A millimeter wave radar installation angle deviation online calibration method and device

Abstract

The present invention discloses a method and device for online calibration of installation angle deviation of millimeter wave radar. When the installation angle deviation is online calibrated, the method obtains the vehicle speed and steering angular velocity information, calculates the radar transverse and longitudinal speeds at the current moment according to the radar sensor installation position parameters; obtains a set of radar measurement data at the current moment; uses each measurement data in the radar measurement data set to calculate the radar installation azimuth deviation, and obtains a first radar installation azimuth deviation set; excludes points from the first radar installation azimuth deviation set to obtain a second radar installation azimuth deviation set; calculates the arithmetic mean of the deviations in the second radar installation azimuth deviation set to obtain a first radar installation azimuth deviation calibration value; repeats three times, performs validity judgment on the obtained first radar installation azimuth deviation calibration value, and takes the median as the second radar installation azimuth deviation calibration value. The present invention can improve the accuracy of installation angle deviation calibration.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving sensor calibration, and in particular to an online calibration method and device for millimeter wave radar installation angle deviation. Background Art

[0002] As a key sensor in FCW, AEB, and ACC systems, millimeter-wave forward radar can effectively reduce the probability of accidents during car driving. Forward millimeter-wave radar is generally installed in the center of the front of the car, requiring the radar normal to coincide with the center line of the vehicle. However, during the installation and subsequent use of the radar, a certain azimuth deviation may occur, resulting in a deviation in the target measurement angle, which may seriously lead to missed reports and false reports of the target. In order not to affect the accuracy of target detection, automatic online calibration of the radar installation azimuth deviation is required during use.

[0003] The existing online azimuth deviation calibration method mainly uses the radar to detect the long straight guardrail to fit the straight line equation, or manually sets static markers such as angular inversion to use the measured radial velocity and azimuth relationship to complete the calibration of the radar installation azimuth deviation.

[0004] The method of fitting a straight line equation for a long straight guardrail requires the assumption that the direction of travel of the ego vehicle is strictly parallel to the guardrail;

[0005] The method of the relationship between radial velocity and azimuth angle requires that the vehicle travels in a straight line, and both have high requirements for the calibration scene. It is very easy to introduce errors because the conditions cannot be strictly met. Summary of the invention

[0006] The embodiment of the present invention provides a method and device for online calibration of installation angle deviation of a millimeter wave radar, which can improve the accuracy of the online calibration result of the installation angle deviation.

[0007] A millimeter wave radar installation angle deviation online calibration method comprises the following steps:

[0008] S1, obtaining the vehicle's motion information from the chassis, and calculating the current lateral and longitudinal speeds of the radar sensor based on the radar sensor's installation position parameters;

[0009] S2, obtaining the radar measurement data set at the current moment;

[0010] S3, using each measurement data in the radar measurement data set obtained in step S2 to calculate the radar installation azimuth deviation, to obtain a first radar installation azimuth deviation set;

[0011] S4, performing a random sampling consensus algorithm on the first radar installation azimuth deviation set to remove external points, to obtain a second radar installation azimuth deviation set obtained by stationary target measurement and calculation;

[0012] S5, calculating the arithmetic mean of the deviations in the second radar installation azimuth deviation set to obtain a calibration value of the first radar installation azimuth deviation;

[0013] S6. Repeat steps S1 to S5 three times, judge the validity of the first radar installation azimuth deviation calibration values ​​obtained three times, take the median as the second radar installation azimuth deviation calibration value, and complete the radar installation azimuth deviation calibration at the current moment.

[0014] Furthermore, the vehicle motion information includes the vehicle speed and steering angular velocity; the sensor installation position parameters include the horizontal and vertical coordinates of the radar installation position in the vehicle coordinate system; and the calculation of the horizontal and vertical speeds of the radar sensor itself at the current moment specifically includes:

[0015] The radar sensor's own longitudinal velocity is calculated using the following formula

[0016] The radar sensor's own lateral velocity is calculated using the following formula

[0017] Among them, v ego is the vehicle speed;

[0018] ω is the steering angular velocity of the vehicle;

[0019] a is the ordinate of the radar sensor installation position;

[0020] b is the horizontal coordinate of the radar sensor installation position;

[0021] is the longitudinal velocity of the radar sensor;

[0022] is the lateral velocity of the radar sensor.

[0023] Furthermore, the radar measurement data includes the distance, azimuth and relative radial velocity of the target detection point from the radar; and the radar measurement data set includes a plurality of radar measurement data detected at the current moment.

[0024] Further, in step S3, the radar installation azimuth deviation is calculated by using each measurement data in the radar measurement data set, specifically including:

[0025] The installation azimuth deviation is calculated using the following formula:

[0026] The first radar installation azimuth deviation set is expressed as

[0027] Among them, Δθ i is the azimuth deviation calculated using measurement i;

[0028] is the longitudinal velocity of the radar sensor;

[0029] is the lateral velocity of the radar sensor;

[0030] θ i Measure the azimuth of i;

[0031] is the radial relative velocity of measurement i;

[0032] N t To measure the number.

[0033] Further, in step S4, performing a random sampling consistent operation on the first radar installation azimuth deviation set to remove external points to obtain a second radar installation azimuth deviation set obtained by stationary target measurement and calculation specifically includes:

[0034] The random sampling consensus algorithm is a general outlier removal algorithm that can estimate the parameters of a mathematical model from a set of observation data containing outliers in an iterative manner. It is an uncertain algorithm with a certain probability of obtaining a reasonable result. In order to increase the probability, the number of iterations must be increased;

[0035] The second radar installation azimuth deviation set is expressed as

[0036] Among them, N ts It is the number of radar installation deviation angles obtained by measuring and calculating stationary targets and selected by a random sampling consensus algorithm.

[0037] Furthermore, step S6 specifically includes:

[0038] The set of calibration values ​​of the first radar installation azimuth deviation obtained by three repetitions is expressed as

[0039] The validity is judged as follows: If the condition is met, the calibration is judged to be successful, and the output median value is Install azimuth calibration values ​​for the second radar;

[0040] TH is the judgment threshold, which is selected as 0.02rad based on experience.

[0041] The present invention further discloses an online calibration device for radar installation azimuth deviation, comprising:

[0042] A first calculation module, a radar data acquisition module, a second calculation module, a third calculation module, a fourth calculation module, and a determination output module, wherein:

[0043] The first calculation module is used to calculate the current moment's lateral and longitudinal speeds of the radar sensor itself based on the vehicle motion information obtained from the vehicle chassis and the radar sensor installation position parameters;

[0044] The radar data acquisition module is used to acquire a set of radar measurement data at the current moment;

[0045] The second calculation module is used to calculate the radar installation azimuth deviation according to each measurement data in the radar measurement data set to obtain a first radar installation azimuth deviation set;

[0046] The third calculation module is used to perform a random sampling consistent algorithm on the first radar installation azimuth deviation set to remove external points, so as to obtain a second radar installation azimuth deviation set obtained by stationary target measurement and calculation;

[0047] The fourth calculation module calculates the arithmetic mean of the deviations in the second radar installation azimuth deviation set to obtain a calibration value of the first radar installation azimuth deviation;

[0048] The determination output module is used to determine the validity of the first radar installation azimuth deviation calibration values ​​obtained three times, take the median as the second radar installation azimuth deviation calibration value, and complete the radar installation azimuth deviation calibration at the current moment.

[0049] The following beneficial effects are achieved by implementing the embodiments of the present invention:

[0050] The embodiment of the present invention provides a method and device for online calibration of millimeter wave radar installation angle deviation. When updating a road model, the method first obtains vehicle motion information from a vehicle chassis, calculates the lateral and longitudinal speeds of the radar sensor itself at the current moment according to the radar sensor installation position parameters, then obtains a set of radar measurement data at the current moment, uses each measurement data in the radar measurement data set to calculate the radar installation azimuth deviation, and obtains a first radar installation azimuth deviation set. Then, the first radar installation azimuth deviation set is randomly extracted by a consistent algorithm to remove external points, and a second radar installation azimuth deviation set obtained by stationary target measurement is obtained. The arithmetic average is calculated to obtain a calibration value of the first radar installation azimuth deviation, and then the above operation is repeated three times. The validity of the calibration values ​​of the first radar installation azimuth deviation obtained three times is judged, and finally the median is taken as the calibration value of the second radar installation azimuth deviation, and the calibration of the radar installation azimuth deviation at the current moment is completed. Compared with the prior art, the present invention uses the vehicle's speed and steering angular velocity information to obtain the lateral and longitudinal velocities of the radar sensor during the online calibration of the installation angle deviation, and considers the sensor's lateral velocity when calculating the relative radial velocity, thereby solving the technical problem that the current online calibration method based on radial relative velocity and azimuth requires the vehicle to travel in a straight line, has high requirements for the calibration scene, and is prone to introducing large errors due to the failure to meet the assumption during calibration implementation; or the online calibration time is long due to the failure to meet the calibration scene requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 The present invention is a flowchart of an online calibration method for a millimeter wave radar installation angle deviation provided in an embodiment of the present invention.

[0052] Figure 2 It is a structural schematic diagram of an online calibration device for millimeter wave radar installation angle deviation provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] like Figure 1 As shown, an embodiment of the present invention provides a method for online calibration of millimeter wave radar installation angle deviation, comprising the following steps:

[0055] Step 101: Obtain vehicle speed and steering angular velocity information from the vehicle chassis, and calculate the current radar lateral and longitudinal velocities based on the radar sensor installation position parameters.

[0056] Step 102: Acquire a set of radar measurement data at the current moment, wherein the radar measurement data includes distance, azimuth, and radial relative velocity information measured by the radar.

[0057] Step 103: Calculate the radar installation azimuth deviation using each measurement data in the radar measurement data set to obtain a first radar installation azimuth deviation set.

[0058] Step 104: Perform a RANSAC operation on the first radar installation azimuth deviation set to remove outliers, and obtain a second radar installation azimuth deviation set calculated by stationary target detection.

[0059] Step 105: Calculate the arithmetic mean of the deviations in the second radar installation azimuth deviation set to obtain the first radar installation azimuth deviation calibration value.

[0060] Step 106: Repeat steps 101 to 105 three times, judge the validity of the obtained first radar installation azimuth deviation calibration value, take the median as the second radar installation azimuth deviation calibration value, and complete the online calibration of the installation azimuth error.

[0061] For step 101, specifically, the vehicle motion information obtained from the chassis includes the vehicle velocity v at the origin of the vehicle coordinate system (the center of the rear axle of the vehicle) ego and steering angular velocity ω; the installation position of the radar sensor in the vehicle coordinate system includes the lateral distance a and longitudinal distance b of the radar sensor from the vehicle coordinate system; the longitudinal velocity of the radar sensor is given by The lateral velocity is calculated from Calculated;

[0062] For step 102, specifically, obtain the radar measurement set at the current moment i represents the i-th radar measurement, N t is the number of radar measurements, z=[r,θ,v r ] represents a single radar measurement data, including the measured distance r, azimuth θ, radial velocity v r It should be noted that the above radar measurement set is all detections reported by the radar at the current moment, including the detection results of all moving or stationary targets within the radar coverage area and a small number of false detections.

[0063] For step 103, in a preferred embodiment, the step of calculating the radar installation azimuth deviation using each measurement data in the radar measurement data set specifically includes:

[0064] The relative radial velocity of stationary target detection can be expressed as:

[0065]

[0066] in, is the longitudinal velocity of the radar sensor; is the lateral velocity of the radar sensor; θ i Measure the azimuth angle of i; Δθ is the true value of the radar sensor azimuth deviation, which is a smaller angle value; is the radial relative velocity of measurement i.

[0067] The above formula can be simplified by using cos(Δθ)≈1,sin(Δθ)≈Δθ

[0068]

[0069] Therefore, the estimated value of the installation azimuth deviation is calculated by the following formula:

[0070] For all N t The first radar installation azimuth deviation set can be expressed as Among them, Δθ i is the azimuth deviation calculated using measurement i.

[0071] For step S104, in a preferred embodiment, performing a RANSAC operation on the first radar installation azimuth deviation set to remove outliers specifically includes:

[0072] The first radar installation azimuth deviation set is a set of radar installation azimuth deviations calculated using all radar measurement speeds. Generally speaking, stationary targets account for the majority of them. For stationary target detection data, the installation azimuth deviation estimate Δθ is obtained. i The deviation estimate can be expressed as Δθ is the true value of the deviation, The estimated error is 0 mean; for moving target detection or false detection data, the estimated value of the installation azimuth deviation is irregular and chaotically distributed, so the random extraction consensus algorithm can be used to filter out such outliers to obtain the second radar installation azimuth deviation set calculated by stationary target detection. Among them, N ts The number of stationary target detections.

[0073] For step S105, in a preferred embodiment, the deviations in the second radar installation azimuth deviation set are calculated to obtain the first radar installation azimuth deviation calibration value.

[0074] For step S106, in a preferred embodiment, the steps 101 to 105 are repeated three times, the validity of the obtained first radar installation azimuth deviation calibration value is judged, and the median is taken as the second radar installation azimuth deviation calibration value, which specifically includes:

[0075] Repeat steps 101 to 105 three times, and express the set of the first radar installation azimuth deviation calibration values ​​obtained by the three repetitions as

[0076] Determine the absolute value of the difference between the maximum and minimum values ​​of the three results Is it less than the fixed threshold of 0.02 radians? If it is not less than the threshold, the three calibration results are considered inconsistent and need to be recalibrated; if it is less than the threshold, the calibration is successful and the middle value of the three results is selected as the calibration value of the second radar installation azimuth deviation.

[0077] Through the online calibration method for radar installation azimuth angle deviation disclosed in the above embodiment of the present invention, the radar installation angular deviation at the current moment is calibrated, and the radar sensor lateral velocity information obtained by calculating the vehicle's offset angular velocity and the radar installation coordinates is introduced. Even when the vehicle is not traveling in a straight line, the relative radial velocity of stationary target detection can be accurately modeled, thereby using the radar detection information to perform online calibration of the radar installation azimuth angle deviation.

[0078] Based on the above method embodiment, the present invention provides a corresponding device embodiment;

[0079] like Figure 2 As shown, an embodiment of the present invention provides an online calibration device for radar installation azimuth deviation, including: a first calculation module, a radar data acquisition module, a second calculation module, a third calculation module, a fourth calculation module, and a determination output module;

[0080] The first calculation module is used to calculate the current moment's lateral and longitudinal speeds of the radar sensor itself based on the vehicle motion information obtained from the vehicle chassis and the radar sensor installation position parameters;

[0081] The radar data acquisition module is used to acquire a set of radar measurement data at the current moment;

[0082] The second calculation module is used to calculate the radar installation azimuth deviation according to each measurement data in the radar measurement data set to obtain a first radar installation azimuth deviation set;

[0083] The third calculation module is used to perform a random sampling consistent algorithm on the first radar installation azimuth deviation set to remove external points, so as to obtain a second radar installation azimuth deviation set obtained by stationary target measurement and calculation;

[0084] The fourth calculation module calculates the arithmetic mean of the deviations in the second radar installation azimuth deviation set to obtain a calibration value of the first radar installation azimuth deviation;

[0085] The determination output module is used to determine the validity of the first radar installation azimuth deviation calibration values ​​obtained three times, take the median as the second radar installation azimuth deviation calibration value, and complete the radar installation azimuth deviation calibration at the current moment.

[0086] It should be noted that the above-mentioned device item embodiment corresponds to the method item embodiment of the present invention, which can implement the road model updating method described in any embodiment of the present invention. In addition, the device embodiment described above is only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the drawings of the device embodiment provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which can be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art can understand and implement it without creative work.

[0087] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A millimeter wave radar installation angle deviation online calibration method, characterized in that: The following steps are involved: S1, obtaining the vehicle's motion information from the chassis, and calculating the current lateral and longitudinal speeds of the radar sensor based on the radar sensor's installation position parameters; S2, obtaining the radar measurement data set at the current moment; S3, using each measurement data in the radar measurement data set obtained in step S2 to calculate the radar installation azimuth deviation, to obtain a first radar installation azimuth deviation set; S4, performing a random sampling consensus algorithm on the first radar installation azimuth deviation set to remove external points, to obtain a second radar installation azimuth deviation set obtained by stationary target measurement and calculation; S5, calculating the arithmetic mean of the deviations in the second radar installation azimuth deviation set to obtain a calibration value of the first radar installation azimuth deviation; S6, repeating steps S1 to S5 three times, judging the validity of the three obtained first radar installation azimuth deviation calibration values, taking the median as the second radar installation azimuth deviation calibration value, and completing the radar installation azimuth deviation calibration at the current moment; In step S3, the radar installation azimuth deviation is calculated using each measurement data in the radar measurement data set, specifically including: The installation azimuth deviation is calculated using the following formula: ; The first radar installation azimuth deviation set is expressed as ; in, is the azimuth deviation calculated using measurement i; is the longitudinal velocity of the radar sensor; is the lateral velocity of the radar sensor; Measure the azimuth of i; is the radial relative velocity of measurement i; To measure the number.

2. The method for online calibration of the installation angle deviation of the millimeter wave radar according to claim 1, characterized in that: The vehicle motion information includes the vehicle speed and steering angular velocity; the sensor installation position parameters include the horizontal and vertical coordinates of the radar installation position in the vehicle coordinate system; the calculation of the current moment's radar sensor's own horizontal and vertical speeds specifically includes: The radar sensor's own longitudinal velocity is calculated using the following formula ; The radar sensor's own lateral velocity is calculated using the following formula ; in, is the vehicle speed; is the steering angular velocity of the vehicle; The vertical coordinate of the installation position of the radar sensor; The horizontal coordinate of the radar sensor installation position; is the longitudinal velocity of the radar sensor; is the lateral velocity of the radar sensor.

3. The method for online calibration of the installation angle deviation of the millimeter wave radar according to claim 1, characterized in that: The radar measurement data includes the distance, azimuth and relative radial velocity of the target detection point from the radar; and the radar measurement data set includes a plurality of radar measurement data detected at the current moment.

4. The method for online calibration of the installation angle deviation of the millimeter wave radar according to claim 1, characterized in that: In step S4, performing a random sampling consistent operation on the first radar installation azimuth deviation set to remove external points to obtain a second radar installation azimuth deviation set obtained by stationary target measurement and calculation specifically includes: The random sampling consensus algorithm is a general outlier removal algorithm that can estimate the parameters of a mathematical model from a set of observation data containing outliers in an iterative manner. It is an uncertain algorithm with a certain probability of obtaining a reasonable result. In order to increase the probability, the number of iterations must be increased; The second radar installation azimuth deviation set is expressed as ; in, It is the number of radar installation deviation angles obtained by measuring and calculating stationary targets and selected by a random sampling consensus algorithm.

5. The method for online calibration of the installation angle deviation of the millimeter wave radar according to claim 1, characterized in that: Step S6 specifically includes: The set of calibration values ​​of the first radar installation azimuth deviation obtained by three repetitions is expressed as ; The validity is judged as follows: If the condition is met, the calibration is judged to be successful, and the output median value is Install azimuth calibration values ​​for the second radar; in For the determination threshold, 0.02rad is selected based on experience.

6. A radar installation azimuth deviation online calibration device, used in the millimeter wave radar installation angle deviation online calibration method according to any one of claims 1 to 5, characterized in that: include: A first calculation module, a radar data acquisition module, a second calculation module, a third calculation module, a fourth calculation module, and a determination output module, wherein: The first calculation module is used to calculate the current moment's lateral and longitudinal speeds of the radar sensor itself based on the vehicle motion information obtained from the vehicle chassis and the radar sensor installation position parameters; The radar data acquisition module is used to acquire a set of radar measurement data at the current moment; The second calculation module is used to calculate the radar installation azimuth deviation according to each measurement data in the radar measurement data set to obtain a first radar installation azimuth deviation set; The third calculation module is used to perform a random sampling consistent algorithm on the first radar installation azimuth deviation set to remove external points, so as to obtain a second radar installation azimuth deviation set obtained by stationary target measurement and calculation; The fourth calculation module calculates the arithmetic mean of the deviations in the second radar installation azimuth deviation set to obtain a calibration value of the first radar installation azimuth deviation; The determination output module is used to determine the validity of the first radar installation azimuth deviation calibration values ​​obtained three times, take the median as the second radar installation azimuth deviation calibration value, and complete the radar installation azimuth deviation calibration at the current moment.

Images