Vehicle following method at intersection based on multi-sensor fusion

Abstract

The present invention provides a method for following a vehicle at an intersection based on multi-sensor fusion, comprising steps S1 to S3. In step S1, information such as the distance between the vehicle and the target vehicle in front, the distance between the vehicle and the traffic light, and the turn signal status of the target vehicle are obtained through radar and camera, and whether to enter the intersection following mode is determined according to the distance and the turn signal status. In step S2, different following modes are selected according to the turn signal signals of the vehicle and the target vehicle and the color of the traffic light, and the relative position and speed of the target vehicle are obtained by radar and camera to calculate the real-time collision time, and the speed of the vehicle and the steering wheel angle are controlled to automatically follow the vehicle under the premise of avoiding collision at the intersection. In step S3, the exit from the intersection following mode is determined according to the road surface markings recognized by the camera. The invention can improve the safety and stability of vehicle driving and is suitable for automatic driving systems.

Description

Technical Field

[0001] The present invention relates to the technical field of autonomous driving, and in particular to a vehicle following method at an intersection based on multi-sensor fusion. Background Art

[0002] With the continuous development of automobile technology, adaptive cruise control has become a standard feature of many high-end cars. This intelligent driving system can obtain real-time information about the vehicle's surrounding environment through sensors such as radar and cameras, and automatically adjust the driving speed and distance according to factors such as vehicle speed and distance, thereby achieving autonomous cruising of the vehicle. The adaptive cruise control system can not only improve driving comfort and safety, but also reduce the risks caused by fatigue driving, giving drivers a better driving experience.

[0003] The current adaptive cruise control system has evolved from the initial high-speed working conditions and road conditions to full-speed working conditions and urban road conditions, but the existing technology is mainly focused on vehicle following and cruising on straight roads. In complex urban road environments, the adaptive cruise control system faces many technical challenges. For example, on urban roads, vehicles need to frequently change lanes, avoid obstacles, and alternate with other vehicles. These operations require accurate perception and analysis of the vehicle's surrounding environment in order to make correct decisions and control. Summary of the invention

[0004] The object of the present invention is to provide a vehicle following method at an intersection based on multi-sensor fusion, so as to realize automatic following of a target vehicle in front under different traffic scenarios at the intersection.

[0005] To solve the above technical problems, the present invention provides a technical solution: a vehicle following method at an intersection based on multi-sensor fusion, comprising the following steps:

[0006] S1. In the straight-ahead phase before turning at an intersection, the radar and camera are used to obtain the distance between the vehicle and the target vehicle in front of the vehicle, the distance between the vehicle and the traffic light at the intersection, and the state of the target vehicle's turn signal. Based on the distance between the vehicle and the target vehicle in front of the vehicle, the distance between the vehicle and the traffic light at the intersection, and the state of the target vehicle's turn signal, it is determined whether to enter the intersection following mode;

[0007] S2. After entering the intersection following mode, different following modes are selected according to whether the turn signal of the own vehicle and the target vehicle is turned on and the color of the forward or turn traffic light; in different following modes, the ego vehicle uses radar and camera to obtain the position and speed of the target vehicle relative to the ego vehicle, and combines the distance between the ego vehicle and the target vehicle in front of the ego vehicle, the distance between the ego vehicle and the traffic light at the intersection, the state of the target vehicle's turn signal, and the driving parameters of the ego vehicle obtained by using radar and camera in S1 to obtain the real-time collision time between the ego vehicle and the target vehicle in different following modes, and then control the speed of the ego vehicle according to the real-time collision time and the speed limit of the ego vehicle's driving section, and adjust the steering wheel angle of the ego vehicle according to the angle between the target vehicle and the driving angle of the ego vehicle;

[0008] S3. Determine whether to exit the intersection following mode according to the road signs recognized by the camera.

[0009] According to the above scheme, the specific judgment process conditions of whether to enter the intersection following mode in S1 are:

[0010] Take the rear bumper center of the target vehicle as the reference point of the target vehicle. In the straight-ahead phase before turning at the intersection, obtain the distance D1 between the vehicle and the traffic light at the intersection through the camera, and obtain the distance D2 between the vehicle and the target vehicle in front of the vehicle through the camera and radar. When the following conditions are met at the same time, enter the intersection following mode:

[0011] 1)D1 <D 1max ;

[0012] 2)D2 <D 2max ;

[0013] 3) The turn signals of the ego vehicle and the target vehicle are consistent;

[0014] Among them, D 1max , D 2max These are the preset distance values.

[0015] According to the above scheme, the following modes described in S2 include the first following mode, the second following mode, the third following mode, and the fourth following mode; in all following modes, the vehicle coordinate system is established, with the direction of the vehicle traveling as the positive x direction and the direction perpendicular to the x direction to the left as the positive y direction; the activation adjustment of each following mode is respectively,

[0016] When both the own vehicle and the target vehicle have no turn signal and the forward traffic light is green, the first following mode is selected; when both the own vehicle and the target vehicle have no turn signal and the forward traffic light is red, the second following mode is selected; when both the own vehicle and the target vehicle have turn signal and the left or right turn signal light is red, the third following mode is selected; when both the own vehicle and the target vehicle have turn signal and the left or right turn signal light is green, the fourth following mode is selected, and the fourth following mode includes a straight-ahead phase and a turning phase.

[0017] According to the above scheme, in the first following mode, the following operation performed by the vehicle is specifically as follows:

[0018] The component dx of the distance between the self-vehicle and the target vehicle in the x-direction, as well as the component vx of the speed of the target vehicle relative to the self-vehicle in the x-direction are obtained through the camera and radar, and TTC1=dx / vx is calculated; the speed of the self-vehicle is controlled so that the speed of the self-vehicle does not exceed the speed limit of the road section, and TTC1 is not lower than a preset time value.

[0019] According to the above scheme, in the second following mode or the third following mode, the following operation performed by the vehicle is specifically:

[0020] The x-component dx of the distance between the ego vehicle and the target vehicle and the x-component vx of the speed of the target vehicle relative to the ego vehicle are obtained through the camera and radar, and TTC1 = dx / vx is calculated; the distance dx between the ego vehicle and the stop line is obtained through the camera t , calculate TTC2 = dx t / egov, where egov is the current vehicle speed; the vehicle speed does not exceed the speed limit of the road section, and both TTC1 and TTC2 are greater than a preset time value.

[0021] According to the above scheme, in the straight-ahead phase of the fourth following mode, the following operation performed by the vehicle is the same as that of the first following mode. In the turning phase of the fourth following mode, the following operation performed by the vehicle is specifically:

[0022] First, the driving parameters vx, vy, dx, and dy of the current target vehicle relative to the ego vehicle are obtained through the camera and radar, and the absolute speed v of the target vehicle relative to the ground is obtained according to the ego vehicle speed egov and the ego vehicle rotation angular velocity w. abs ;Specifically,

[0023]

[0024] Among them, vy is the component of the target vehicle's velocity relative to the ego vehicle in the y direction, and dy is the component of the distance between the ego vehicle and the target vehicle in the y direction;

[0025] The target vehicle’s turning action at the intersection is regarded as v abs The target car moves in a circle at a speed of, and the turning radius L is,

[0026]

[0027] The absolute angular velocity of the target vehicle relative to the ground is w abs =v abs / L;

[0028] T is the duration of a sensor detection cycle. Assuming that the vehicle does not move from the current moment to the next sensor detection cycle after T time, the driving parameters of the target vehicle relative to the vehicle after T time are predicted to be dx′, dy′, vx′, and vy′. Specifically,

[0029]

[0030] Since the vehicle will follow the target vehicle in the next sensor detection cycle from the current time to T time later, the target vehicle’s driving parameter relative to the vehicle after T time is predicted to be dx T ,dy T 、vx T ,vy T , specifically,

[0031] dx T =dx′-L*sin(w*T)

[0032] dy T =dy′-L+L*cos(w*T)

[0033]

[0034] The radar and camera measurements are then matched to the turning target vehicle. Specifically,

[0035] If the distance matching threshold in the straight-ahead phase of the fourth following mode is GATE, and half of the vehicle width length of the target vehicle is obtained as width, then the distance matching threshold in the turning phase of the fourth following mode is GATE+width;

[0036] Among all objects detected by the camera, objects whose distance to the vehicle is less than the distance matching threshold are regarded as candidate targets, and among all candidate targets, the candidate targets with the same target category as the target vehicle in the previous sensor detection cycle and the closest distance to the vehicle are regarded as target vehicles;

[0037] During the radar matching process, all objects detected by the radar whose distance to the vehicle is less than the distance matching threshold are selected as candidate targets. The radar measurement of the candidate targets is translated. If the target vehicle's turning direction predicted by the previous sensor detection cycle is a left turn, it is translated to the lower left, and if it is a right turn, it is translated to the lower right.

[0038] The horizontal translation distance is

[0039] The vertical translation distance is

[0040] The radar measurements before and after the translation that are closest to the target vehicle's driving parameters relative to the self-vehicle predicted in the previous sensor detection cycle are taken as the matching result;

[0041] Finally, the target vehicle's driving parameters relative to the vehicle after matching are combined with the target vehicle's driving parameters relative to the vehicle predicted in the previous sensor detection cycle, and the driving parameters dx1, dy1, vx1, and vy1 of the target vehicle relative to the vehicle in this sensor detection cycle are obtained. The absolute speed v of the target vehicle relative to the ground in this sensor detection cycle is obtained. abs1 , the turning radius L1 of the target vehicle, the absolute rotational angular velocity w of the target vehicle relative to the ground abs1 , combined with the vehicle speed egov1 and the angular velocity w1 of the vehicle during this sensor detection cycle; assuming that the target vehicle is on the driving trajectory of the vehicle, then

[0042]

[0043] Make the vehicle speed not exceed the maximum speed required for turning, and make TTC3 greater than a preset time value. When following, adjust the steering wheel angle of the vehicle according to the angle between the vehicle and the target vehicle.

[0044] According to the above scheme, S3 specifically includes exiting the intersection following mode when the following conditions are met at the same time:

[0045] 1) The camera recognizes the lane line;

[0046] 2) The lane length is greater than D 2max ;

[0047] 3) There is no stop line within the camera's recognition range.

[0048] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned intersection vehicle following method based on multi-sensor fusion are implemented.

[0049] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned intersection vehicle following method based on multi-sensor fusion.

[0050] A car comprises the computer device described above.

[0051] The beneficial effects of the present invention are:

[0052] 1. It realizes the automation of vehicle following at intersections, improving road traffic efficiency and traffic safety.

[0053] 2. By using radar and camera fusion data, the location, speed and other information of the target vehicle can be obtained more accurately, and the turn signal status and traffic light color can be combined to determine whether to enter the intersection following mode, avoiding the need for manual intervention.

[0054] 3. In different following modes, the speed of the vehicle is controlled according to the real-time collision time and the speed limit of the road section where the vehicle is traveling, and the steering wheel angle of the vehicle is adjusted according to the angle between the target vehicle and the driving angle of the vehicle, so as to achieve more accurate vehicle following and improve the following effect.

[0055] 4. By identifying road signs, it is determined whether to exit the intersection following mode, which avoids unnecessary following delays and improves the operability of intersection following.

[0056] 5. The driving parameters of the target vehicle relative to the own vehicle obtained in each sensor detection cycle are obtained by fusing the measurement data of the camera and radar in the current sensor detection cycle with the predicted driving parameters of the target vehicle relative to the own vehicle output in the previous sensor detection cycle, thereby improving the accuracy of the detection of the target vehicle's motion. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 is a schematic diagram of parameters required for enabling conditions of the intersection following mode according to an embodiment of the present invention;

[0058] Figure 2 is a schematic diagram of a vehicle coordinate system according to an embodiment of the present invention;

[0059] Figure 3 is a schematic diagram of the predicted driving parameters of a target vehicle relative to the own vehicle according to an embodiment of the present invention;

[0060] Figure 4 It is a flow chart of a vehicle following method at an intersection based on multi-sensor fusion according to an embodiment of the present invention. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution of the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings of the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the described embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0062] See also Figure 4 ,The vehicle following method at intersection based on multi-sensor fusion includes the following steps:

[0063] S1. See Figure 1, taking the rear bumper center of the target vehicle as the reference point of the target vehicle, in the straight-ahead phase before turning at the intersection, the distance D1 between the ego vehicle and the traffic light at the intersection is obtained through the camera, and the distance D2 between the ego vehicle and the target vehicle in front of the ego vehicle is obtained through the camera and radar;

[0064] When the following conditions are met at the same time, the intersection following mode is entered.

[0065] 1) D1<D 1max ;

[0066] 2) D2<D 2max ;

[0067] 3) The turn signals of the ego vehicle and the target vehicle are consistent;

[0068] Among them, D 1max , D 2max All are preset distance values;

[0069] S2. After entering the intersection following mode, determine the following mode category;

[0070] See also Figure 2 , establish the vehicle coordinate system, with the vehicle's travel direction as the positive x direction, and the direction perpendicular to the x direction to the left as the positive y direction;

[0071] When neither the host vehicle nor the target vehicle has a turn signal, and the forward traffic light is green, the first following mode is selected and S201 is executed;

[0072] When neither the host vehicle nor the target vehicle has a turn signal, and the forward traffic light is red, the second following mode is selected and S202 is executed;

[0073] When both the host vehicle and the target vehicle have turn signal lights, and the left turn or right turn signal light is red, the third following mode is selected and S202 is executed;

[0074] When both the host vehicle and the target vehicle have turn signal signals, and the left or right turn signal light is green, the fourth following mode is selected. The fourth following mode includes a straight phase and a turning phase. The straight phase of the fourth following mode executes S201, and the turning phase of the fourth following mode executes S203;

[0075] S201, obtaining the component dx of the distance between the self-vehicle and the target vehicle in the x direction and the component vx of the speed of the target vehicle relative to the self-vehicle in the x direction through the camera and the radar, and calculating TTC1=dx / vx; controlling the speed of the self-vehicle so that the speed of the self-vehicle does not exceed the speed limit of the road section, and TTC1 is not less than a preset time value;

[0076] S202, obtain the component dx of the distance between the self-vehicle and the target vehicle in the x direction and the component vx of the speed of the target vehicle relative to the self-vehicle in the x direction through the camera and radar, and calculate TTC1=dx / vx; obtain the distance dx between the self-vehicle and the stop line through the camera t , calculate TTC2 = dx t / egov, where egov is the current speed of the vehicle; the speed of the vehicle does not exceed the speed limit of the road section, and TTC1 and TTC2 are both greater than a preset time value;

[0077] S203, execute S2031 to S2033 in sequence;

[0078] S2031, obtain the current driving parameters vx, vy, dx, dy of the target vehicle relative to the ego vehicle through the camera and radar, and obtain the absolute speed v of the target vehicle relative to the ground according to the ego vehicle speed egov and the ego vehicle rotation angular velocity w abs ;Specifically,

[0079]

[0080] Among them, vy is the component of the target vehicle's velocity relative to the ego vehicle in the y direction, and dy is the component of the distance between the ego vehicle and the target vehicle in the y direction;

[0081] The turning action of the target vehicle at the intersection is regarded as v abs The target car moves in a circle at a speed of, and the turning radius L is,

[0082]

[0083] The absolute angular velocity of the target vehicle relative to the ground is w abs =v abs / L;

[0084] T is the duration of a sensor detection cycle. Assuming that the vehicle does not move from the current moment to the next sensor detection cycle after T time, the driving parameters of the target vehicle relative to the vehicle after T time are predicted to be dx′, dy′, vx′, and vy′. Specifically,

[0085]

[0086] See also Figure 3 Since the vehicle will follow the target vehicle in the next sensor detection cycle from the current time to T time later, the predicted driving parameter of the target vehicle relative to the vehicle after T time is dx T ,dy T 、vx T ,vy T , specifically,

[0087] dx T =dx′-L*sin(w*T)

[0088] dy T =dy′-L+L*coS(w*T)

[0089]

[0090] S2032. In the first, second, third following modes and the straight-ahead phase of the fourth following mode of this embodiment, the detection point of the target vehicle by the camera and the radar is the center of the rear bumper of the target vehicle. In the turning phase of the fourth following mode, the posture of the target vehicle changes due to the turning, and the detection point obtained by the camera will jump or disappear, and the detection point obtained by the radar will move to the left rear corner or the right rear corner. At this time, it is necessary to match all objects detected by the radar and the camera to find the target vehicle to follow in the straight-ahead phase and obtain the formal parameters of the target vehicle relative to the self-vehicle; the matching is specifically,

[0091] If the distance matching threshold in the straight-ahead phase of the fourth following mode is GATE, and half of the vehicle width length of the target vehicle is obtained as width, then the distance matching threshold in the turning phase of the fourth following mode is GATE+width;

[0092] Among all objects detected by the camera, objects whose distance to the vehicle is less than the distance matching threshold are regarded as candidate targets, and among all candidate targets, the candidate targets with the same target category as the target vehicle in the previous sensor detection cycle and the closest distance to the vehicle are regarded as target vehicles;

[0093] During the radar matching process, all objects detected by the radar whose distance to the vehicle is less than the distance matching threshold are selected as candidate targets. The radar measurement of the candidate targets is translated. If the target vehicle's turning direction predicted by the previous sensor detection cycle is a left turn, it is translated to the lower left, and if it is a right turn, it is translated to the lower right.

[0094] The horizontal translation distance is

[0095] The vertical translation distance is

[0096] The radar measurements before and after the translation that are closest to the target vehicle's driving parameters relative to the self-vehicle predicted in the previous sensor detection cycle are taken as the matching result;

[0097] The matching algorithm can use the Hungarian algorithm or the nearest neighbor matching method for matching;

[0098] S2033, the target vehicle's driving parameters relative to the vehicle obtained after matching are combined with the target vehicle's driving parameters relative to the vehicle predicted in the previous sensor detection cycle, and the driving parameters dx1, dy1, vx1, vy1 of the target vehicle relative to the vehicle in this sensor detection cycle are obtained; thus, the absolute vehicle speed v of the target vehicle relative to the ground in this sensor detection cycle is obtained. abs1 , the turning radius L1 of the target vehicle, the absolute rotational angular velocity w of the target vehicle relative to the ground abs1 , combined with the vehicle speed egov1 and the angular velocity w1 of the vehicle during this sensor detection cycle; assuming that the target vehicle is on the driving trajectory of the vehicle, then

[0099]

[0100] Make the vehicle speed not exceed the maximum speed required for turning, and make TTC3 greater than a preset time value, and adjust the steering wheel angle of the vehicle according to the angle between the vehicle and the target vehicle when following;

[0101] S3. When the following conditions are met at the same time, exit the intersection following mode.

[0102] 1) The camera recognizes the lane line;

[0103] 2) The lane length is greater than D 2max ;

[0104] 3) There is no stop line within the camera recognition range;

[0105] S4. After exiting the intersection following mode, the vehicle switches to the cruise mode or other automatic driving mode.

[0106] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A vehicle following method at an intersection based on multi-sensor fusion, characterized in that: The following steps are included: S1. In the straight-ahead phase before turning at an intersection, the radar and camera are used to obtain the distance between the vehicle and the target vehicle in front of the vehicle, the distance between the vehicle and the traffic light at the intersection, and the state of the target vehicle's turn signal. Based on the distance between the vehicle and the target vehicle in front of the vehicle, the distance between the vehicle and the traffic light at the intersection, and the state of the target vehicle's turn signal, it is determined whether to enter the intersection following mode; S2. After entering the intersection following mode, different following modes are selected according to whether the turn signal of the own vehicle and the target vehicle is turned on and the color of the forward or turn traffic light; in different following modes, the ego vehicle uses radar and camera to obtain the position and speed of the target vehicle relative to the ego vehicle, and combines the distance between the ego vehicle and the target vehicle in front of the ego vehicle, the distance between the ego vehicle and the traffic light at the intersection, the state of the target vehicle's turn signal, and the driving parameters of the ego vehicle obtained by using radar and camera in S1 to obtain the real-time collision time between the ego vehicle and the target vehicle in different following modes, and then control the speed of the ego vehicle according to the real-time collision time and the speed limit of the ego vehicle's driving section, and adjust the steering wheel angle of the ego vehicle according to the angle between the target vehicle and the driving angle of the ego vehicle; S3. Determine whether to exit the intersection following mode according to the road signs recognized by the camera.

2. The vehicle following method at an intersection based on multi-sensor fusion according to claim 1, characterized in that: The specific judgment process conditions for whether to enter the intersection following mode in S1 are: Take the rear bumper center of the target vehicle as the reference point of the target vehicle. In the straight-ahead phase before turning at the intersection, obtain the distance D1 between the vehicle and the traffic light at the intersection through the camera, and obtain the distance D2 between the vehicle and the target vehicle in front of the vehicle through the camera and radar. When the following conditions are met at the same time, enter the intersection following mode: 1)D1<D 1max ; <h2 style=";text-align:left;direction:ltr">2)D2 <D <h2 style=";text-align:left;direction:ltr"> 2max <h2 style=";text-align:left;direction:ltr"> ; 3) The turn signals of the ego vehicle and the target vehicle are consistent; Among them, D 1max , D 2max These are the preset distance values.

3. The vehicle following method at an intersection based on multi-sensor fusion according to claim 1, characterized in that: The following modes described in S2 include the first following mode, the second following mode, the third following mode, and the fourth following mode; in all following modes, the vehicle coordinate system is established, with the direction of the vehicle traveling as the positive x direction and the direction perpendicular to the x direction to the left as the positive y direction; the activation adjustment of each following mode is respectively, When both the own vehicle and the target vehicle have no turn signal and the forward traffic light is green, the first following mode is selected; when both the own vehicle and the target vehicle have no turn signal and the forward traffic light is red, the second following mode is selected; when both the own vehicle and the target vehicle have turn signal and the left or right turn signal light is red, the third following mode is selected; when both the own vehicle and the target vehicle have turn signal and the left or right turn signal light is green, the fourth following mode is selected, and the fourth following mode includes a straight-ahead phase and a turning phase.

4. The vehicle following method at an intersection based on multi-sensor fusion according to claim 2, characterized in that: In the first following mode, the following operation performed by the vehicle is specifically as follows: The component dx of the distance between the self-vehicle and the target vehicle in the x-direction, as well as the component vx of the speed of the target vehicle relative to the self-vehicle in the x-direction are obtained through the camera and radar, and TTC1=dx / vx is calculated; the speed of the self-vehicle is controlled so that the speed of the self-vehicle does not exceed the speed limit of the road section, and TTC1 is not lower than a preset time value.

5. The vehicle following method at an intersection based on multi-sensor fusion according to claim 2, characterized in that: In the second following mode or the third following mode, the following operation performed by the vehicle is specifically: The x-component dx of the distance between the ego vehicle and the target vehicle and the x-component vx of the speed of the target vehicle relative to the ego vehicle are obtained through the camera and radar, and TTC1 = dx / vx is calculated; the distance dx between the ego vehicle and the stop line is obtained through the camera t , calculate TTC2 = dx t / egov, where egov is the current vehicle speed; the vehicle speed does not exceed the speed limit of the road section, and both TTC1 and TTC2 are greater than a preset time value.

6. The vehicle following method at an intersection based on multi-sensor fusion according to claim 4, characterized in that: In the straight-ahead phase of the fourth following mode, the following operation performed by the vehicle is the same as that in the first following mode. In the turning phase of the fourth following mode, the following operation performed by the vehicle is specifically: First, the driving parameters vx, vy, dx, and dy of the current target vehicle relative to the ego vehicle are obtained through the camera and radar, and the absolute speed v of the target vehicle relative to the ground is obtained according to the ego vehicle speed egov and the ego vehicle rotation angular velocity w. abs ;Specifically, Among them, vy is the component of the target vehicle's velocity relative to the ego vehicle in the y direction, and dy is the component of the distance between the ego vehicle and the target vehicle in the y direction; The target vehicle’s turning action at the intersection is regarded as v abs The target car moves in a circle at a speed of, and the turning radius L is, The absolute angular velocity of the target vehicle relative to the ground is w abs =v abs / L; T is the duration of a sensor detection cycle. Assuming that the vehicle does not move from the current moment to the next sensor detection cycle after T time, the driving parameters of the target vehicle relative to the vehicle after T time are predicted to be dx′, dy′, vx′, and vy′. Specifically, Since the vehicle will follow the target vehicle in the next sensor detection cycle from the current time to T time later, the target vehicle’s driving parameter relative to the vehicle after T time is predicted to be dx T ,dy T 、vx T ,vy T , specifically, dx T =dx′-L*sin(w*T) of T =dy′-L+L*cos(w*T) The radar and camera measurements are then matched to the turning target vehicle. Specifically, If the distance matching threshold in the straight-ahead phase of the fourth following mode is GATE, and half of the vehicle width length of the target vehicle is obtained as width, then the distance matching threshold in the turning phase of the fourth following mode is GATE+width; Among all objects detected by the camera, objects whose distance to the vehicle is less than the distance matching threshold are regarded as candidate targets, and among all candidate targets, the candidate targets with the same target category as the target vehicle in the previous sensor detection cycle and the closest distance to the vehicle are regarded as target vehicles; During the radar matching process, all objects detected by the radar whose distance to the vehicle is less than the distance matching threshold are selected as candidate targets. The radar measurement of the candidate targets is translated. If the target vehicle's turning direction predicted by the previous sensor detection cycle is a left turn, it is translated to the lower left, and if it is a right turn, it is translated to the lower right. The horizontal translation distance is The vertical translation distance is The radar measurements before and after the translation that are closest to the target vehicle's driving parameters relative to the self-vehicle predicted in the previous sensor detection cycle are taken as the matching result; Finally, the target vehicle's driving parameters relative to the vehicle after matching are combined with the target vehicle's driving parameters relative to the vehicle predicted in the previous sensor detection cycle, and the driving parameters dx1, dy1, vx1, and vy1 of the target vehicle relative to the vehicle in this sensor detection cycle are obtained. The absolute speed v of the target vehicle relative to the ground in this sensor detection cycle is obtained. abs1 , the turning radius L1 of the target vehicle, the absolute rotational angular velocity w of the target vehicle relative to the ground abs1 , combined with the vehicle speed egov1 and the angular velocity w1 of the vehicle during this sensor detection cycle; assuming that the target vehicle is on the driving trajectory of the vehicle, then Make the vehicle speed not exceed the maximum speed required for turning, and make TTC3 greater than a preset time value. When following, adjust the steering wheel angle of the vehicle according to the angle between the vehicle and the target vehicle.

7. The vehicle following method at an intersection based on multi-sensor fusion according to claim 2, characterized in that: S3 specifically includes exiting the intersection following mode when the following conditions are met at the same time: 1) The camera recognizes the lane line; 2) The lane length is greater than D 2max ; 3) There is no stop line within the camera's recognition range.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the intersection vehicle following method based on multi-sensor fusion according to any one of claims 1-7 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the intersection vehicle following method based on multi-sensor fusion according to any one of claims 1 to 7 are implemented.

10. An automobile, characterized in that: Comprising a computer device as claimed in claim 8.

Images