, including: Based on the standard geometric features of the limit block, the straight line equation features of the bottom edge of the parking space are obtained, and the feature parameters are extracted. Iterate through the slope-intercept form of the line segment containing the obstacle in each frame to obtain the equation parameters; Calculate the feature difference between the equation parameters and the feature parameters, and determine whether the feature difference is less than a preset feature threshold. If so, determine that the measurable endpoint coordinates are a limit block region. Obtain the coordinate information of the corresponding measurable endpoint coordinates, and integrate the coordinate information to obtain the limit block information.

[0015] In a further embodiment, the equation characteristic of the linear equation is: The characteristic parameters include slope. and intercept The equation of the slope-intercept form of the line is: The equation parameters include the slope k and the intercept b; the preset feature threshold includes a first feature threshold. Second feature threshold .

[0016] This scheme is based on the fundamental principle of two points forming a line. By calculating the slope-intercept form of the line segment where the obstacle is located in the current frame, the slope and intercept of the line equation on the corresponding line segments of the two measurable endpoints are directly obtained. These are then compared with the feature parameters of the line equation of the bottom edge of the parking space (the limit block is inside the parking space and at a fixed distance from the bottom edge). This allows for an intuitive and accurate determination of whether the corresponding obstacle pixel is in the typical position of the limit block, thereby effectively filtering out noise caused by environmental interference or false detection, and further improving the accuracy of limit block positioning and system robustness.

[0017] In a further implementation scheme, the method further includes: arranging a multi-sensor array, with two ultrasonic radars symmetrically positioned on the left and right sides of the rear bumper of the vehicle; collecting ambient noise, and dynamically adjusting a preset error threshold based on the ambient noise.

[0018] This solution uses two existing ultrasonic radars on the vehicle body for limit block recognition without incurring additional costs. Furthermore, by dynamically adjusting the preset error threshold based on environmental noise, it compensates for recognition deviations caused by temperature changes, eliminates environmental interference, and reduces the false judgment rate.

[0019] In a further implementation, the formula for calculating the theoretical cross-echo path is as follows: In the formula, Indicates the theoretical cross-echo path, , d represents the measured cross-echo paths acquired by the two ultrasonic radars; d is the distance between the two ultrasonic radars.

[0020] This solution uses geometric relationships to directly calculate the theoretical position of the cross echo based on the direct echo distance from two ultrasonic radars, which can effectively identify and verify real rod-shaped / linear objects, thereby improving the accuracy of detection and anti-interference capability.

[0021] In a further implementation scheme, the method also includes: verifying the limit block based on a PLA sensor, specifically as follows: A PLA sensor is installed at the rear of the vehicle to obtain the longitudinal distance between the vehicle and the obstacle collected by the PLA sensor, which is used as verification data. The longitudinal distance between the vehicle and the obstacle obtained from the limit block information is used as the data to be measured. Calculate the longitudinal difference between the verification data and the data to be tested, and determine whether the longitudinal difference is greater than a preset longitudinal threshold. If so, discard the limit block information and trigger a re-detection.

[0022] This solution further supplements the PLA sensor for limit block verification, which is used to collect the longitudinal distance (L) between the vehicle and the limit block to provide redundant distance verification information, cross-verify the reliability of ultrasonic data, and filter out abnormal distance values ​​caused by ultrasonic multipath reflection or environmental interference, so as to improve the recognition accuracy of the limit block.

[0023] The present invention also provides a limit block recognition system based on ultrasonic radar, for implementing the limit block recognition method based on ultrasonic radar as described above, characterized in that: it includes a main controller and a radar detection module connected thereto, the radar detection module including two ultrasonic radars symmetrically arranged on the left and right sides of the rear bumper of the vehicle.

[0024] In a further embodiment, a PLA sensor connected to the main controller is also included, the PLA sensor being arranged in the middle of the rear bumper of the vehicle, between the two ultrasonic radars. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating the process of a limit block identification method based on ultrasonic radar provided in an embodiment of the present invention. Figure 2 This is an installation diagram of the ultrasonic radar provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the measurement model provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the direct echo measurement model provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the equivalent model of indirect echo provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of an obstacle in an actual ultrasonic array system, as provided in an embodiment of the present invention. Figure 7 This is a schematic diagram of the vehicle coordinate system provided in an embodiment of the present invention; Figure 8 This is a schematic diagram of a parking space in the vehicle coordinate system provided in an embodiment of the present invention. Detailed Implementation

[0026] The embodiments of the present invention are described in detail below with reference to the accompanying drawings. The embodiments are given for illustrative purposes only and should not be construed as limiting the present invention. The accompanying drawings are for reference and illustration only and do not constitute a limitation on the scope of patent protection of the present invention, because many changes can be made to the present invention without departing from the spirit and scope of the present invention.

[0027] Example 1 This invention provides a method and system for identifying limit blocks based on ultrasonic radar, such as... Figures 1-8 As shown, this embodiment includes the following steps: S01. Arrange the multi-sensor array, placing two ultrasonic radars symmetrically on the left and right sides of the vehicle's rear bumper, such as... Figure 2 The positions of probe 1 and probe 2 in the diagram.

[0028] S02. Collect real-time weather data and dynamically adjust the preset error threshold based on the real-time weather. For example, on rainy days, the relative error value can be adjusted to ≤8%. The specific value can be obtained through experimental calibration, and this embodiment does not impose any restrictions.

[0029] In other embodiments, adaptive switching between high and low temperature modes is also included: based on temperature data collected by a temperature sensor, the ultrasonic propagation speed model (formula: v=331.5+0.6T, where T is the air temperature) is dynamically adjusted to compensate for ranging deviations caused by temperature changes.

[0030] Under extreme temperature conditions (such as -20°C or 50°C), the system will also activate a temperature drift correction algorithm to ensure that the ranging accuracy is not affected by the environment through historical data backtracking and real-time calibration.

[0031] This embodiment uses two existing ultrasonic radars on the vehicle body for limit block identification without incurring additional costs. Furthermore, by dynamically adjusting the preset error threshold based on environmental noise, it compensates for ranging deviations caused by temperature changes, eliminates environmental interference, and reduces the misjudgment rate.

[0032] S1. Obstacle data of obstacles in the parking space are collected based on ultrasonic radar detection. The obstacle data includes measured cross echo paths and at least one set of direct echo distances. Specifically, the vehicle sensor system includes at least two ultrasonic radars (PDC sensors), with four typically deployed in a vehicle. This embodiment uses two ultrasonic radars (PDC sensors) as an example. The PDC sensors are symmetrically arranged on the left and right sides of the rear bumper of the vehicle, with a fixed known spacing d. Therefore, during data acquisition, the two sets of direct echo distances of the limiting block can be directly collected. , ) and measured cross-echo path length ( ).

[0033] That is, after the vehicle enters the garage reversing mode, the PDC sensors on both sides of the rear bumper simultaneously emit ultrasonic waves and receive the direct echo distance from the limit block ( , ) and cross-echo distance ( ); S2. Construct a direct echo measurement model based on the echo signal, and then construct an equivalent model of indirect echo based on the direct echo measurement model, and calculate the theoretical cross echo path, including: S21. Establish a Cartesian coordinate system with the midpoint of the rear bumper of the vehicle as the origin (i.e., the midpoint of the line connecting the two sensors as the origin O). S22. Construct a direct echo measurement model: With the ultrasonic radar as the center, construct a circle on a plane rectangular coordinate system with the direct echo distance as the radius; The actual location of the obstacle could be anywhere within the sector measured on the circle. Therefore, the equivalent obstacle echo reflection point is located at the center of the sector during calculation. Equating the reflection point to the center of the sector is an engineering approximation method used to balance computational complexity and accuracy. This approximation introduces some system positioning error, but this error can be effectively suppressed and compensated for through subsequent multi-frame data fusion and least-squares fitting steps, thus meeting the final recognition accuracy requirements.

[0034] S23. Construct an equivalent model of indirect echo: Construct an ellipse on a plane rectangular coordinate system with the two ultrasonic radars as the focal points. From the definition of indirect echo, we know that The ultrasonic radar sensor 1 emits ultrasonic waves, and the ultrasonic radar sensor 2 receives the ultrasonic waves reflected by the obstacle to measure the distance. That is, the location of the obstacle should be within the focal point of the two sensors (ultrasonic radar). Let be any point on the ellipse with its major axis, and let the obstacle be in the area where the measurements of the two sensors intersect.

[0035] like Figure 5 The black rectangle represents the actual location of the obstacle, and the black circle represents the equivalent location of the obstacle. Let be the equivalent minor axis length of the ellipse containing the obstacle.

[0036] S24. Calculate the theoretical cross echo path.

[0037] Specifically, such as Figure 6 , Measurement of direct echo for ultrasonic radar sensor1 The equivalent position of the obstacle, Measurement of direct echo for ultrasonic radar sensor2 The equivalent position of the obstacle, The indirect echo measured when ultrasonic radar sensor1 transmits and ultrasonic radar sensor2 receives. The equivalent position of the obstacle.

[0038] Sensor1 coordinates:

[0039] Sensor2 coordinates:

[0040] Then in middle

[0041]

[0042]

[0043] In summary, solving the simultaneous equations yields:

[0044] but:

[0045] Therefore, in this embodiment, the calculation formula for the theoretical cross-echo path is as follows: In the formula, Indicates the theoretical cross-echo path, , d represents the measured cross-echo paths acquired by the two ultrasonic radars; d is the distance between the two ultrasonic radars.

[0046] This embodiment uses geometric relationships to directly calculate the theoretical position of the cross echo based on the direct echo distances fed back by two ultrasonic radars, which can effectively identify and verify real rod-shaped / linear objects, thereby improving the accuracy of detection and anti-interference ability.

[0047] S3. Based on the measured cross-echo path and the theoretical cross-echo path, determine whether the obstacle meets the preset shape characteristics. If so, proceed to the next step of feature matching, including: Calculate the relative error value based on the measured cross-echo path and the theoretical cross-echo path; Determine whether the relative error value is greater than a preset error threshold. If so, determine that the obstacle meets the preset morphological characteristics and proceed to the next step of feature matching.

[0048] The formula for calculating the relative error value is as follows:

[0049] like If the percentage is ≤5%, the target is determined to be a rod-shaped / linear object (meeting the geometric characteristics of the limiting block), and the process proceeds to the next step of feature matching. If the error is greater than 5%, it is judged as a non-target (such as a wall or columnar object, where the echo path does not satisfy the triangular relationship).

[0050] This embodiment is based on the principle that the echo intensity of rod / line objects is unevenly distributed along the axial direction due to reflection along the length direction. By calculating the relative error value between the measured cross echo path and the theoretical cross echo path, the difference in echo amplitude is used to determine whether the obstacle meets the preset shape characteristics. If so, it is identified as a rod / line object and a preliminary screening judgment is performed. Since it uses the inherent physical characteristics of rod-shaped objects for identification, it is less affected by environmental clutter, has strong anti-interference ability, and high identification accuracy.

[0051] S4. Traverse the obstacle information of each frame and perform feature matching based on the standard geometric features of the limit block. If the match is consistent, output the coordinate information and integrate the coordinate information to obtain the limit block information.

[0052] In this embodiment, traversing each frame of obstacle information includes: A1. Based on the obstacle information collected by the ultrasonic radar in each frame, and combined with the installation position of the ultrasonic radar, determine the coordinates of the two measurable endpoints of the line segment where the obstacle is located in each frame. like Figure 7 The coordinates of the two measurable endpoints of the line segment containing the obstacle: , .

[0053] A2. Based on the coordinate system transformation matrix, perform coordinate system transformation to convert the coordinates of the measurable endpoints into the vehicle coordinate system; Assuming a left-handed global coordinate system is established with the moment the vehicle begins to reverse as the origin and the direction of travel as the positive y-axis, the coordinates of the theoretical endpoint of the line segment where the ultimately identified rod-shaped object is located are then determined in the sensor coordinate system. Convert the position to the vehicle coordinate system. The coordinate system is as follows: Figure 7 : Assuming the global coordinates of ultrasonic radar sensor1 are in the global coordinate system... Global coordinates of ultrasonic radar sensor2 Then, the coordinates of any point in the sensor's local coordinate system under this global coordinate system can be obtained. To the coordinates of this global coordinate system The transformation relationship is as follows:

[0054] in,

[0055]

[0056]

[0057] The coordinates of the start and end points of the obstacle line segment can then be transformed to the global coordinate system. Let's assume the transformed coordinates are: ; ; Recorded as .

[0058] A3. Based on the coordinates of the two measurable endpoints after coordinate system transformation, perform line fitting to calculate the slope-intercept equation of the line segment where the obstacle is located in the current frame. The formula for the slope-intercept equation is as follows:

[0059] in:

[0060]

[0061] The obstacle information includes measurable endpoint coordinates and the slope-intercept equation of a line.

[0062] Based on the three-dimensional structural characteristics of the limiting block, this embodiment performs multi-frame data acquisition. By acquiring the coordinates of the two measurable endpoints of the line segment where the obstacle is located in different frames, the data is converged into a dense set of pixels. This significantly improves the reliability of the limiting block contour recognition and effectively overcomes occlusion and noise interference in single-frame measurement, achieving a more complete and stable three-dimensional perception of the structural region.

[0063] In this embodiment, feature matching is performed based on the standard geometric features of the limiting block. If a match is found, coordinate information is output. The coordinate information is then integrated to obtain the limiting block information, including: B1. Based on the standard geometric features of the limiting block, obtain the straight line equation features of the bottom edge of the parking space and extract the feature parameters; In this embodiment, the characteristic equation of the straight line equation is: The characteristic parameters include slope. and intercept .

[0064] For example, the equation of the straight line where the parking space baseline is located, identified by other sensing methods (such as a vehicle surround view camera system or a reversing camera system), in the global coordinate system.

[0065] B2. Traverse the slope-intercept form of the line segment containing the obstacle in each frame and obtain the equation parameters; In this embodiment, the equation of the slope-intercept form of the line is: The parameters of the equation include the slope k and the intercept b.

[0066] As the vehicle moves and multiple frames of data accumulate, the starting and ending coordinates (theoretical endpoints) of the target output in step A3 are determined. and the slope-intercept form of the line. It is compared with a pre-stored feature library to determine its position in the parking space.

[0067] B3. Calculate the feature difference between the equation parameters and the feature parameters, and determine whether the feature difference is less than a preset feature threshold. If so, determine that the measurable endpoint coordinates are the limit block region. See Figure 8 The black hollow rectangles are parking space frames, and the black solid rectangles are parking limit blocks.

[0068] The preset feature threshold includes a first feature threshold. Second feature threshold The logic for determining whether a region is a limit block is as follows:

[0069] Wherein, constant c is the typical position parameter of the bottom edge of the compensation parking space and the corresponding limit block behind the parking space. and To determine whether an identified obstacle is a limit block, the slope threshold and intercept threshold are obtained through calibration. Based on this, it is determined whether the measured obstacle position is the location of a typical limit block. If the difference is too large, the corresponding pixel / region is identified as noise and deleted; if the difference is within a certain range, the obstacle is determined to be a limit block, and the obstacle endpoint is added to the limit block library.

[0070] B4. Obtain the coordinate information of the corresponding measurable endpoint coordinates, and integrate the coordinate information to obtain the limit block information.

[0071] Specifically, after a period of measurement, the obtained obstacle database contains the endpoints of n obstacles, which are respectively By fitting the line using least squares, a new line is obtained as follows: This process continues, continuously outputting the line where the newly generated limit block is located, as the limit block information.

[0072] This embodiment is based on the fundamental principle of two points forming a line. By calculating the slope-intercept form of the line segment where the obstacle is located in the current frame, the slope and intercept of the line equation on the corresponding line segments of the two measurable endpoints are directly obtained. This is then compared with the feature parameters of the line equation of the bottom edge of the parking space (the limit block is inside the parking space and at a fixed distance from the bottom edge). This allows for an intuitive and accurate determination of whether the corresponding obstacle pixel is at the typical position of the limit block, thereby effectively filtering out noise caused by environmental interference or false detection, and further improving the accuracy of the limit block positioning and the robustness of the system.

[0073] In this embodiment, it also includes: S5. Limit block verification based on PLA sensor, specifically: S51. Install a PLA sensor on the rear side of the vehicle and obtain the longitudinal distance between the vehicle and the obstacle collected by the PLA sensor as verification data. S52. The longitudinal distance between the vehicle and the obstacle obtained based on the limit block information is used as the data to be measured. S53. Calculate the longitudinal difference between the verification data and the test data, and determine whether the longitudinal difference is greater than the preset longitudinal threshold. If so, discard the limit block information and trigger re-detection.

[0074] For example, the preset longitudinal threshold is selected as 0.1m according to the requirements. When it is greater than 0.1m, the limit block information is discarded and a re-detection is triggered.

[0075] Prior to S51, dynamic threshold adaptation for vehicle speed was also possible, specifically: PLA adaptation verification based on vehicle speed: When the vehicle speed is less than 5 km / h (typical reversing scenario), the distance measurement error threshold of longitudinal distance L is relaxed to ±0.05m to avoid misjudgment due to slight vehicle movement. When the vehicle speed is ≥5km / h, the standard error threshold of ±0.02m is restored to ensure recognition accuracy at high speeds.

[0076] In other embodiments, acceleration sensor data may be introduced as needed, and the error threshold may be dynamically adjusted to adapt to measurement fluctuations caused by instantaneous speed changes when the vehicle is in a state of rapid acceleration or deceleration.

[0077] In this embodiment, it also includes: S6, time consistency verification, specifically: determining whether the same limit block is continuously detected within 10 seconds, otherwise it is regarded as an invalid target.

[0078] In other embodiments, step S6 of this embodiment further includes: introducing a historical data backtracking function. When it is determined that the target has been marked as valid in multiple detections, the position fluctuation is further calculated. If the position fluctuation is large, the anomaly analysis module is activated to investigate sensor failures or environmental interference factors.

[0079] This embodiment further supplements the PLA sensor for limit block verification, which is used to collect the longitudinal distance (L) between the vehicle and the limit block to provide redundant distance verification information, cross-verify the reliability of ultrasonic data, and filter out abnormal distance values ​​caused by ultrasonic multipath reflection or environmental interference, so as to improve the recognition accuracy of the limit block.

[0080] S7. Execute emergency warning when a system fault is detected.

[0081] The specific principles of emergency early warning are as follows: I. Sensor Failure Handling: If a single PDC sensor (ultrasonic radar) fails, the system automatically switches to single-sensor mode, predicting the position of the limit block based on historical data, but the accuracy drops to ±0.1m, and an alarm is triggered.

[0082] In scenarios where the dual PDC sensors (ultrasonic radar) fail, the system will attempt to use the PLA sensor alone and combine it with the vehicle motion model to estimate the position. Although the accuracy is further reduced to ±0.15m, it can still provide basic obstacle avoidance functions.

[0083] II. Extreme Environment Response: In scenarios such as strong winds and heavy rain, if the system fails to detect parking five times in a row, it will suspend the automatic parking function and switch to manual intervention mode. At the same time, the system will display "Environmental interference, please operate manually" through the HMI.

[0084] Steps S5, S6, and S7 are optional verification procedures, and are not limited in this embodiment.

[0085] Example 2 This invention also provides a limit block recognition system based on ultrasonic radar, used to implement the limit block recognition method based on ultrasonic radar as described above. The system is characterized by including a main controller and a radar detection module connected thereto, wherein the radar detection module includes two ultrasonic radars symmetrically arranged on the left and right sides of the rear bumper of the vehicle.

[0086] In this embodiment, a PLA sensor connected to the main controller is also included. The PLA sensor is arranged in the middle of the rear bumper of the vehicle, between the two ultrasonic radars.

[0087] It also includes an environmental risk assessment module connected to the main controller. When the environmental risk assessment module detects continuous severe weather, it will automatically extend the detection interval, reduce the probability of false triggering, and provide a detailed environmental status report on the HMI interface to help the driver make decisions.

[0088] This invention employs multiple ultrasonic radars to collaboratively collect direct echo data and cross-echo data. Based on the measured and theoretical cross-echo paths, it determines whether obstacles meet preset morphological characteristics, i.e., it verifies the preset morphological characteristics using the triangulation principle for initial screening. Furthermore, through limit block feature matching, it analyzes the distribution and location characteristics of obstacles for secondary positioning and screening, thereby achieving accurate identification of low and small limit blocks. Through limit block pose detection, the safety and reliability of automatic parking are further improved.

[0089] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. A method for identifying limit blocks based on ultrasonic radar, characterized in that, Including the following steps: Obstacle data on obstacles in parking spaces are collected based on ultrasonic radar detection. The obstacle data includes measured cross echo paths and at least one set of direct echo distances. A direct echo measurement model is constructed based on the echo signal, and then an equivalent model of indirect echo is constructed based on the direct echo measurement model to calculate the theoretical cross echo path. Based on the measured cross-echo path and the theoretical cross-echo path, determine whether the obstacle meets the preset shape characteristics; if so, proceed to the next step of feature matching. Traverse the obstacle information of each frame and perform feature matching based on the standard geometric features of the limit block. If the match is consistent, output the coordinate information and integrate the coordinate information to obtain the limit block information. Among them, the direct echo measurement model is constructed: with the ultrasonic radar as the center, a circle is constructed on the plane rectangular coordinate system with the direct echo distance as the radius; Construct an equivalent model of indirect echo: Construct an ellipse on a Cartesian coordinate system with the two ultrasonic radars as the focal points; in, Measurement of direct echo for ultrasonic radar sensor1 The equivalent position of the obstacle, Measurement of direct echo for ultrasonic radar sensor2 The equivalent position of the obstacle, The indirect echo measured when ultrasonic radar sensor1 transmits and ultrasonic radar sensor2 receives. The equivalent position of the obstacle; Sensor1 coordinates: Sensor2 coordinates: Then in middle In summary, solving the simultaneous equations yields: Therefore, the formula for calculating the theoretical cross-echo path is as follows: In the formula, Indicates the theoretical cross-echo path, , d represents the direct echo distance collected by the two ultrasonic radars; d is the distance between the two ultrasonic radars.

2. The method for identifying limit blocks based on ultrasonic radar as described in claim 1, characterized in that, Based on the measured cross-echo path and the theoretical cross-echo path, determine whether the obstacle meets the preset shape characteristics. If so, proceed to the next step of feature matching, including: Calculate the relative error value based on the measured cross-echo path and the theoretical cross-echo path; Determine whether the relative error value is greater than a preset error threshold. If so, determine that the obstacle meets the preset morphological characteristics and proceed to the next step of feature matching.

3. The limit block identification method based on ultrasonic radar as described in claim 2, characterized in that, Traversing each frame of obstacle information, including: Based on the obstacle information collected by the ultrasonic radar in each frame, and combined with the installation location of the ultrasonic radar, the coordinates of the two measurable endpoints of the line segment where the obstacle is located in each frame are determined. Coordinate system transformation is performed based on the coordinate system transformation matrix to convert the coordinates of the measurable endpoints into the vehicle coordinate system; Based on the coordinates of the two measurable endpoints after coordinate system transformation, a straight line fit is performed to calculate the slope-intercept equation of the line segment where the obstacle is located in the current frame; The obstacle information includes measurable endpoint coordinates and the slope-intercept equation of a line.

4. The limit block recognition method based on ultrasonic radar as described in claim 3, characterized in that, Based on the standard geometric features of the limiting block, feature matching is performed. If a match is found, coordinate information is output. The coordinate information is then integrated to obtain the limiting block information, including: Based on the standard geometric features of the limit block, the straight line equation features of the bottom edge of the parking space are obtained, and the feature parameters are extracted. Iterate through the slope-intercept form of the line segment containing the obstacle in each frame to obtain the equation parameters; Calculate the feature difference between the equation parameters and the feature parameters, and determine whether the feature difference is less than a preset feature threshold. If so, determine that the measurable endpoint coordinates are a limit block region. Obtain the coordinate information of the corresponding measurable endpoint coordinates, and integrate the coordinate information to obtain the limit block information.

5. The limit block identification method based on ultrasonic radar as described in claim 4, characterized in that: The characteristic equation of the linear equation is: The characteristic parameters include slope. and intercept The equation of the slope-intercept form of the line is: The equation parameters include the slope k and the intercept b; the preset feature threshold includes a first feature threshold. Second feature threshold .

6. The method for identifying limit blocks based on ultrasonic radar as described in claim 2, characterized in that, Also includes: A multi-sensor array layout was adopted, with two ultrasonic radars symmetrically arranged on the left and right sides of the rear bumper of the vehicle. Collect ambient noise and dynamically adjust a preset error threshold based on the ambient noise.

7. The limit block identification method based on ultrasonic radar as described in claim 4, characterized in that, Also includes: Limit block verification based on PLA sensors is specifically as follows: A PLA sensor is installed at the rear of the vehicle to obtain the longitudinal distance between the vehicle and the obstacle collected by the PLA sensor, which is used as verification data. The longitudinal distance between the vehicle and the obstacle obtained from the limit block information is used as the data to be measured. Calculate the longitudinal difference between the verification data and the data to be tested, and determine whether the longitudinal difference is greater than a preset longitudinal threshold. If so, discard the limit block information and trigger a re-detection.

8. A limit block recognition system based on ultrasonic radar, used to implement the limit block recognition method based on ultrasonic radar as described in any one of claims 1 to 7, characterized in that: It includes a main controller and a radar detection module connected thereto, wherein the radar detection module includes two ultrasonic radars symmetrically arranged on the left and right sides of the rear bumper of the vehicle.

9. The limit block recognition system based on ultrasonic radar as described in claim 8, characterized in that: It also includes a PLA sensor connected to the main controller, the PLA sensor being arranged in the middle of the rear bumper of the vehicle, between the two ultrasonic radars.

Images