On-line calibration method and device for external parameters of vehicle-mounted radar and automatic driving vehicle

Abstract

The invention provides an external parameter online calibration method and device of a vehicle-mounted radar and an automatic driving vehicle, and relates to the technical field of computers, in particular to the technical fields of automatic driving, multi-sensor fusion, external parameter calibration of the vehicle-mounted radar and the like. According to the specific implementation scheme, the method comprises the following steps: extracting a static target from a data frame obtained by a vehicle-mounted radar; in a world coordinate system, calibrating radar external parameters of the vehicle-mounted radar by minimizing a speed residual error between a radar speed measurement value and a reference speed value of the static target; wherein the reference speed value is the speed of the static target determined based on pre-fusion data; the radar speed measurement value is converted into the world coordinate system through the radar external parameters.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and in particular to the fields of autonomous driving, multi-sensor fusion, and external parameter calibration of vehicle radar. Background Technology

[0002] In the field of driving, vehicle radar, as one of the core sensors of the vehicle's environmental perception system, can detect obstacles in the vehicle's surrounding environment in real time and provide some key parameters of the obstacles, such as distance and speed, providing key basis for vehicle driving decisions and safety control.

[0003] However, if the external parameters of the vehicle radar become inaccurate, it will lead to deviations in the calculation of parameters such as the position and speed of obstacles, which in turn will reduce the environmental perception accuracy of the multi-sensor fusion system and ultimately affect the reliability and safety of vehicle driving decisions and safety control. Summary of the Invention

[0004] This disclosure provides a method, apparatus, and autonomous vehicle for online calibration of extrinsic parameters of vehicle-mounted radar.

[0005] According to one aspect of this disclosure, an online extrinsic parameter calibration method for vehicle-mounted radar is provided, comprising: Static targets are extracted from data frames acquired by vehicle-mounted radar; In the world coordinate system, the radar extrinsic parameters of the vehicle-mounted radar are calibrated by minimizing the velocity residual between the radar velocity measurement value and the reference velocity value of the static target; where the reference velocity value is the velocity of the static target determined based on the pre-fusion data; the radar velocity measurement value is transformed into the world coordinate system through the radar extrinsic parameters.

[0006] According to another aspect of this disclosure, an online extrinsic parameter calibration device for vehicle-mounted radar is provided, comprising: The extraction module is used to extract static targets from the data frames acquired by the vehicle-mounted radar; The calibration module is used to calibrate the radar extrinsic parameters of the vehicle-mounted radar in the world coordinate system by minimizing the velocity residual between the radar velocity measurement value and the reference velocity value of the static target; wherein, the reference velocity value is the velocity of the static target determined based on the pre-fusion data; the radar velocity measurement value is transformed to the world coordinate system through the radar extrinsic parameters.

[0007] According to another aspect of this disclosure, an electronic device is provided, comprising: At least one processor; and The memory is communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform any of the methods described in the present disclosure.

[0008] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are used to cause the computer to perform any of the methods according to embodiments of this disclosure.

[0009] According to another aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements any of the methods according to embodiments of this disclosure.

[0010] According to another aspect of this disclosure, a vehicle is provided, including the electronic equipment provided in this disclosure.

[0011] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0012] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein: Figure 1 This is a flowchart illustrating an online external parameter calibration method for vehicle-mounted radar according to an embodiment of the present disclosure; Figure 2 This is a schematic diagram of a process for extracting static targets from data frames acquired by vehicle-mounted radar according to an embodiment of the present disclosure; Figure 3 This is a flowchart illustrating the process of determining a candidate target as a static target according to an embodiment of the present disclosure; Figure 4 This is a schematic diagram of a process for filtering candidate targets based on credibility according to an embodiment of the present disclosure; Figure 5 This is a schematic diagram of the structure of an online external parameter calibration device for vehicle-mounted radar according to an embodiment of the present disclosure; Figure 6 This is a block diagram of an electronic device used to implement the online external parameter calibration method for vehicle-mounted radar according to embodiments of the present disclosure. Detailed Implementation

[0013] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0014] The terms “first,” “second,” etc., used in this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion, such as including a series of steps or units. A method, system, product, or apparatus is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or apparatuses.

[0015] It should be noted that, unless it is explicitly stated that there is a sequential order of execution between different operations, or that there is a sequential order of execution between different operations in terms of technical implementation, the execution order between multiple operations may not be significant, and multiple operations may be executed simultaneously.

[0016] The extrinsic parameters of vehicle-mounted radar describe the spatial transformation relationship between the radar coordinate system and the vehicle coordinate system (such as an IMU (Inertial Measurement Unit) cell). Its core parameters typically include a rotation matrix describing the attitude mapping and a translation vector characterizing the position offset. Accurate calibration of these parameters is crucial for the accurate processing and subsequent applications of radar data.

[0017] If there are errors in the radar extrinsic parameters, it will cause deviations in the target position and speed information detected by the radar, thereby affecting the performance and safety of the entire vehicle in driving scenarios.

[0018] In real-world driving scenarios, the vehicle's motion is constantly changing due to the complex and ever-changing environment. Traditional offline calibration methods often struggle to adapt to these complex environments, causing the calibrated extrinsic parameters to gradually deviate from their true values, thus introducing new measurement errors.

[0019] In other words, the extrinsic parameters of automotive radar will slowly accumulate some errors after offline calibration. Especially in scenarios where the vehicle is in motion, these errors need to be corrected through online calibration to remove the deficiencies of offline calibration.

[0020] In view of this, this disclosure provides an online calibration method for the extrinsic parameters of vehicle-mounted radar. By optimizing the extrinsic parameters of the vehicle-mounted radar, the accuracy and reliability of the data provided by the vehicle-mounted radar are ensured, thereby improving the stability and safety of vehicle driving.

[0021] like Figure 1 The diagram shown is a flowchart illustrating the online external parameter calibration method for vehicle-mounted radar provided in this disclosure, including the following: S101, extract static targets from data frames acquired by vehicle-mounted radar.

[0022] Vehicle-mounted radar is a sensor installed on a vehicle to detect information about its surrounding environment. Common types include millimeter-wave radar and lidar. Vehicle-mounted radar periodically collects data about the surrounding environment and organizes this data into data frames.

[0023] A data frame is the basic unit for transmitting data by an onboard radar, containing relevant information about all targets detected by the radar at a given moment. After analysis and processing, the data frames of the onboard radar can be used to obtain obstacle information, such as target ID (Identifier), distance, speed, and angle.

[0024] Static targets are those with zero or near-zero velocity in the world coordinate system, such as utility poles, buildings, and road signs.

[0025] Static targets are extracted from the data frames, with each static target corresponding to multiple points in the data frame of the vehicle-mounted radar.

[0026] S102, in the world coordinate system, the radar extrinsic parameters of the vehicle-mounted radar are calibrated by minimizing the velocity residual between the radar velocity measurement value and the reference velocity value of the static target; wherein, the reference velocity value is the velocity of the static target determined based on the pre-fusion data; the radar velocity measurement value is transformed into the world coordinate system through the radar extrinsic parameters.

[0027] The world coordinate system is a unified, fixed coordinate system used to describe the relative positions of a vehicle and all objects in its surrounding environment.

[0028] Radar speed measurement value refers to the speed of a target object relative to the radar, which is directly measured by the vehicle-mounted radar. It is the speed representation in the radar coordinate system.

[0029] The baseline velocity value is the velocity of a static target in the world coordinate system determined based on pre-fusion data. Pre-fusion data refers to the data obtained by fusing data from multiple sensors (such as cameras, LiDAR, and inertial measurement units) in an early stage. By fusing data from multiple sensors, the velocity of the static target in the world coordinate system can be estimated more accurately and used as the baseline velocity.

[0030] Velocity residual refers to the difference between the radar velocity measurement value of the same static target and its reference velocity value after being transformed to the world coordinate system using radar extrinsic parameters. During the calibration of vehicle-mounted radar extrinsic parameters, minimizing the velocity residual allows the radar velocity measurement value to be as close as possible to the reference velocity, thereby determining the optimal radar extrinsic parameters.

[0031] In practice, the radar extrinsic parameters of the vehicle-mounted radar are calibrated by minimizing the velocity residual between the radar velocity measurement value and the reference velocity value of the static target. This can be described by expression (1): (1) In expression (1), i represents the i-th static target. There are a total of N static targets, where N is a positive integer greater than 1. These N static targets can come from the same data frame of the vehicle radar or from multiple data frames of the vehicle radar. This represents the rotation matrix in the calibrated radar extrinsic parameters; This represents the translation vector in the calibrated radar extrinsic parameters; This means finding the optimal radar extrinsic parameters for the vehicle-mounted radar to minimize the velocity residual between the radar velocity measurement value and the reference velocity value of a static target; This represents the summation of N static targets; ||·||2 represents the Euclidean distance. This represents the reference velocity value of the i-th static target; This represents the Jacobian matrix, used to optimize radar extrinsic parameters; This represents the speed of the i-th static target in the world coordinate system, as measured by radar velocity.

[0032] In this embodiment, extracting static targets from the vehicle-mounted radar data frames allows for focusing on relatively stable and easily analyzable target objects, minimizing complex interference factors caused by dynamic targets. Minimizing velocity residuals leverages the comprehensive advantages of pre-fused data to ensure the accuracy and reliability of the reference velocity value. Transforming radar velocity measurements to a world coordinate system using radar extrinsic parameters allows for comparison of radar velocity measurements and reference velocity values ​​within the same coordinate system, enabling more precise adjustment of radar extrinsic parameters and improving the accuracy of vehicle-mounted radar measurements, thereby enhancing the reliability of the vehicle-mounted radar in real-world driving scenarios. This method is suitable for radar extrinsic parameter calibration in vehicle operation scenarios and enables real-time online calibration.

[0033] In this embodiment of the disclosure, the radar velocity measurement value is transformed to the world coordinate system through radar extrinsic parameters, which can be achieved based on the following steps: Step A1: Based on the radar extrinsic parameters, the radar velocity measurement value is transformed into the world coordinate system to obtain the intermediate value; Step A2: Multiply the intermediate value by the Jacobian matrix to obtain the speed of the radar velocity measurement in the world coordinate system.

[0034] During implementation, the velocity of the radar velocity measurement value in the world coordinate system can be described by expression (2): (2) In expression (2), This represents the speed of the radar velocity measurement in the world coordinate system. Represents the Jacobian matrix; This represents an intermediate value obtained by converting radar speed measurements to the world coordinate system based on radar extrinsic parameters. In practice, the radar speed measurements are converted to the world coordinate system based on the radar extrinsic parameters and the vehicle's pose in the world coordinate system.

[0035] The Jacobian matrix can be described by expression (3): (3) In expression (3), Represents the Jacobian matrix; This represents the transformation matrix from the Novatel coordinate system (which can be understood as the vehicle coordinate system) to the world coordinate system, used to transform target points in the Novatel coordinate system to the world coordinate system; This represents the transformation matrix from the radar coordinate system to the Novatel coordinate system, used to transform target points in the radar coordinate system to the Novatel coordinate system. This indicates the coordinates of the target point in the radar coordinate system; The symbol represents the partial derivative, used to calculate the partial derivative with respect to a variable.

[0036] In this embodiment, the radar velocity measurement value is transformed to the world coordinate system using radar extrinsic parameters to obtain an intermediate value, thus mapping the radar velocity measurement value to a unified world coordinate system. This intermediate value is then multiplied by the Jacobian matrix. The Jacobian matrix J is the partial derivative of the velocity residual with respect to the optimization variable (R, t), and its physical meaning is: when the extrinsic parameters undergo a small perturbation, along what direction and at what rate will the velocity residual change? By introducing the Jacobian matrix, the problem can be transformed from a difficult brute-force search into a "fast-converging nonlinear least squares" solution, thereby obtaining a more accurate velocity measurement value in the world coordinate system.

[0037] In this embodiment of the disclosure, static targets are extracted from the data frames acquired by the vehicle-mounted radar, such as... Figure 2 As shown, it includes the following: S201, Obtain the initial speed measurement value of the candidate target in the data frame by the vehicle radar.

[0038] Candidate targets are potential static targets selected from the data frames. These candidate targets may require further analysis to confirm whether they are indeed static targets. These candidate targets are objects detected in the radar's data frames.

[0039] The initial velocity value of the candidate target is the velocity of the candidate target relative to the radar, which is directly measured by the vehicle-mounted radar.

[0040] When pulse Doppler radar measures the radial velocity of a target, the periodicity of Doppler frequency sampling causes repetitive ambiguity in the velocity measurement values. In simpler terms, this is the maximum range of target velocities that the radar can unambiguously distinguish; this range can be called the radar's velocity ambiguity period.

[0041] Therefore, the initial test values ​​measured by the vehicle-mounted radar may have some deviation, and thus need to be corrected.

[0042] S202 generates multiple hypothetical speeds based on the initial speed measurement value and the speed ambiguity period of the vehicle radar.

[0043] During implementation, based on the initial speed measurement value and the speed ambiguity period of the vehicle radar, multiple hypothetical speeds are generated, which can be described by expression (4): (4) In expression (4), This represents the k-th assumed velocity; Indicates actual speed; Indicates the initial speed measurement value; The velocity ambiguity period of the vehicle-mounted radar is represented by k; k represents the hypothesis index.

[0044] Specifically, when k=-1, the actual speed is one speed ambiguity period smaller than the initial speed measurement value; when k=0, the initial speed measurement value is the actual speed; and when k=1, the actual speed is one speed ambiguity period larger than the initial speed measurement value.

[0045] S203, based on the prior velocity of the candidate target, determine the probability that multiple hypothetical velocities are the actual velocities; the prior velocity is the velocity predicted based on the historical trajectory of the candidate target.

[0046] The prior velocity of the candidate target is generated based on the candidate target's historical trajectory information. During implementation, the prior velocity of the candidate target at the current moment can be predicted using a prediction algorithm.

[0047] Taking Kalman filtering as an example, Kalman filtering is an optimal recursive filter used to estimate the state of a system. It can optimally estimate the true state of the system based on a series of noisy measurements. Its core logic is to achieve the optimal estimation of the target state through an iterative process of "prediction-update".

[0048] In practice, a state model can be established based on the historical trajectory data of the candidate target. The optimal velocity estimate of the previous moment can be used to predict the prior velocity value of the current moment. At the same time, an observation model can be established by combining the measurement noise characteristics of the vehicle-mounted radar. The above prediction results can be iteratively corrected and error compensated by the measurement data of the candidate target velocity by the vehicle-mounted radar in the historical frames. Finally, the prior velocity of the candidate target at the current moment can be output.

[0049] In practice, the probability that multiple hypothetical velocities are the actual velocities can be determined by expression (5): (5) In expression (5), This represents the probability that the assumed speed under the k-th assumption is the actual speed. Indicates the prior velocity of the candidate target; This represents the assumed velocity under the k-th assumption; The variance of velocity; The kernel of the Gaussian likelihood function is used to represent the similarity between the assumed velocity and the prior velocity. Let represent the Gaussian likelihood values ​​for all hypotheses j; where j is the index of the hypothesis, corresponding to the same set of hypotheses as k.

[0050] S204. Select the hypothetical velocity with the highest probability as the actual velocity of the candidate target.

[0051] S205, if the actual velocity of the candidate target satisfies the static condition, the candidate target is determined to be a static target.

[0052] That is, after obtaining the probability that the assumed velocity under each assumption is the true velocity, the assumed velocity with the highest probability is selected as the true velocity of the candidate target. When the true velocity of the candidate target satisfies the stationary condition, the candidate target is determined as a static target. These static targets can be used for vehicle-mounted radar extrinsic parameter calibration tasks.

[0053] In this embodiment, multiple hypothetical velocities are generated based on initial velocity measurements and velocity ambiguity periods, comprehensively considering the possible true velocities of candidate targets. By utilizing the prior velocities of candidate targets, the probability that each hypothetical velocities corresponds to a true velocity is determined. This, combined with the motion history information of candidate targets, makes the judgment of hypothetical velocities more reasonable. Selecting the hypothetical velocity with the highest probability as the true velocity of the candidate target ensures that the obtained true velocity is closest to the actual situation. When the true velocity of a candidate target satisfies the condition of being stationary, it is determined to be a static target, which improves the accuracy of static target extraction and thus enhances the accuracy of radar extrinsic parameter calibration operations.

[0054] In this embodiment of the disclosure, if the actual velocity of the candidate target satisfies the stationary condition, the candidate target is determined to be a static target, such as... Figure 3 As shown, it includes the following: S301, obtain the vehicle's jerkiness.

[0055] During implementation, the jerkiness of the vehicle can be expressed by expression (6): (6) In expression (6), Indicates the vehicle's swiftness; This represents the vehicle acceleration output by the IMU (Inertial Measurement Unit) at time t. This represents the vehicle acceleration output by the IMU at time t-1; This represents the absolute value of the acceleration difference between adjacent moments; Indicates the time interval between two frames; This indicates the urgency threshold, which can be set according to actual needs.

[0056] During implementation, the relationship between the vehicle's jerkiness and the jerkiness threshold can be determined first: if the vehicle's jerkiness is greater than the jerkiness threshold, the vehicle's speed can be considered inaccurate, and the operation of compensating for the true speed based on the vehicle's speed can be skipped. Otherwise, proceed with S302.

[0057] S302, when the vehicle's jerkiness is less than the jerkiness threshold, compensates for the actual speed based on the vehicle's speed to obtain the calibrated speed.

[0058] Calibration speed refers to the speed value obtained by correcting the true speed of the candidate target using the vehicle's own speed. Its core function is to eliminate the interference of the vehicle's own motion state on the measurement of the true speed of the candidate target, so that the obtained true speed of the candidate target is more consistent with the actual motion characteristics of the candidate target in the world coordinate system, thereby providing an accurate basis for subsequent static target judgment.

[0059] S303, when the calibration speed is less than the speed threshold, determine that the true speed of the candidate target meets the stationary condition, and identify the candidate target as a static target.

[0060] The velocity threshold is a preset value used to determine whether a candidate target is stationary. If the calibrated velocity is less than the threshold, it indicates that the candidate target's velocity in the world coordinate system is sufficiently small, and it can be considered stationary; conversely, if the calibrated velocity is greater than or equal to the threshold, the candidate target is determined to be a moving target. The velocity threshold can be set according to the actual application scenario, and this disclosure does not limit it.

[0061] In this embodiment, the vehicle's jerkiness is acquired. Jerkiness reflects the rate of change of vehicle acceleration; a lower jerkiness indicates a relatively stable vehicle motion. When the vehicle's jerkiness is less than a threshold, a calibrated speed is obtained by compensating the candidate target's true speed based on the vehicle's speed. This eliminates the influence of the vehicle's own motion on the candidate target's true speed, making the candidate target's true speed more accurate. When the calibrated speed is less than a speed threshold, determining that the candidate target's true speed meets the stationary condition and classifying it as a static target more effectively eliminates misjudgments caused by vehicle motion and measurement errors. It identifies the true static target from numerous candidate targets, providing more accurate static target data for the vehicle-mounted radar's extrinsic parameter calibration operation, thereby improving the reliability of the vehicle-mounted radar.

[0062] In some embodiments, the calibrated speed is obtained by compensating the actual speed based on the vehicle's speed. Alternatively, the calibrated speed can be obtained by subtracting the vehicle's speed from the actual speed.

[0063] In practice, the calibrated speed is obtained by subtracting the vehicle's speed from the actual speed, and can be described by expression (7): (7) In expression (7), This indicates the obtained calibration speed; This represents the true velocity of the candidate target; Indicates the speed of the vehicle.

[0064] In this embodiment, the calibration speed is obtained by subtracting the vehicle's speed from the candidate target's actual speed. This eliminates interference from vehicle motion, allowing the calibration speed to more accurately reflect the actual motion of the candidate target. This improves the reliability of subsequent determinations based on the calibration speed to determine whether a target is static.

[0065] In other embodiments, the calibrated speed is obtained by compensating the real speed based on the vehicle's speed, which can also be achieved based on the following steps: Step B1: Map the jerkiness to the weight value range to obtain the weight value corresponding to the jerkiness. In implementation, any one of the following methods can be used to map swiftness to a weight value of 0 to 1: linear threshold, cosine soft switching, or sigmoid exponential decay.

[0066] Taking the linear threshold as an example, the weight value corresponding to the jerkiness can be obtained and described by expression (8): (8) In expression (8), This represents the weight value corresponding to the obtained jerkiness; Indicates the current level of urgency; This represents the fully credible threshold, used to indicate the lower limit critical value of jirkiness; This represents the completely unreliable threshold, used to indicate the upper limit critical value of urgency.

[0067] Step B2: The compensation value is obtained by multiplying the weight value by the vehicle speed; Step B3: Subtract the compensation value from the actual speed to obtain the calibrated speed.

[0068] In practice, the calibrated speed is obtained by subtracting the compensation value from the actual speed, which can be described by expression (9): (9) In expression (9), This indicates the obtained calibration speed; This represents the true velocity of the candidate target; This represents the weight value corresponding to the degree of urgency; Indicates the vehicle's speed; This represents the compensation value obtained.

[0069] In this embodiment, the jerkiness is mapped to a weight range to obtain a corresponding weight value, allowing for flexible adjustment of the compensation level based on the magnitude of the vehicle's jerkiness. A higher jerkiness indicates unstable vehicle movement, resulting in a larger weight value and greater compensation. Conversely, a lower jerkiness indicates more stable vehicle movement, leading to a smaller weight value and reduced compensation. The compensation value is obtained by multiplying this weight value by the vehicle speed, and the calibration speed is obtained by subtracting the compensation value from the actual speed. This comprehensive approach considers the dynamic characteristics of vehicle movement, thereby more accurately determining whether a candidate target is a static target.

[0070] In rainy or foggy weather, environmental factors can cause significant interference and misidentification of candidate targets detected by vehicle-mounted radar. For example, high humidity and suspended particulate matter can significantly scatter radar signals, causing the radar to misidentify rain and fog echoes or environmental clutter as candidate targets. To avoid this problem, target credibility assessment can be used to filter out interfering targets that are prone to causing errors, such as… Figure 4 As shown, it includes the following: S401, in rainy or foggy weather, acquire the radar point amplitude and initial speed measurement value of the vehicle-mounted radar for the candidate target.

[0071] Radar point amplitude refers to the intensity of the target echo signal received by the radar. It reflects the target's ability to reflect radar electromagnetic waves and is usually related to factors such as the target's size, shape, material, and the distance between the target and the radar.

[0072] The initial speed measurement value is the speed of the candidate target relative to the radar, which is directly measured by the vehicle-mounted radar.

[0073] It should be noted that vehicle-mounted radar typically measures the relative velocity of a target (i.e., Doppler velocity) based on the Doppler effect: after the electromagnetic waves emitted by the radar are reflected by the target, the frequency of the reflected waves will produce a Doppler frequency shift; by measuring this frequency shift and combining it with radar system parameters, such as the frequency of the emitted wave, the radial velocity of the target along the radar's line of sight can be calculated.

[0074] In this embodiment of the disclosure, the radial velocity of the candidate target calculated by the radar can be used as the initial velocity measurement value of the candidate target.

[0075] S402 determines the credibility of candidate targets based on radar point amplitude and initial speed measurement value; credibility is positively correlated with radar point amplitude and negatively correlated with target speed difference, which is the absolute difference between the initial speed measurement value and the vehicle speed.

[0076] During implementation, the credibility of candidate targets is determined based on radar point amplitude and initial velocity values, which can be achieved through the following steps: Step C1: Determine the mean of the normalized values ​​of the radar point amplitudes of multiple radar points of the candidate target to obtain the first parameter; Normalization maps the amplitude values ​​of radar points to a specific range, such as [0, 1], to eliminate differences in amplitude values ​​caused by factors such as the measurement environment.

[0077] In practice, the radar amplitude values ​​of multiple radar points can be normalized using methods such as linear normalization. Then, the average of all normalized radar point amplitude values ​​is taken to obtain the first parameter. The first parameter reflects the overall radar signal strength characteristics of the candidate target. For example, if the radar signal reflected by the candidate target is strong, the value of the first parameter will also be relatively large.

[0078] Step C2: Determine the ratio between the target velocity difference and the preset condition parameters to obtain the second parameter; Step C3: Determine the value of the inverse proportional exponential function of the second parameter, and multiply it with the first parameter to obtain the confidence level.

[0079] In this embodiment, the mean of the normalized amplitude values ​​of multiple radar points of the candidate target is determined as the first parameter. Normalization eliminates differences in amplitude values ​​caused by factors such as the measurement environment, allowing the first parameter to more objectively reflect the overall intensity of the candidate target's reflected signal. A stronger signal indicates a higher likelihood of the candidate target's existence. The target speed difference reflects the degree of speed difference between the candidate target and the vehicle. A second parameter is obtained by calculating the ratio of the target speed difference to a preset condition parameter, which measures the relative magnitude of the speed difference. The confidence level is obtained by multiplying the inverse proportional exponential function value of the second parameter by the first parameter. The inverse proportional exponential function highlights the positive impact of a small target speed difference on confidence. Combined with the first parameter, it comprehensively considers the reflected signal intensity and speed characteristics of the candidate target, thereby enabling a more accurate assessment of the candidate target's confidence level.

[0080] During implementation, the credibility of candidate targets can be determined using expression (10): (10) In expression (10), Indicates the credibility of the candidate target; This indicates the radar point amplitude corresponding to the candidate target detected by the vehicle-mounted radar; Indicates the maximum detectable amplitude of the radar; This represents the normalized reflection intensity factor, also known as the first parameter. The larger the radar point amplitude, the closer the first parameter is to 1, and the higher the reliability. This represents the natural exponential function, used to convert the target velocity difference into a decay factor between 0 and 1, ensuring that the impact of velocity consistency on credibility decreases smoothly. Indicates the initial speed measurement value; Indicates the vehicle's own speed; Indicates preset condition parameters; This indicates the second parameter.

[0081] S403: Filter candidate targets if the confidence level is less than the confidence level threshold.

[0082] In this embodiment of the disclosure, by determining the credibility of candidate targets based on radar point amplitude and initial speed measurement value, and filtering candidate targets based on the credibility, false targets caused by interference in rainy and foggy weather can be effectively eliminated, thereby improving the accuracy of vehicle-mounted radar in detecting real targets.

[0083] To further improve the accuracy of target selection, the credibility of the identified candidate targets can be further optimized based on the following steps: Step D1: Acquire visual images captured by the vehicle's vision sensors; Visual sensors typically refer to devices such as cameras mounted on vehicles. Their core function is to collect image information of the surrounding environment and convert this information into visual image data that can be used for subsequent target recognition.

[0084] Step D2: Extract multiple target objects from the visual image; In practice, object detection algorithms can be used to process visual images and identify various target objects present in the visual images, such as pedestrians, buildings, other vehicles or non-motorized vehicles.

[0085] Step D3: Project the static target in the data frame into the visual space of the visual sensor; In practice, the static target can be projected onto the visual space of the visual sensor using the calibrated radar extrinsic parameters of the vehicle-mounted radar, and then matched with multiple target objects extracted from the visual image.

[0086] Step D4: In the absence of a target object matching the static target in the visual space, the credibility is corrected based on a preset credibility optimization strategy, wherein the corrected credibility is lower than the original credibility.

[0087] That is, if no target object corresponding to the static target is matched in the visual space, the credibility of the static target is reduced according to the preset credibility optimization strategy.

[0088] It should be noted that visual information is used as an independent verification method. If the visual sensor does not detect a corresponding target object, the reliability of the target detected by the radar is low. Therefore, its reliability can be reduced to avoid making incorrect decisions based on the static target.

[0089] For example, false targets obtained by radar reflections can be verified in this way to determine whether they are real targets, thereby eliminating false targets in the data frames collected by vehicle-mounted radar.

[0090] In this embodiment, multiple target objects are extracted from visual images collected by a vision sensor. Simultaneously, static targets in the data frame are projected into the visual space of the vision sensor. If no matching target object is found in the visual space, it indicates that the static target detected by the radar may be a false target or that a detection error exists. The credibility of the static target is corrected based on a preset credibility optimization strategy, making the corrected credibility lower than the original. This avoids decision-making errors caused by erroneous target information, thereby improving the reliability and accuracy of vehicle driving decisions.

[0091] Based on the same technical concept, this disclosure also provides an online external parameter calibration device 500 for vehicle-mounted radar, such as... Figure 5 As shown, it includes: Extraction module 501 is used to extract static targets from data frames acquired by vehicle radar; The calibration module 502 is used to calibrate the radar extrinsic parameters of the vehicle-mounted radar in the world coordinate system by minimizing the velocity residual between the radar velocity measurement value and the reference velocity value of the static target; wherein, the reference velocity value is the velocity of the static target determined based on the pre-fusion data; the radar velocity measurement value is transformed into the world coordinate system through the radar extrinsic parameters.

[0092] In some embodiments, a conversion module is also included for converting radar velocity values ​​to the world coordinate system using radar extrinsic parameters based on the following method: Based on the radar extrinsic parameters, the radar velocity measurement values ​​are transformed into the world coordinate system to obtain intermediate values; Multiplying the intermediate value by the Jacobian matrix yields the radar velocity value in the world coordinate system.

[0093] In some embodiments, the extraction module includes: The acquisition unit is used to acquire the initial speed measurement values ​​of candidate targets in the data frame by the vehicle-mounted radar. The generation unit is used to generate multiple hypothetical speeds based on the initial speed measurement value and the speed ambiguity period of the vehicle radar. The first determining unit is used to determine the probability that multiple hypothetical velocities are the actual velocities based on the prior velocities of candidate targets; the prior velocities are the velocities predicted based on the historical trajectories of candidate targets. The selection unit is used to select the hypothetical velocity with the highest probability as the true velocity of the candidate target. The second determining unit determines the candidate target as a static target if the candidate target's actual velocity satisfies the static condition.

[0094] In some embodiments, the second determining unit is specifically used for: Obtain the vehicle's jerkiness; When the vehicle's jerkiness is less than the jerkiness threshold, the actual speed is compensated based on the vehicle's speed to obtain the calibrated speed. If the calibration speed is less than the speed threshold, the true speed of the candidate target is determined to meet the static condition, and the candidate target is determined to be a static target.

[0095] In some embodiments, the second determining unit is specifically used for: The calibrated speed is obtained by subtracting the vehicle's speed from the actual speed.

[0096] In some embodiments, the second determining unit is specifically used for: Mapping the urgency to a weight range yields the corresponding weight value for the urgency. The compensation value is obtained by multiplying the weight value by the vehicle's speed; The calibrated speed is obtained by subtracting the compensation value from the actual speed.

[0097] In some embodiments, a filtering module is also included for: In rainy or foggy weather, obtain the radar point amplitude and initial speed measurement value of the vehicle-mounted radar for the candidate target; The credibility of candidate targets is determined based on radar point amplitude and initial speed measurement value. Credibility is positively correlated with radar point amplitude and negatively correlated with target speed difference, which is the absolute difference between the initial speed measurement value and the vehicle speed. If the confidence level is less than the confidence threshold, filter out candidate targets.

[0098] In some embodiments, a correction module is further included, for: Acquire visual images captured by the vehicle's vision sensors; Extracting multiple target objects from a visual image; Project static targets in the data frame into the visual space of the vision sensor; In the absence of a target object matching the static target in the visual space, the credibility is corrected based on a preset credibility optimization strategy, wherein the corrected credibility is lower than the original credibility.

[0099] In some embodiments, the correction module is specifically used for: The first parameter is obtained by taking the mean of the normalized values ​​of the radar point amplitudes of multiple radar points of the candidate target. The ratio between the target velocity difference and the preset condition parameters is determined to obtain the second parameter; The confidence level is obtained by multiplying the inverse proportional exponential function value of the second parameter with the first parameter.

[0100] The specific functions and examples of each module and submodule of the apparatus in this disclosure can be found in the relevant descriptions of the corresponding steps in the above method embodiments, and will not be repeated here.

[0101] The acquisition, storage, and application of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0102] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0103] Figure 6 A schematic block diagram of an example electronic device 600 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0104] like Figure 6 As shown, device 600 includes a computing unit 601, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 602 or a computer program loaded into random access memory (RAM) 603 from storage unit 608. RAM 603 may also store various programs and data required for the operation of device 600. The computing unit 601, ROM 602, and RAM 603 are interconnected via bus 604. Input / output (I / O) interface 605 is also connected to bus 604.

[0105] Multiple components in device 600 are connected to I / O interface 605, including: input unit 606, such as keyboard, mouse, etc.; output unit 607, such as various types of monitors, speakers, etc.; storage unit 608, such as disk, optical disk, etc.; and communication unit 609, such as network card, modem, wireless transceiver, etc. Communication unit 609 allows device 600 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0106] The computing unit 601 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 601 performs the various methods and processes described above, such as the online extrinsic parameter calibration method for vehicle-mounted radar. For example, in some embodiments, the online extrinsic parameter calibration method for vehicle-mounted radar can be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as storage unit 608. In some embodiments, part or all of the computer program can be loaded and / or installed on device 600 via ROM 602 and / or communication unit 609. When the computer program is loaded into RAM 603 and executed by the computing unit 601, one or more steps of the online extrinsic parameter calibration method for vehicle-mounted radar described above can be performed. Alternatively, in other embodiments, the computing unit 601 may be configured by any other suitable means (e.g., by means of firmware) to perform an online extrinsic calibration method for vehicle radar.

[0107] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0108] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0109] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0110] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0111] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0112] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0113] Based on the aforementioned electronic devices, this disclosure also provides a vehicle that may include electronic devices, and may also include communication components, a display screen for realizing a human-machine interface, and an information collection device for collecting information about the surrounding environment, etc., wherein the communication components, the display screen, the information collection device and the electronic devices are communicatively connected.

[0114] According to embodiments of this disclosure, the electronic device can be integrated with the communication component, display screen, and information acquisition device, or it can be separately configured with the communication component, display screen, and information acquisition device.

[0115] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0116] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for online calibration of extrinsic parameters of vehicle-mounted radar, comprising: Static targets are extracted from data frames acquired by vehicle-mounted radar; In the world coordinate system, the radar extrinsic parameters of the vehicle-mounted radar are calibrated by minimizing the velocity residual between the radar velocity measurement value and the reference velocity value of the static target; wherein, the reference velocity value is the velocity of the static target determined based on the pre-fusion data; the radar velocity measurement value is transformed into the world coordinate system through the radar extrinsic parameters.

2. The method according to claim 2, further comprising converting the radar velocity value to the world coordinate system using the radar extrinsic parameters based on the following method: Based on the radar extrinsic parameters, the radar velocity measurement value is converted to the world coordinate system to obtain an intermediate value; Multiplying the intermediate value by the Jacobian matrix yields the speed of the radar velocity measurement in the world coordinate system.

3. The method according to claim 1, wherein, The static target extracted from the data frame acquired from the vehicle-mounted radar includes: Obtain the initial speed measurement value of the candidate target in the data frame by the vehicle-mounted radar; Based on the initial speed measurement value and the speed ambiguity period of the vehicle-mounted radar, multiple hypothetical speeds are generated; Based on the prior velocity of the candidate target, the probability that the plurality of hypothetical velocities are the actual velocities is determined; the prior velocity is the velocity predicted based on the historical trajectory of the candidate target. The assumed speed with the highest probability is selected as the true speed of the candidate target; If the actual velocity of the candidate target satisfies the static condition, the candidate target is determined to be the static target.

4. The method according to claim 3, wherein, Determining a candidate target as a static target when its actual velocity satisfies the stationary condition includes: Obtain the vehicle's jerkiness; If the jerkiness of the vehicle is less than the jerkiness threshold, the actual speed is compensated based on the vehicle's speed to obtain a calibrated speed. If the calibration speed is less than the speed threshold, the true speed of the candidate target is determined to satisfy the stationary condition, and the candidate target is determined to be the static target.

5. The method according to claim 4, wherein, The process of compensating the real speed based on the vehicle's speed to obtain a calibrated speed includes: The calibrated speed is obtained by subtracting the vehicle's speed from the actual speed.

6. The method according to claim 4, wherein, The process of compensating the real speed based on the vehicle's speed to obtain a calibrated speed includes: The jerkiness is mapped to a weight value range to obtain the weight value corresponding to the jerkiness; The compensation value is obtained by multiplying the weight value by the vehicle's speed; The calibrated speed is obtained by subtracting the compensation value from the actual speed.

7. The method according to any one of claims 3-6, further comprising: In rainy or foggy weather, the radar point amplitude and initial speed measurement value of the vehicle-mounted radar for the candidate target are obtained; Based on the radar point amplitude and the initial speed measurement value, the credibility of the candidate target is determined; the credibility is positively correlated with the radar point amplitude and negatively correlated with the target speed difference, which is the absolute difference between the initial speed measurement value and the vehicle speed. If the confidence level is less than the confidence threshold, the candidate targets are filtered.

8. The method according to claim 7, further comprising: Acquire visual images captured by the vehicle's vision sensors; Extract multiple target objects from the visual image; The static target in the data frame is projected into the visual space of the visual sensor; If there is no target object matching the static target in the visual space, the credibility is corrected based on a preset credibility optimization strategy, wherein the corrected credibility is lower than the original credibility.

9. The method according to claim 7, wherein, Determining the credibility of the candidate target based on the radar point amplitude and the initial velocity value includes: The mean of the normalized values ​​of the radar point amplitudes of multiple radar points of the candidate target is determined to obtain the first parameter; The ratio between the target speed difference and the preset condition parameter is determined to obtain the second parameter; The confidence level is obtained by multiplying the inverse proportional exponential function value of the second parameter with the first parameter.

10. An online extrinsic parameter calibration device for vehicle-mounted radar, comprising: The extraction module is used to extract static targets from the data frames acquired by the vehicle-mounted radar; The calibration module is used to calibrate the radar extrinsic parameters of the vehicle-mounted radar in the world coordinate system by minimizing the velocity residual between the radar velocity measurement value and the reference velocity value of the static target; wherein the reference velocity value is the velocity of the static target determined based on the pre-fusion data; the radar velocity measurement value is transformed into the world coordinate system through the radar extrinsic parameters.

11. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-9.

12. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-9.

13. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-9.

14. An autonomous vehicle, including the electronic equipment as claimed in claim 11.

Images