Longitudinal control method and device for vehicle
By using the vehicle local prediction reference lines and obstacle prediction trajectory displacement-time-space relationship in vertical control of the vehicle, the problem of low longitudinal control accuracy in the prior art is solved, and more accurate automatic emergency braking and adaptability are achieved Cruise function improves driving safety.
Patent Information
- Application Number
- CN202411298415.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-18
- Publication Date
- 2024-11-19
AI Technical Summary
In the vertical control of vehicles, the accuracy of behavioral decision-making and longitudinal motion control is low in terms of the relationship between the current position of the bicycle and the obstacle in front of the lane, especially in curved scenes. The problem of not being able to brake or slow down in time.
The motion relationship between the local prediction reference line of the bicycle and the obstacle prediction trajectory in the displacement-time space is used to determine whether braking is needed. By obtaining the local prediction reference line of the bicycle and the obstacle prediction trajectory projection and mapping to the displacement-time space, Calculate and execute acceleration.
It improves the accuracy of behavioral decision-making and longitudinal motion control, ensures the accuracy of automatic emergency braking and adaptive cruise in complex scenarios such as curves, and improves driving safety.
Smart Images

Figure CN118977708A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of automobile technology, and more specifically, to a longitudinal control method and device for a vehicle. Background Art
[0002] In the prior art, the vehicle's longitudinal behavior decision and longitudinal motion control are completed by calculating the obstacle in front of the lane based on the current position of the vehicle, or predicting the time distance of the obstacle that will arrive in front of the lane, or adding obstacle trajectory prediction. However, relying only on the relationship between the current position of the vehicle and the obstacle in front of the lane, without the predicted trajectory of the vehicle as a reference, the accuracy of behavior decision and longitudinal motion control is low, especially in the case of automatic emergency braking, adaptive cruise control and other auxiliary driving functions in the curve scene, there are problems such as automatic emergency braking cannot brake in time and adaptive cruise control does not slow down in time. Summary of the invention
[0003] The present application provides a method and device for longitudinal control of a vehicle, which uses the motion relationship between the local predicted reference line of the vehicle and the predicted trajectory of the obstacle in the displacement-time space to determine whether braking is required, thereby improving the accuracy of behavioral decision-making and longitudinal motion control.
[0004] The present application provides a longitudinal control method for a vehicle, comprising:
[0005] Obtaining a local prediction reference line of the ego vehicle corresponding to the real-time position of the ego vehicle;
[0006] Obtaining a first obstacle predicted trajectory based on real-time perception information of the obstacle;
[0007] Projecting the local prediction reference line of the ego vehicle and all the first obstacle prediction trajectories into the ego vehicle coordinate system to obtain a second obstacle prediction trajectory that overlaps with the local prediction reference line of the ego vehicle;
[0008] Map the local predicted reference line of the vehicle and all the predicted trajectories of the second obstacles into the displacement-time space to determine whether acceleration or braking is needed at the moment;
[0009] If so, the acceleration required for acceleration or braking is calculated and sent to the controlled-by-wire chassis for execution.
[0010] Preferably, obtaining a local predicted reference line of the vehicle corresponding to the real-time position of the vehicle specifically includes:
[0011] Fitting a local predicted trajectory corresponding to the real-time position of the ego vehicle based on the current motion state of the ego vehicle;
[0012] The local predicted trajectory is used as the local predicted reference line of the ego vehicle.
[0013] Preferably, obtaining the local predicted reference line of the vehicle corresponding to the real-time position of the vehicle further includes:
[0014] The global prediction trajectory is combined with the local prediction trajectory to form the local prediction reference line of the vehicle.
[0015] Preferably, obtaining the local predicted reference line of the vehicle corresponding to the real-time position of the vehicle further includes:
[0016] The global prediction trajectory and road topology map are integrated with the local prediction trajectory to form the local prediction reference line of the vehicle.
[0017] Preferably, the global predicted trajectory is a global route recommended by a navigation map or a global route on a familiar road map.
[0018] The present application also provides a longitudinal control device for a vehicle, comprising a reference line acquisition module, an obstacle trajectory prediction module, a projection module, a judgment module, and an acceleration calculation module;
[0019] The reference line acquisition module is used to obtain a local prediction reference line of the ego vehicle corresponding to the real-time position of the ego vehicle;
[0020] The obstacle trajectory prediction module is used to obtain a first obstacle prediction trajectory based on real-time perception information of the obstacle;
[0021] The projection module is used to project the local prediction reference line of the ego vehicle and all the first obstacle prediction trajectories into the ego vehicle coordinate system to obtain the second obstacle prediction trajectory that overlaps with the local prediction reference line of the ego vehicle;
[0022] The judgment module is used to map the local prediction reference line of the vehicle and all the second obstacle prediction trajectories into the displacement-time space to judge whether acceleration or braking is needed at present;
[0023] The acceleration calculation module is used to calculate the acceleration required for acceleration or braking when acceleration or braking is required, and send it to the wire-controlled chassis for execution.
[0024] Preferably, the reference line acquisition module is used to fit a local predicted trajectory corresponding to the real-time position of the ego vehicle based on the current motion state of the ego vehicle, and use the local predicted trajectory as the local predicted reference line of the ego vehicle.
[0025] Preferably, the reference line acquisition module is used to merge the global prediction trajectory with the local prediction trajectory to form a local prediction reference line of the ego vehicle.
[0026] Preferably, the reference line acquisition module is used to fuse the global prediction trajectory and the road topology map with the local prediction trajectory to form a local prediction reference line of the vehicle.
[0027] Preferably, the global predicted trajectory is a global route recommended by a navigation map or a global route on a familiar road map.
[0028] Other features and advantages of the present application will become apparent from the following detailed description of exemplary embodiments of the present application with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the application and, together with the description, serve to explain the principles of the application.
[0030] Figure 1 A flow chart of a longitudinal control method for a vehicle provided in the present application;
[0031] Figure 2 This is a structural diagram of the longitudinal control device of the vehicle provided in this application. DETAILED DESCRIPTION
[0032] Various exemplary embodiments of the present application will now be described in detail with reference to the accompanying drawings. It should be noted that unless otherwise specifically stated, the relative arrangement of components and steps, numerical expressions and numerical values set forth in these embodiments do not limit the scope of the present application.
[0033] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the present application, its application, or uses.
[0034] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.
[0035] In all examples shown and discussed herein, any specific values should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values.
[0036] The present application provides a longitudinal control method and device for a vehicle, which uses the motion relationship between the local prediction reference line of the vehicle and the obstacle prediction trajectory in the displacement-time space to determine whether braking is required, thereby improving the accuracy of behavioral decision-making and longitudinal motion control. In addition, in the process of determining the local prediction reference line of the vehicle, the local prediction trajectory and the global prediction trajectory based on the vehicle kinematic model are combined, and it is preferred to further combine the road topology map based on visual perception to determine the local prediction reference line, so that the local prediction reference line is more accurate, and a longer time and more global prediction can be achieved, thereby improving the accuracy of behavioral decision-making and longitudinal motion control, especially in the case of no guide line and curve at the intersection, the auxiliary driving functions such as adaptive cruise and automatic emergency braking can achieve early deceleration or emergency braking, thereby improving driving safety. Finally, the present application combines the local prediction reference line of the vehicle, the obstacle prediction trajectory and the traffic light signal information to determine whether deceleration or emergency braking is currently required, and the longitudinal motion control of the vehicle is achieved through more comprehensive information, which can better liberate the driver.
[0037] It should be noted that the longitudinal control method and device of the vehicle provided in this application are applicable to an autonomous driving domain controller.
[0038] like Figure 1 As shown, the longitudinal control method of the vehicle provided by the present application includes:
[0039] S110: Obtaining a local prediction reference line of the vehicle corresponding to the real-time position of the vehicle.
[0040] Specifically, obtaining a local prediction reference line of the vehicle corresponding to the real-time position of the vehicle includes:
[0041] S1101: Fitting a local predicted trajectory corresponding to the real-time position of the vehicle based on the current motion state of the vehicle.
[0042] Specifically, the current motion state includes the current front wheel angle , the current yaw rate , Current vehicle speed , current longitudinal velocity vs_last, current longitudinal acceleration as and current longitudinal position s_last.
[0043] Each point on the local prediction trajectory is given by the curvature of the point , the longitudinal predicted speed vs_next and the longitudinal predicted position s_next are determined:
[0044] When the vehicle speed is less than 1m / s, ;
[0045] When the vehicle speed is greater than 1.5m / s, ;
[0046] When the vehicle speed is between 1m / s and 1.5m / s, the curvature It is calculated by taking the weighted average of the curvatures when the vehicle speed is 1m / s and 1.5m / s, and the respective weights are determined by the difference between the current vehicle speed and 1m / s and 1.5m / s respectively;
[0047] In the prediction time domain:
[0048] vs_next = vs_last + as * △t,
[0049] s_next = s_last + 0.5*(vs_last + vs_next) * △t;
[0050] The current longitudinal acceleration as of the vehicle has an available time. When the available time is exceeded, as=0.
[0051] S1102: Using the local predicted trajectory as a local predicted reference line for the vehicle.
[0052] Preferably, in S1102, the global prediction trajectory is merged with the local prediction trajectory to form a local prediction reference line of the vehicle.
[0053] As an embodiment, the global predicted trajectory is a global route recommended by a navigation map, specifically a global route ahead of the current position of the vehicle.
[0054] As another embodiment, the global predicted trajectory is a global route on a familiar road map, specifically a global route in front of the current position of the vehicle. The familiar road map is obtained by recording the routes that the driver has frequently driven before. In this application, the autonomous driving domain controller obtains the driving path of a certain distance by matching it with the recorded familiar road map.
[0055] As an embodiment, a weighted average of the distances between corresponding points on the global prediction trajectory and the local prediction trajectory and the X-axis (the length direction of the vehicle, ie, the longitudinal direction) of the current vehicle may be calculated.
[0056] More preferably, in S1102, the global prediction trajectory and the road topology map are fused with the local prediction trajectory to form a local prediction reference line of the vehicle.
[0057] The road topology map is a completely graph-free solution (such as Huawei's PCR road topology inference network) that uses a model to perceive the current road topology structure and generates it in real time, with higher positioning accuracy.
[0058] As an embodiment, the road topology map is matched with the navigation map to locate the current lane position of the vehicle, generate a global prediction trajectory, and then perform a weighted average with the local prediction trajectory to obtain the local prediction reference line of the vehicle.
[0059] S120: Obtain a first obstacle predicted trajectory based on real-time perception information of the obstacle.
[0060] Specifically, the surrounding obstacles are sensed by sensors such as cameras, lidars, and millimeter-wave radars to obtain the current lateral position l_last, current lateral speed v1, current longitudinal speed vs_last2, current longitudinal acceleration as2, and current longitudinal position s_last2 of the obstacle.
[0061] Based on the above, the first obstacle prediction trajectory is determined by the obstacle's lateral prediction position l_next, longitudinal prediction speed vs_next2 and longitudinal prediction position s_next2:
[0062] In the prediction time domain:
[0063] l_next = l_last + vl * △t;
[0064] The current horizontal position l_last has available time. When the available time is exceeded, it is considered that vl=0;
[0065] vs_next2 = vs_last2 + as2 * △t,
[0066] s_next2 = s_last2 + 0.5*(vs_last2 + vs_next2) * △t;
[0067] The current longitudinal acceleration as2 of the obstacle has an available time. When the available time is exceeded, as2=0.
[0068] S130: Projecting the local prediction reference line of the own vehicle and all the first obstacle prediction trajectories into the own vehicle coordinate system to obtain the second obstacle prediction trajectories that overlap with the local prediction reference line of the own vehicle (i.e., have a collision risk), thereby filtering out the first obstacle prediction trajectories that have no collision risk with the local prediction reference line of the own vehicle.
[0069] S140: Map the local prediction reference line of the vehicle and all the second obstacle prediction trajectories into the displacement-time space (ST space) to determine whether acceleration or braking (including normal deceleration and emergency braking) is currently required. If so, execute S150; otherwise, drive according to the local prediction reference line of the vehicle.
[0070] In ST space, the ST diagram (displacement-time diagram) is used to describe the longitudinal movement of obstacles around the local prediction reference line, and the obstacle avoidance strategy is determined based on it. The current behavior decision is determined based on the avoidance strategy, that is, whether to accelerate, brake or maintain the vehicle speed.
[0071] Preferably, in S140, by combining the local predicted reference line of the vehicle, the second obstacle predicted trajectory and the traffic light signal information, it is determined whether acceleration or braking is currently required. In this preferred embodiment, not only obstacle avoidance is considered, but also the impact of traffic lights on behavioral decision-making is considered, making the behavioral decision more comprehensive and improving its reliability.
[0072] S150: Calculate the acceleration required for acceleration or braking, and send it to the control-by-wire chassis for execution. Specifically, the vehicle is driven according to the planned speed of the acceleration. It should be noted that the planned speed needs to be lower than the maximum speed set by the driver.
[0073] Based on the above, if Figure 2 As shown, the present application also provides a longitudinal control device for a vehicle, including a reference line acquisition module 210 , an obstacle trajectory prediction module 220 , a projection module 230 , a judgment module 240 and an acceleration calculation module 250 .
[0074] The reference line obtaining module 210 is used to obtain a local predicted reference line of the vehicle corresponding to the real-time position of the vehicle.
[0075] The obstacle trajectory prediction module 220 is used to obtain a first obstacle prediction trajectory based on real-time perception information of the obstacle.
[0076] The projection module 230 is used to project the local prediction reference line of the ego vehicle and all the first obstacle prediction trajectories into the ego vehicle coordinate system to obtain a second obstacle prediction trajectory that overlaps with the local prediction reference line of the ego vehicle.
[0077] The judgment module 240 is used to map the local prediction reference line of the vehicle and all the second obstacle prediction trajectories into the displacement-time space to judge whether acceleration or braking is currently required.
[0078] The acceleration calculation module 250 is used to calculate the acceleration required for acceleration or braking when acceleration or braking is required, and send it to the control-by-wire chassis for execution.
[0079] Preferably, the reference line obtaining module 210 is used to fit a local predicted trajectory corresponding to the real-time position of the ego vehicle based on the current motion state of the ego vehicle, and use the local predicted trajectory as the local predicted reference line of the ego vehicle.
[0080] Preferably, the reference line obtaining module 210 is used to merge the global prediction trajectory with the local prediction trajectory to form a local prediction reference line of the ego vehicle.
[0081] Preferably, the reference line acquisition module 210 is used to merge the global prediction trajectory and the road topology map with the local prediction trajectory to form a local prediction reference line of the vehicle.
[0082] Preferably, the global predicted trajectory is a global route recommended by a navigation map or a global route on a familiar road map.
[0083] Preferably, the judgment module 240 is used to determine whether acceleration or braking is currently required by combining the local predicted reference line of the vehicle, the second obstacle predicted trajectory and the traffic light signal information.
[0084] Although some specific embodiments of the present application have been described in detail by way of example, it should be understood by those skilled in the art that the above examples are only for illustration, not for limiting the scope of the present application. It should be understood by those skilled in the art that the above embodiments may be modified without departing from the scope and spirit of the present application. The scope of the present application is defined by the appended claims.
Claims
1. A longitudinal control method for a vehicle, characterized in that: include: Obtaining a local prediction reference line of the ego vehicle corresponding to the real-time position of the ego vehicle; Obtaining a first obstacle predicted trajectory based on real-time perception information of the obstacle; Projecting the local prediction reference line of the ego vehicle and all the first obstacle prediction trajectories into the ego vehicle coordinate system to obtain a second obstacle prediction trajectory that overlaps with the local prediction reference line of the ego vehicle; Mapping the local predicted reference line of the vehicle and all the predicted trajectories of the second obstacles into the displacement-time space to determine whether acceleration or braking is currently required; If so, the acceleration required for acceleration or braking is calculated and sent to the controlled-by-wire chassis for execution.
2. The longitudinal control method of a vehicle according to claim 1, characterized in that: Obtaining the local prediction reference line of the vehicle corresponding to the real-time position of the vehicle, specifically including: Fitting a local predicted trajectory corresponding to the real-time position of the ego vehicle based on the current motion state of the ego vehicle; The local predicted trajectory is used as the local predicted reference line of the vehicle.
3. The longitudinal control method of a vehicle according to claim 2, characterized in that: Obtaining a local prediction reference line of the ego vehicle corresponding to the real-time position of the ego vehicle, further comprising: The global prediction trajectory is merged with the local prediction trajectory to form the local prediction reference line of the vehicle.
4. The longitudinal control method of a vehicle according to claim 2, characterized in that: Obtaining a local prediction reference line of the ego vehicle corresponding to the real-time position of the ego vehicle, further comprising: The global prediction trajectory and the road topology map are merged with the local prediction trajectory to form the local prediction reference line of the vehicle.
5. The longitudinal control method of a vehicle according to claim 3 or 4, characterized in that: The global predicted trajectory is a global route recommended by a navigation map or a global route on a familiar road map.
6. A longitudinal control device for a vehicle, characterized in that: It includes a reference line acquisition module, an obstacle trajectory prediction module, a projection module, a judgment module and an acceleration calculation module; The reference line acquisition module is used to obtain a local prediction reference line of the vehicle corresponding to the real-time position of the vehicle; The obstacle trajectory prediction module is used to obtain a first obstacle prediction trajectory based on real-time perception information of the obstacle; The projection module is used to project the local prediction reference line of the ego vehicle and all the first obstacle prediction trajectories into the ego vehicle coordinate system to obtain a second obstacle prediction trajectory that overlaps with the local prediction reference line of the ego vehicle; The judgment module is used to map the local prediction reference line of the vehicle and all the second obstacle prediction trajectories into the displacement-time space to judge whether acceleration or braking is currently required; The acceleration calculation module is used to calculate the acceleration required for acceleration or braking when acceleration or braking is required, and send it to the control-by-wire chassis for execution.
7. The longitudinal control device of a vehicle according to claim 6, characterized in that: The reference line acquisition module is used to fit a local prediction trajectory corresponding to the real-time position of the vehicle based on the current motion state of the vehicle, and use the local prediction trajectory as the local prediction reference line of the vehicle.
8. The longitudinal control device for a vehicle according to claim 7, characterized in that: The reference line acquisition module is used to merge the global prediction trajectory with the local prediction trajectory to form the local prediction reference line of the vehicle.
9. The longitudinal control device for a vehicle according to claim 7, characterized in that: The reference line acquisition module is used to fuse the global prediction trajectory and the road topology map with the local prediction trajectory to form the local prediction reference line of the vehicle.
10. The longitudinal control device of a vehicle according to claim 8 or 9, characterized in that: The global predicted trajectory is a global route recommended by a navigation map or a global route on a familiar road map.
