Vehicle control methods, automated driving control systems and vehicles
By uniformly modeling vehicles and obstacles within a four-dimensional transportation system, the problem of dynamic interaction and influence of multiple obstacles in complex traffic environments is solved, thereby improving the accuracy and safety of vehicle trajectory prediction.
Patent Information
- Application Number
- CN202511209238.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-27
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2045-08-27
AI Technical Summary
Existing technologies struggle to accurately reflect the dynamic interactions between multiple obstacles in complex traffic environments, leading to delayed responses in autonomous vehicle control strategies and posing safety hazards.
A four-dimensional transportation system is constructed by uniformly modeling vehicles, target obstacles, and traffic facilities in a four-dimensional spatial coordinate system, comprehensively considering the dynamic behavioral interaction relationships between obstacles, generating the predicted movement trajectory of multiple target obstacles within a preset time period in the future, and deriving the predicted movement trajectory of the vehicle.
It improves the accuracy and rationality of predicting the movement trajectory of various target obstacles and autonomous vehicles in complex traffic environments, and enhances the safety and foresight of autonomous vehicle control decisions.
Smart Images

Figure CN120735799B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of vehicle technology, and in particular to a vehicle control method, an automatic driving control system, and a vehicle. Background Technology
[0002] With the development of intelligent driving technology, vehicles' ability to perceive their surroundings during driving has been significantly enhanced. This is especially true in autonomous driving or advanced driver assistance systems (ADAS), which often utilize multi-sensor fusion techniques such as LiDAR, millimeter-wave radar, and cameras to detect and track surrounding obstacles. Existing technologies have already achieved the ability to predict the movement trajectory of obstacles over a short period and adjust the vehicle's driving behavior accordingly to improve driving safety.
[0003] However, in complex traffic environments, such as urban roads, multi-vehicle intersections, and high-speed lane changes, obstacles (including other vehicles, non-motorized vehicles, or pedestrians) often exhibit dynamic behavioral interactions. This means that a change in the behavior of one obstacle can influence the movement decisions of surrounding obstacles. Existing technologies are mostly based on trajectory prediction of single obstacles, making it difficult to accurately reflect the dynamic evolution of these interactions between multiple obstacles, and consequently, to comprehensively reflect the dynamic interactive changes in the overall traffic scenario. If an obstacle undergoes a sudden behavioral change (such as rapid acceleration or emergency lane changes), this limitation prevents the system from fully understanding the dynamic interaction relationships between multiple obstacles and their impact on the vehicle's path, leading to a lag in the vehicle's control strategy response and potential safety hazards. Summary of the Invention
[0004] This invention aims to at least solve one of the technical problems existing in the prior art. To this end, one objective of this invention is to propose a vehicle control method that realizes the construction of a four-dimensional traffic system, comprehensively considering the impact of the behavior changes of one obstacle on the behavior decisions of other obstacles, improving the accuracy and rationality of predicting the movement trajectories of various target obstacles and the vehicle in complex traffic environments, thereby enhancing the safety of the vehicle's driving decisions.
[0005] The second objective of this invention is to propose an automatic driving control system.
[0006] The third objective of this invention is to provide a vehicle.
[0007] To achieve the above objectives, a vehicle control method according to a first aspect of the present invention includes: obtaining the predicted movement trajectories of various target obstacles within a future preset time period under a four-dimensional traffic system, wherein the four-dimensional traffic system is constructed based on obstacle perception information and traffic facilities in a four-dimensional spatial coordinate system of an autonomous driving perception system of a vehicle, and various obstacles and the vehicle are respectively presented in the four-dimensional traffic system, and the future preset time period is a set time period after the current moment; the predicted movement trajectories of various target obstacles within the future preset time period are generated based on the dynamic behavioral interaction relationships between the various obstacles, and the dynamic behavioral interaction relationships include the influence of the behavior change of one obstacle on the behavior decisions of other obstacles; determining the predicted movement trajectory of the vehicle based on the predicted movement trajectories of each target obstacle within the future preset time period, and controlling the driving of the vehicle based on the predicted movement trajectory of the vehicle.
[0008] According to the vehicle control method of this invention, a unified four-dimensional traffic system is constructed, and the vehicle, target obstacles, and traffic facilities are uniformly modeled in a four-dimensional spatial coordinate system. This enables the vehicle to have a holistic perception capability of environmental spatiotemporal changes, thereby providing an accurate and continuous spatiotemporal reference basis for subsequent behavior analysis and trajectory prediction. In this four-dimensional traffic system, the predicted movement trajectory of each target obstacle is not only generated based on its own historical behavior data, but also comprehensively considers the influence of the behavior changes of other obstacles on its motion decision, realizing the modeling and coupled prediction of dynamic behavior interaction relationships. This allows the trajectory prediction model to accurately generate the predicted movement trajectories of multiple target obstacles under the mutual influence of multiple target obstacles within a preset future time. Furthermore, the predicted movement trajectories of these obstacles are used to deduce the predicted movement trajectory of the vehicle within a preset future time. This allows the vehicle's decision path to be based not only on its own state and static environment, but also on the dynamic response to the interaction changes between surrounding obstacles. This improves the accuracy and rationality of the predicted movement trajectories of various target obstacles and the vehicle in complex traffic environments, better reflects the actual dynamic interaction behavior in complex traffic environments, and ultimately enhances the safety and foresight of the vehicle's control decisions.
[0009] In some embodiments, the four-dimensional spatial coordinate system is constructed with the vehicle as the origin, the lane line direction as the x-axis, the direction perpendicular to the lane line direction as the y-axis, the vehicle's height direction as the z-axis, and the time axis.
[0010] In some embodiments, the vehicle control method further includes: determining the degree of danger posed by each dynamic obstacle to the vehicle based on the motion state information of obstacles in the four-dimensional traffic system; and determining the target obstacle based on the degree of danger.
[0011] In some embodiments, the target obstacle is a dynamic obstacle whose danger level is higher than a preset danger level threshold.
[0012] In some embodiments, the vehicle control method further includes: determining the predicted movement trajectory of each of the target obstacles based on the degree of danger posed by each target obstacle to the vehicle.
[0013] In some embodiments, determining the predicted movement trajectory of each target obstacle based on the degree of danger posed by each target obstacle to the vehicle includes: the predicted movement trajectory of a first target obstacle is determined based on the historical movement trajectory of the first target obstacle, the position of a static obstacle, and the historical movement trajectory of the vehicle, wherein the historical movement trajectory of the first target obstacle is the movement trajectory of the first target obstacle within a preset historical time period, and the preset historical time period is a set time period prior to the current moment; the predicted movement trajectory of a second target obstacle is determined based on the already determined predicted movement trajectory of a target obstacle, the historical movement trajectory of the second target obstacle, the position of the static obstacle, and the historical movement trajectory of the vehicle; the historical movement trajectory is determined based on the movement speed, obstacle shape characteristics, and movement direction of the target obstacle or the vehicle within the preset historical time period.
[0014] In some embodiments, the first target obstacle is the target obstacle with the highest degree of danger to the vehicle among the plurality of target obstacles, or, when any dynamic obstacle changes and the predicted movement trajectory of the target obstacle and the vehicle is re-determined, the first target obstacle is the changed obstacle; the second target obstacle is a target obstacle other than the first target obstacle among the plurality of target obstacles.
[0015] In some embodiments, the predicted movement trajectories of each of the target obstacles are presented as multiple parallel movement trajectory curves.
[0016] To achieve the above objectives, an autonomous driving control system according to a second aspect of the present invention includes: a perception system, the perception system including a lidar, a forward-facing camera, a forward-facing mid-range camera, a forward-facing close-range camera, a side-view camera and a rear-view camera, a forward-facing millimeter-wave radar, an angular millimeter-wave radar, a rear-facing millimeter-wave radar and a moving speed detector; and a controller connected to the perception system for executing the vehicle control method according to any one of claims 1-8.
[0017] According to an embodiment of the autonomous driving control system of the present invention, the controller is connected to the perception system and adopts the vehicle control method described in the above embodiment. By constructing a unified four-dimensional traffic system, the vehicle, target obstacles, and traffic facilities are uniformly modeled in a four-dimensional spatial coordinate system, enabling the vehicle to have a holistic perception capability of environmental spatiotemporal changes, thereby providing an accurate and continuous spatiotemporal reference basis for subsequent behavior analysis and trajectory prediction. In this four-dimensional traffic system, the predicted movement trajectory of each target obstacle is not only generated based on its own historical behavior data, but also comprehensively considers the influence of the behavior changes of other obstacles on its motion decision, realizing the modeling and coupling prediction of dynamic behavior interaction relationships. This allows the trajectory prediction model to accurately generate the predicted movement trajectory of multiple target obstacles under the mutual influence of multiple target obstacles within a preset time period in the future. Furthermore, the predicted movement trajectory of the vehicle within a preset time period in the future is derived using the predicted movement trajectory of these obstacles. This makes the vehicle's decision path not only based on its own state and static environment, but also dynamically responding to the interaction changes between surrounding obstacles, thereby improving the accuracy and rationality of the predicted movement trajectory of each target obstacle and the vehicle in complex traffic environments, more closely conforming to the actual dynamic interaction behavior in complex traffic environments, and ultimately enhancing the safety and foresight of the vehicle's control decision.
[0018] To achieve the above objectives, a vehicle according to a third aspect of the present invention is provided, which is used to implement the vehicle control method described in the above embodiments, or the vehicle includes the automatic driving control system described in the above embodiments.
[0019] According to embodiments of the present invention, the vehicle employs the vehicle control method described in the above embodiments, or the autonomous driving control system described in the above embodiments. By constructing a unified four-dimensional traffic system, the vehicle, target obstacles, and traffic facilities are uniformly modeled in a four-dimensional spatial coordinate system. This enables the vehicle to possess a holistic perception capability of environmental spatiotemporal changes, thereby providing an accurate and continuous spatiotemporal reference basis for subsequent behavior analysis and trajectory prediction. In this four-dimensional traffic system, the predicted movement trajectory of each target obstacle is not only generated based on its own historical behavior data, but also comprehensively considers the influence of the behavior changes of other obstacles on its motion decision, realizing the modeling and coupled prediction of dynamic behavior interaction relationships. This allows the trajectory prediction model to accurately generate the predicted movement trajectories of multiple target obstacles under the mutual influence of multiple target obstacles within a preset future time period. Furthermore, by using the predicted movement trajectories of these obstacles to deduce the predicted movement trajectory of the vehicle within a preset time period in the future, the vehicle's decision path is not only based on its own state and static environment, but also dynamically responds to the changes in the interaction between surrounding obstacles. This improves the accuracy and rationality of the predicted movement trajectories of various target obstacles and the vehicle in complex traffic environments, making it more in line with the actual dynamic interaction behavior in complex traffic environments, and ultimately enhancing the safety and foresight of the vehicle's control decisions.
[0020] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description
[0021] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a flowchart of a vehicle control method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a target obstacle and a vehicle in a four-dimensional transportation system according to an embodiment of the present invention; Figure 3 This is a block diagram of an autonomous driving control system according to an embodiment of the present invention; Figure 4 This is a block diagram of a vehicle according to an embodiment of the present invention.
[0022] Figure label: 100 vehicles; Automated driving control system 1; Sensing system 11; controller 12. Detailed Implementation
[0023] The embodiments of the present invention are described in detail below. The embodiments described with reference to the accompanying drawings are exemplary. The embodiments of the present invention are described in detail below.
[0024] The following is for reference. Figures 1-2 A vehicle control method according to an embodiment of the present invention is described.
[0025] Figure 1 This is a flowchart of a vehicle control method according to an embodiment of the present invention, such as... Figure 1 As shown, the vehicle control method of this embodiment includes at least steps S1-S2.
[0026] S1 obtains the predicted movement trajectory of each target obstacle in the four-dimensional transportation system within a preset time period in the future.
[0027] In some embodiments, a four-dimensional transportation system can refer to a dynamic spatiotemporal transportation model constructed by adding a time dimension to a three-dimensional spatial coordinate system (X, Y, Z). The four-dimensional transportation system is constructed based on obstacle perception information and traffic facilities from the autonomous driving perception system of the vehicle in a four-dimensional spatial coordinate system. Various obstacles (such as other vehicles, pedestrians, non-motorized vehicles, etc.) and the vehicle are presented in a unified four-dimensional transportation system, thus possessing temporal continuity and spatial consistency, and being able to express its historical movement trajectory and predicted movement trajectory.
[0028] In some embodiments, the future preset duration is a set duration after the current moment. That is, a predicted time window extending from the current moment into the future, for example, 5 to 10 seconds, and this duration can be dynamically adjusted according to the complexity of the traffic scenario.
[0029] In some embodiments, obstacle perception information may include, but is not limited to, the obstacle's moving speed, obstacle shape features, moving direction, and location.
[0030] In some embodiments, traffic facilities may include, but are not limited to: road structural facilities (such as lane lines, ramps, central dividers, etc.), traffic safety facilities (such as warning posts, guardrails, reflective markers, speed limit devices, etc.), and traffic control facilities (such as traffic lights, traffic signs, etc.).
[0031] In some embodiments, the predicted movement trajectory of each target obstacle within a preset time period is generated based on the dynamic behavioral interaction relationships between various obstacles. These dynamic behavioral interaction relationships include the impact of a change in the behavior of one obstacle (such as a sudden lane change, rapid acceleration or deceleration) on the behavioral decisions of other obstacles (such as triggering avoidance, deceleration, or yielding by other obstacles). For example, if a vehicle in front suddenly changes lanes, the vehicle behind may follow suit by changing lanes or decelerating; if a pedestrian approaches the lane, a vehicle may anticipate their crossing and decelerate in advance; if a non-motorized vehicle suddenly cuts in, vehicles on both sides may also take evasive action. Therefore, the trajectory prediction of each obstacle should consider not only its own historical movement trajectory but also the predicted movement trajectories of other obstacles.
[0032] S2 determines the predicted movement trajectory of the vehicle based on the predicted movement trajectory of each target obstacle within a preset time period in the future, and controls the vehicle's movement based on the predicted movement trajectory of the vehicle.
[0033] In some embodiments, controlling the vehicle's movement may include, but is not limited to, actions such as acceleration, deceleration, avoidance, and lane changing. This control decision is based on the vehicle's predicted trajectory, which in turn references the predicted trajectories of multiple target obstacles within a preset timeframe in a four-dimensional traffic system, thereby greatly improving the accuracy and rationality of the prediction.
[0034] According to the vehicle control method of this invention, a unified four-dimensional traffic system is constructed, and the vehicle, target obstacles, and traffic facilities are uniformly modeled in a four-dimensional spatial coordinate system. This enables the vehicle to have a holistic perception capability of environmental spatiotemporal changes, thereby providing an accurate and continuous spatiotemporal reference basis for subsequent behavior analysis and trajectory prediction. In this four-dimensional traffic system, the predicted movement trajectory of each target obstacle is not only generated based on its own historical behavior data, but also comprehensively considers the influence of the behavior changes of other obstacles on its motion decision, realizing the modeling and coupled prediction of dynamic behavior interaction relationships. This allows the trajectory prediction model to accurately generate the predicted movement trajectories of multiple target obstacles under the mutual influence of multiple target obstacles within a preset future time. Furthermore, the predicted movement trajectories of these obstacles are used to deduce the predicted movement trajectory of the vehicle within a preset future time. This allows the vehicle's decision path to be based not only on its own state and static environment, but also on the dynamic response to the interaction changes between surrounding obstacles. This improves the accuracy and rationality of the predicted movement trajectories of various target obstacles and the vehicle in complex traffic environments, better reflects the actual dynamic interaction behavior in complex traffic environments, and ultimately enhances the safety and foresight of the vehicle's control decisions.
[0035] In some embodiments, the four-dimensional spatial coordinate system is constructed with the vehicle as the origin, the lane line direction as the x-axis, the direction perpendicular to the lane line direction as the y-axis, the vehicle's height direction as the z-axis, and the time axis.
[0036] In some embodiments, the purpose of using the vehicle as the origin of the four-dimensional spatial coordinate system is to facilitate the accurate description of the relative positional relationship between the vehicle and surrounding obstacles, and to ensure that the system dynamically updates the positional data of each obstacle with the vehicle as the center during real-time calculation. This allows the coordinate system to be updated as the vehicle moves (similar to a "vehicle local coordinate system"), thereby enhancing the robustness and real-time performance of trajectory prediction.
[0037] In some embodiments, the x-axis, defined by the lane line direction, can refer to the direction of extension of the road where the vehicle is currently located (e.g., the direction of travel on a highway) as the x-axis of the coordinate system, corresponding to the vehicle's forward direction. The advantage of this axial definition is that it aligns with the road structure, facilitates modeling the relative positions of obstacles and the road, and allows for a more natural description of the obstacle's movement along the lane line direction, such as parallel driving, acceleration, and deceleration. It also facilitates the differentiation and modeling of obstacles within and between lanes.
[0038] In some embodiments, the y-axis is defined as the direction perpendicular to the lane lines, indicating that the direction of the y-axis is perpendicular to the vehicle's direction of travel and represents the lateral offset between the obstacle and the lane centerline. In multi-lane scenarios, the y-axis can effectively capture lateral dynamic behaviors such as vehicle lane changes, lateral avoidance, non-motorized vehicle crossings, and pedestrian crossings, which is of great significance for improving the system's ability to identify lateral collision risks.
[0039] In some embodiments, the z-axis is defined as the height direction of the vehicle. The z-axis is used to describe the vertical height position of obstacles in space, such as overpasses, height restriction poles, ramps, pedestrian bridges, and dump truck bed lifting scenarios. The advantage of this definition is that it makes the traffic environment modeling more complete, avoids misjudgments caused by using only planar coordinates, and enhances the system's ability to identify and control obstacles in complex three-dimensional traffic environments.
[0040] In some embodiments, a "time axis (t)" is added as a fourth dimension to record the spatial state changes of obstacles over time. Each target can be represented in this four-dimensional coordinate system as a set of three-dimensional position points (x, y, z, t) that change over time, thereby enabling modeling of its motion trend and trajectory prediction. The introduction of the time dimension gives the system dynamic understanding capabilities, meaning it can not only acquire the state at a certain moment but also continuously track the behavior trajectory of obstacles, supporting trajectory prediction and behavior analysis.
[0041] By constructing the aforementioned four-dimensional coordinate system, unified and dynamic modeling of traffic participants and infrastructure can be achieved within the autonomous driving control system of the vehicle. The system can continuously track the behavioral evolution of multiple targets in four-dimensional space, thereby providing more accurate and timely perception data support for the vehicle's path planning and driving control.
[0042] In some embodiments, the vehicle control method further includes: determining the degree of danger posed by each dynamic obstacle to the vehicle based on the motion state information of obstacles in a four-dimensional traffic system, and determining the target obstacle based on the degree of danger.
[0043] In some embodiments, the obstacle's motion status information may include, but is not limited to, the obstacle's current speed, direction of movement, and positional changes. This information can be collected in real time by vehicle-mounted sensors such as lidar, millimeter-wave radar, and cameras.
[0044] In some embodiments, the degree of danger posed by each dynamic obstacle to the vehicle can be expressed as the risk of potential conflict, interference, or collision between the vehicle and the obstacle within a preset time period in the future. This degree of danger can be quantitatively assessed using a defined danger rating model or collision probability assessment mechanism.
[0045] In some embodiments, a target obstacle may refer to a dynamic obstacle that has a significant impact on the safety of a vehicle's driving path in a four-dimensional transportation system. A target obstacle may be one or more dynamic obstacles.
[0046] Specifically, the system can use various sensors mounted on the vehicle (such as LiDAR, cameras, millimeter-wave radar, etc.) to collect real-time motion status information (such as the current speed, direction of movement, and position changes of obstacles) of various obstacles in the four-dimensional traffic system. Based on this information, the system can obtain the historical movement trajectory of each obstacle and predict its future trajectory. At the same time, it can calculate the relative distance, relative speed, and trajectory overlap between each obstacle and the vehicle, and input these parameters into a preset danger level scoring model. The model can output a score for each obstacle that reflects its danger level and compare it with a preset danger level threshold to determine whether to include it in the target obstacle set.
[0047] In some embodiments, the target obstacle is a dynamic obstacle whose danger level is higher than a preset danger level threshold. A dynamic obstacle can refer to an object that exhibits a tendency to change position over time, i.e., a non-stationary target. Examples include other vehicles in motion, pedestrians or animals walking or running, electric vehicles or bicycles being ridden, and objects with a tendency to fall or roll.
[0048] In some embodiments, the preset danger level threshold can be a reference standard set by the system to classify "high-risk obstacles" and "negligible obstacles". This threshold can be statically configured or dynamically adjusted according to the actual driving environment (such as road type, weather conditions and other real-time factors) to improve the robustness and sensitivity of the system.
[0049] In some embodiments, the preset danger threshold can be set as a score, such as 0.6. When the danger score of a dynamic obstacle is higher than 0.6, the obstacle can be identified as a target obstacle and considered to pose a high threat to the vehicle. The system will then focus on its dynamic behavior in the path planning and control strategy to ensure the vehicle's driving safety.
[0050] In some embodiments, the vehicle control method further includes: determining the predicted movement trajectory of each target obstacle based on the degree of danger posed by each target obstacle to the vehicle.
[0051] Specifically, the system scores and ranks the hazards posed by each obstacle to the vehicle, generating an ordered list from highest to lowest hazard level. This list represents the order of the potential risk posed by each obstacle to the vehicle. The predicted trajectories of each obstacle are then determined sequentially according to the ranking. In other words, the system first predicts the trajectories of obstacles ranked highest in hazard level, and then uses these trajectories as environmental constraints to input into the trajectory prediction model of the next obstacle. This step-by-step, interactive trajectory prediction approach more realistically simulates the behavioral impact between obstacles in complex traffic scenarios, improving overall prediction accuracy and the reliability of the vehicle's control strategy.
[0052] In some embodiments, multiple target obstacles may have dynamic behavioral coupling relationships, such as following relationships, avoidance relationships, and the influence of traffic yielding rules. Therefore, when predicting the future trajectory of a target obstacle, the trajectories of other obstacles that have already been predicted and are ranked higher can be referenced to form a "progressive trajectory reasoning". For example, if the first-ranked target obstacle is an overtaking vehicle, the system first predicts its lane-changing trajectory; the second-ranked target obstacle is the vehicle being overtaken, and its trajectory may be affected by the former (such as slowing down or yielding to the right); the third-ranked obstacle is a cyclist, and its path may be affected by the trajectories of the above two. This approach can effectively capture the traffic interaction relationships between obstacles and improve the overall prediction rationality.
[0053] In some embodiments, determining the predicted movement trajectory of each target obstacle based on its degree of danger to the vehicle includes: the predicted movement trajectory of a first target obstacle is determined based on the historical movement trajectory of the first target obstacle, the position of a static obstacle, and the historical movement trajectory of the vehicle, wherein the historical movement trajectory is the movement trajectory within a preset historical time period, and the preset historical time period is a set time period prior to the current moment. The predicted movement trajectory of a second target obstacle is determined based on the already determined predicted movement trajectory of the target obstacle, the position of a static obstacle, and the historical movement trajectory of the vehicle, wherein the historical movement trajectory is determined based on the movement speed, obstacle shape characteristics, and movement direction of the target obstacle or the vehicle within the preset historical time period.
[0054] The first target obstacle is the one with the highest danger level to the vehicle among multiple target obstacles. Specifically, when all target obstacles and the vehicle's operating state are relatively stable (i.e., no significant sudden changes in movement, such as sharp acceleration, deceleration, turning, or lane change, occur at the current time), the system can rank all target obstacles based on a "danger level scoring mechanism," and the target obstacle with the highest danger level score is set as the "first target obstacle." Based on this, a sequential prediction mechanism is used to construct the predicted movement trajectory of each target obstacle. The predicted movement trajectory of subsequent target obstacles will refer to the already predicted movement trajectories to avoid overlapping or collisions, thereby achieving more realistic and robust traffic modeling.
[0055] In some embodiments, the second target obstacle is a target obstacle other than the first target obstacle among a plurality of target obstacles. The second target obstacle can be one or more target obstacles. When the second target obstacle is multiple target obstacles, these target obstacles also need to be ranked according to their degree of danger to the vehicle to determine the order in which the predicted movement trajectories of each target obstacle are determined. Each "sub-predicted target" must consider the impact of "all predicted targets" to improve the rationality and safety of the behavior modeling.
[0056] For example, such as Figure 2 As shown, the system detects four target obstacles at the current moment: obstacle 1, obstacle 2, obstacle 3, and obstacle 4. By acquiring the motion state information of these four obstacle obstacles, the system uses a preset hazard rating model to score the degree of danger between them and the vehicle. The system first uses the motion status information (including speed, direction, shape, and relative position) of each obstacle obtained from the four-dimensional traffic system to score the degree of danger between the obstacle and the vehicle using a pre-set hazard assessment model. The resulting order of danger from highest to lowest is: Target Obstacle 1 > Target Obstacle 2 > Target Obstacle 3 > Target Obstacle 4. Target Obstacle 1 is the first target obstacle, and Target Obstacles 2, 3, and 4 are the second target obstacles.
[0057] Furthermore, by monitoring the moving speed, shape features, and direction of movement of target obstacle 1 within a preset historical time period (e.g., 3-5 seconds), a historical trajectory of target obstacle 1 can be constructed using a four-dimensional spatial coordinate system (including the time axis). Similarly, by monitoring the moving speed, shape features, and direction of movement of the vehicle within a preset historical time period (e.g., 3-5 seconds), a historical trajectory of the vehicle can be constructed using a four-dimensional spatial coordinate system (including the time axis). Combining the spatial position of static obstacles in the coordinate system, and inputting these parameters into a prediction trajectory model (e.g., LSTM, Transformer, Bayesian prediction network, etc.), a predicted trajectory of target obstacle 1 within a future preset time period (e.g., 5-10 seconds) can be obtained.
[0058] Furthermore, by monitoring the moving speed, shape characteristics, and direction of movement of the target obstacle 2 within a preset historical time period (e.g., 3-5 seconds), the historical movement trajectory of the target obstacle 2 can be constructed using a four-dimensional spatial coordinate system (including the time axis). Combining the spatial position of the static obstacle in the coordinate system, the historical movement trajectory of the vehicle, and the predicted movement trajectory of the target obstacle 1 within a preset future time period (e.g., 5-10 seconds), the above parameters are input into the prediction trajectory model to obtain the predicted movement trajectory of the target obstacle 2 within the preset future time period (e.g., 5-10 seconds).
[0059] Furthermore, by monitoring the moving speed, shape characteristics, and direction of movement of target obstacle 3 within a preset historical time period (e.g., 3-5 seconds), the historical movement trajectory of target obstacle 3 can be constructed using a four-dimensional spatial coordinate system (including the time axis). Combining the spatial position of static obstacles in the coordinate system, the historical movement trajectory of the vehicle, and the predicted movement trajectories of target obstacle 1 and target obstacle 2 within a future preset time period (e.g., 5-10 seconds), the above parameters are input into the prediction trajectory model to obtain the predicted movement trajectory of target obstacle 3 within the future preset time period (e.g., 5-10 seconds).
[0060] Furthermore, by monitoring the moving speed, shape characteristics, and direction of movement of target obstacle 4 within a preset historical time period (e.g., 3-5 seconds), the historical movement trajectory of target obstacle 4 can be constructed using a four-dimensional spatial coordinate system (including the time axis). Combining the spatial position of static obstacles in the coordinate system, the historical movement trajectory of the vehicle, and the predicted movement trajectories of target obstacle 1, target obstacle 2, and target obstacle 3 within a future preset time period (e.g., 5-10 seconds), inputting these parameters into the prediction trajectory model, the predicted movement trajectory of target obstacle 4 within the future preset time period (e.g., 5-10 seconds) can be obtained.
[0061] Furthermore, after obtaining the predicted movement trajectories of all target obstacles within a preset time period (e.g., 5-10 seconds), the system can use them as model input, combine them with the current state of the vehicle, and output the predicted movement trajectory of the vehicle within a preset time period (e.g., 5-10 seconds). Based on the predicted movement trajectory of the vehicle, the system can control the vehicle to complete decision-making actions such as acceleration, deceleration, avoidance, and lane changing to ensure the driving safety of the vehicle.
[0062] In some embodiments, obstacle shape features may refer to the geometric structural features of the obstacle or vehicle in three-dimensional space, including but not limited to: shape type, size parameters, etc. These features help to more accurately predict the trajectory, space occupied, and obstacle avoidance strategies of the obstacle or vehicle.
[0063] In some embodiments, static obstacles may include, but are not limited to: traffic facilities (such as traffic barriers, water-filled barriers, traffic cones, signposts, road signs, traffic light poles, etc.), road structures (such as guardrails, walls, bridge piers, curbs, roadside flower beds, etc.), temporary items (such as temporarily parked vehicles, construction materials, unattended express delivery boxes, street stalls, etc.), and natural obstacles (such as fallen tree branches, snowdrifts, rocks, large puddles, etc.). The system can jointly sense and identify the location of static obstacles using radar and cameras, and incorporate them into the environmental boundary constraints of the predictive model.
[0064] In some embodiments, when any dynamic obstacle changes, and the predicted movement trajectories of the target obstacle and the vehicle are re-determined, the first target obstacle is the obstacle that has changed. Specifically, when any dynamic obstacle undergoes a significant change in state (such as acceleration, deceleration, steering, or lane change) at the current moment, the system preferentially re-identifies this obstacle as the "first target obstacle," and re-obtains the predicted movement trajectory of the obstacle within a preset future time period, as well as the predicted movement trajectories of other target obstacles within a preset future time period, according to the method described in the above embodiments, thereby improving the response efficiency to sudden hazards and the adaptability of path prediction.
[0065] In some embodiments, the predicted trajectories of each target obstacle are presented as multiple parallel trajectory curves. This indicates that although the predicted trajectories of each target obstacle are generated separately, the dynamic constraints with other target obstacles (including the vehicle) are considered during the generation process, particularly to avoid intersections or conflicts between the predicted trajectories. This parallel trajectory prediction method helps improve prediction robustness and decision-making safety, providing more forward-looking support for subsequent vehicle path planning and dynamic obstacle avoidance.
[0066] The following is for reference. Figure 3 An autonomous driving control system according to an embodiment of the present invention is described.
[0067] Figure 3This is a block diagram of an automatic driving control system according to an embodiment of the present invention, such as... Figure 3 As shown, the autonomous driving control system 1 includes a perception system 11 and a controller 12.
[0068] In some embodiments, the sensing system 11 includes a lidar, a forward-facing camera, a forward-facing mid-range camera, a forward-facing close-range camera, a side-view camera and a rear-view camera, a forward-facing millimeter-wave radar, an angular millimeter-wave radar, a rear-facing millimeter-wave radar and a motion speed detector.
[0069] In some embodiments, lidar can provide high-precision point cloud data, which can be used to construct obstacle outlines, determine the spatial position and relative height of objects, and is particularly suitable for the identification and modeling of static and dynamic obstacles.
[0070] In some embodiments, the camera can detect appearance information of obstacles, such as color, shape, etc.
[0071] In some embodiments, millimeter-wave radar can provide the ability to detect dynamic obstacles at medium and long distances, and is particularly robust in visually limited environments such as nighttime, rain, and snow. It can be used to detect parameters such as relative speed, distance, and angle.
[0072] In some embodiments, the motion speed detector is used to acquire the vehicle's speed and acceleration in real time for motion trajectory prediction and vehicle motion state estimation.
[0073] In some embodiments, the autonomous driving control system 1 constructs a unified four-dimensional traffic system by fusing multi-source information from the perception system 11. That is, based on the four-dimensional spatial coordinates of (X, Y, Z, t), it uniformly models the information of the vehicle, surrounding obstacles (static and dynamic) and traffic facilities, which has spatiotemporal continuity, can represent the historical movement trajectory and future trend of each target object, and can accurately judge the collision relationship within the vehicle's travel range, enabling the machine to understand the physical world and changes like the human brain, thereby improving the system's decision-making efficiency and accuracy.
[0074] In some embodiments, the controller 12 is connected to the sensing system 11 to perform the vehicle control method described in the above embodiments.
[0075] In some embodiments, controller 12 may refer to an electronic control unit (ECU), which may be a stand-alone autonomous driving domain controller or integrated with the vehicle's central processing system (central computing platform).
[0076] According to an embodiment of the present invention, the autonomous driving control system 1, with controller 12 connected to perception system 11, employs the vehicle control method described in the previous embodiment. By constructing a unified four-dimensional traffic system, the vehicle, target obstacles, and traffic facilities are uniformly modeled in a four-dimensional spatial coordinate system, enabling the vehicle to possess an overall perception capability of environmental spatiotemporal changes. This provides an accurate and continuous spatiotemporal reference basis for subsequent behavior analysis and trajectory prediction. In this four-dimensional traffic system, the predicted movement trajectory of each target obstacle is not only generated based on its own historical behavior data but also comprehensively considers the influence of the behavior changes of other obstacles on its motion decision, realizing the modeling and coupled prediction of dynamic behavior interaction relationships. This allows the trajectory prediction model to accurately generate the predicted movement trajectories of multiple target obstacles under the mutual influence of multiple target obstacles within a preset future time period. Furthermore, by using the predicted movement trajectories of these obstacles to deduce the predicted movement trajectory of the vehicle within a preset time period in the future, the vehicle's decision path is not only based on its own state and static environment, but also dynamically responds to the changes in the interaction between surrounding obstacles. This improves the accuracy and rationality of the predicted movement trajectories of various target obstacles and the vehicle in complex traffic environments, making it more in line with the actual dynamic interaction behavior in complex traffic environments, and ultimately enhancing the safety and foresight of the vehicle's control decisions.
[0077] The following is for reference. Figure 4 A vehicle according to an embodiment of the present invention is described.
[0078] Figure 4 This is a block diagram of a vehicle according to an embodiment of the present invention, such as... Figure 4 As shown, the vehicle 100 includes the automatic driving control system 1 described in the above embodiment.
[0079] In some embodiments, vehicle 100 may be an intelligent connected vehicle such as a passenger car, commercial vehicle, logistics delivery vehicle, sweeper, or park shuttle vehicle equipped with autonomous driving capabilities, which can achieve autonomous driving functions from L2 to L4 to a certain extent.
[0080] In some embodiments, the vehicle 100 is used to implement the vehicle control method described in the above embodiments.
[0081] According to an embodiment of the present invention, the vehicle 100 employs the vehicle control method described in the above embodiment, or the autonomous driving control system 1 described in the above embodiment. By constructing a unified four-dimensional traffic system, the vehicle, target obstacles, and traffic facilities are uniformly modeled in a four-dimensional spatial coordinate system, enabling the vehicle 100 to have a holistic perception capability of environmental spatiotemporal changes. This provides an accurate and continuous spatiotemporal reference basis for subsequent behavior analysis and trajectory prediction. In this four-dimensional traffic system, the predicted movement trajectory of each target obstacle is not only generated based on its own historical behavior data, but also comprehensively considers the influence of the behavior changes of other obstacles on its motion decision, realizing the modeling and coupled prediction of dynamic behavior interaction relationships. This allows the trajectory prediction model to accurately generate the predicted movement trajectories of multiple target obstacles under the mutual influence of multiple target obstacles within a preset future time period. Furthermore, by using the predicted movement trajectories of these obstacles to deduce the predicted movement trajectory of the vehicle within a preset time period in the future, the vehicle's decision path is not only based on its own state and static environment, but also dynamically responds to the changes in the interaction between surrounding obstacles. This improves the accuracy and rationality of the predicted movement trajectories of various target obstacles and the vehicle in complex traffic environments, making it more in line with the actual dynamic interaction behavior in complex traffic environments, and ultimately enhancing the safety and foresight of the vehicle's control decisions.
[0082] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example.
[0083] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.
Claims
1. A vehicle control method, characterized in that, include: The system obtains the predicted movement trajectories of various target obstacles within a preset future timeframe in a four-dimensional transportation system. This four-dimensional transportation system is constructed based on obstacle perception information and traffic facilities from the vehicle's autonomous driving perception system in a four-dimensional spatial coordinate system. Various obstacles and the vehicle are respectively presented within the four-dimensional transportation system. The preset future timeframe is a set duration after the current moment. The predicted movement trajectories of each target obstacle within the preset future timeframe are generated based on the dynamic behavioral interaction relationships between the various obstacles. These dynamic behavioral interaction relationships include the influence of the behavior change of one obstacle on the behavioral decisions of other obstacles. The predicted movement trajectory of the vehicle is determined based on the predicted movement trajectory of each of the target obstacles within a preset time period in the future, and the driving of the vehicle is controlled based on the predicted movement trajectory of the vehicle. The vehicle control method further includes determining the predicted movement trajectory of each of the target obstacles based on the degree of danger posed by each target obstacle to the vehicle, wherein... The predicted movement trajectory of the first target obstacle is determined based on the historical movement trajectory of the first target obstacle, the position of the static obstacle, and the historical movement trajectory of the vehicle. The historical movement trajectory of the first target obstacle is the movement trajectory within a preset historical time period, and the preset historical time period is the set time period before the current moment. The predicted movement trajectory of the second target obstacle is determined based on the already determined predicted movement trajectory of the target obstacle, the historical movement trajectory of the second target obstacle, the position of the static obstacle, and the historical movement trajectory of the vehicle; The historical movement trajectory is determined based on the target obstacle or the vehicle's speed, obstacle shape characteristics, and movement direction within the preset historical time period.
2. The vehicle control method according to claim 1, characterized in that, The four-dimensional spatial coordinate system is constructed with the vehicle as the origin, the lane line direction as the x-axis, the direction perpendicular to the lane line direction as the y-axis, the vehicle's height direction as the z-axis, and the time axis.
3. The vehicle control method according to claim 1, characterized in that, The vehicle control method further includes: Based on the motion state information of obstacles in the four-dimensional transportation system, the degree of danger posed by each dynamic obstacle to the vehicle is determined; The target obstacle is determined based on the degree of danger.
4. The vehicle control method according to claim 3, characterized in that, The target obstacle is a dynamic obstacle whose danger level is higher than a preset danger level threshold.
5. The vehicle control method according to claim 1, characterized in that, The first target obstacle is the one among the multiple target obstacles that poses the highest level of danger to the vehicle; or, when any dynamic obstacle changes and the predicted movement trajectory of the target obstacle and the vehicle is re-determined, the first target obstacle is the obstacle that has changed. The second target obstacle is any target obstacle other than the first target obstacle among the plurality of target obstacles.
6. The control method according to claim 1, characterized in that, The predicted movement trajectories of each of the target obstacles are presented as multiple parallel movement trajectory curves.
7. An automatic driving control system, characterized in that, include: The perception system includes a lidar, a forward-facing camera, a forward-facing mid-range camera, a forward-facing close-range camera, a side-view camera and a rear-view camera, a forward-facing millimeter-wave radar, an angular millimeter-wave radar, a rear-facing millimeter-wave radar and a motion speed detector. A controller, connected to the sensing system, for executing the vehicle control method according to any one of claims 1-6.
8. A vehicle, characterized in that, The vehicle is used to implement the vehicle control method according to any one of claims 1-6, or the vehicle includes the automatic driving control system according to claim 7.
Citation Information
Patent Citations
Vehicle track prediction modeling method and device based on dynamic space-time interaction diagram
CN116595871A
Automatic driving vehicle trajectory planning method
CN119283896A
Commercial vehicle auxiliary driving method and device based on environment perception
CN120517406A
