An autonomous vehicle pose transformation control system and method based on terrain perception

Abstract

The application discloses a kind of based on terrain perception autonomous vehicle posture transformation control system and method, control system includes environment perception system, terrain estimation system, vehicle posture planning system, vehicle suspension control system and line control suspension system;Environment perception system is used to obtain the point cloud of road surface and front obstacle, image data;Terrain estimation system processes point cloud, image data, calculates current terrain state;Vehicle posture planning system receives current terrain state, according to current vehicle attitude, generates corresponding action sequence based on terrain state and vehicle posture mode mapping relationship;Vehicle suspension control system receives action sequence and solves, obtains the control quantity required for adjusting each suspension stroke, stiffness change;Line control suspension system is analyzed according to control quantity, and relevant signal is forwarded to corresponding actuator, the action execution of different vehicle posture target is realized, and vehicle posture autonomous transformation is completed.The application can autonomously adjust vehicle posture, adapt to environmental change and overcome road obstacle.

Description

Technical Field

[0001] This invention relates to the field of unmanned vehicle perception and control technology, specifically to an autonomous vehicle posture transformation control system and method based on terrain perception. Background Technology

[0002] With the promotion and application of unmanned and intelligent technologies, functional drones and unmanned vehicles have entered people's daily lives, providing strong support for the delivery of supplies. In special mission scenarios in the field, unmanned vehicles can efficiently and quickly replace personnel, reducing the loss of manpower and resources. However, whether in urban delivery or field missions, the varied and complex terrain and environment affect the driving range of unmanned vehicles, limiting their functionality and preventing them from fully realizing their potential to liberate manpower and execute tasks efficiently. Therefore, unmanned vehicles with high environmental adaptability and autonomous attitude change capabilities will have a wide range of applications.

[0003] Patent CN110345117B discloses "a modular integrated hydraulic device," which integrates the hydraulic control system and braking system of an oil-pneumatic suspension, realizing the online control system integration and design on a small unmanned platform. Patent CN110962522A discloses "design of an oil-pneumatic spring locking function for unmanned wheeled vehicles crossing trenches," realizing the design of basic hydraulic components for unmanned vehicles crossing trenches. It is evident that the integration design capability of components has been achieved in this field, but the design related to autonomous terrain judgment and intelligent control is still insufficient, resulting in cumbersome system operation procedures, a sharp increase in the operational pressure on operators, and an inability to effectively implement the system in real-world application scenarios. Summary of the Invention

[0004] In view of this, the present invention provides an autonomous vehicle posture change control system and method based on terrain perception, which can autonomously adjust vehicle posture, adapt to environmental changes and overcome road obstacles.

[0005] The technical solution adopted in this invention is as follows:

[0006] An autonomous vehicle posture transformation control system based on terrain perception includes an environmental perception system, a terrain estimation system, a vehicle posture planning system, a vehicle suspension control system, and a steerable suspension system.

[0007] The environmental perception system is used to obtain point cloud and image data of the road surface and obstacles ahead;

[0008] The terrain estimation system processes the point cloud and image data based on clustering and recognition algorithms, extracts road surface features, and calculates the current terrain state.

[0009] The vehicle posture planning system receives the current terrain state, generates a corresponding action sequence based on the current vehicle posture and the mapping relationship between terrain state and vehicle posture mode, and controls the vehicle's center of gravity to be within its stable area.

[0010] The vehicle suspension control system receives the action sequence and calculates it to obtain the control quantities required to adjust the travel and stiffness changes of each suspension component.

[0011] After parsing the control quantity, the steerable suspension system forwards the relevant signals to the corresponding actuators to realize the action execution of different vehicle posture targets and complete the autonomous transformation of vehicle posture.

[0012] Furthermore, the terrain estimation system employs a scene modeling algorithm to fuse the 3D digital high-rise map of the scene obtained by laser information and the road boundary information provided by vision into a scene model, forming a probability model of passable areas, effectively determining the terrain category, and providing input conditions for autonomous vehicle posture transformation.

[0013] Furthermore, the vehicle posture planning system combines vehicle status and terrain information to make intelligent and autonomous judgments. Under normal conditions, it plans vehicle posture behavior actions and generates corresponding action sequences. When abnormal or dangerous scenarios occur, it will pause vehicle posture action planning and prompt the operator.

[0014] Furthermore, the vehicle suspension control system includes a suspension adjustment controller and a suspension adjustment actuator; the suspension adjustment controller is used to receive the control quantity, calculate the execution command of each suspension adjustment actuator, and transmit it to the corresponding suspension adjustment actuator via a bus; the suspension adjustment actuator is used to adjust the suspension travel and stiffness changes.

[0015] Furthermore, the drive-by-wire suspension system includes a vehicle posture controller, a power unit assembly, an oil supply valve block, a wheel locking integrated valve block, an oil filling and releasing integrated valve block, an information bus, hydraulic lines, and a gas spring;

[0016] The oil supply valve block, the wheel locking integrated valve block, and the oil filling and discharging integrated valve block form a valve group; the power unit assembly is connected to each valve block and the oil spring in the valve group through hydraulic pipelines; the vehicle posture controller is connected to the power unit assembly and the valve group through an information bus.

[0017] The vehicle posture controller is a software control element of the hydraulic system, used for information reception, acquisition and analysis;

[0018] The power unit assembly serves as the power oil source for the entire steer-by-wire suspension system.

[0019] The oil supply valve block is used to control the pressure, flow and direction of the hydraulic system of the entire steerable suspension system, so as to realize the oil supply distribution control of the vehicle posture adjustment action;

[0020] The wheel-lifting locking integrated valve block is used to realize the locking and wheel-lifting functions of the vehicle suspension;

[0021] The integrated oil filling and discharging valve block is used to control the oil supply for vehicle attitude adjustment.

[0022] Furthermore, the action sequence includes:

[0023] The action sequence for a trench terrain scenario is: lower the vehicle's center of gravity to its lowest point → adjust the suspension system stiffness to rigid → keep the tires suspended.

[0024] The action sequence for sloping terrain scenarios is: lower the vehicle's center of gravity to its lowest point → maintain the suspension system's stiffness at a flexible level;

[0025] The action sequence corresponding to the scenario of facing a vertical wall and severe off-road terrain is as follows: the vehicle's center of gravity descends to the lowest point → the spring locks → the vehicle's center of gravity rises to the highest point → the spring unlocks → the suspension system stiffness becomes flexible.

[0026] An autonomous vehicle attitude change control method based on terrain perception, employing the aforementioned control system, comprises the following steps:

[0027] Step 1: Obtain point cloud and image data of the road surface and obstacles ahead;

[0028] Step 2: Process and calculate the point cloud and image data based on clustering and recognition algorithms, extract road surface features, and calculate the current terrain status;

[0029] Step 3: Based on the mapping relationship between terrain state and vehicle posture pattern obtained by the learning algorithm, and combined with the current terrain state and current vehicle posture, generate the corresponding action sequence to control the vehicle's center of gravity within its stable region.

[0030] Step 4: Based on the action sequence and calculation, obtain the control quantities required to adjust the travel and stiffness changes of each suspension component;

[0031] Step 5: After parsing the information based on the control quantity, the relevant signals are forwarded to the corresponding actuators to realize the action execution of different vehicle posture targets and complete the autonomous transformation of vehicle posture.

[0032] Furthermore, in step three, the method for generating the action sequence is as follows:

[0033] Step 301: Based on experimental testing, establish an expert experience base for motion mode switching, including the mapping relationship between terrain conditions and vehicle posture movements;

[0034] Step 302: The terrain state based on natural language description is fuzzified and transformed into the universe of discourse of the input;

[0035] Step 303: Based on the expert experience base data, for the fuzzy quantity of the input terrain state, perform action sequence reasoning and permutation combination on the terrain state-vehicle posture action mapping relationship, and solve for the optimal solution based on the objective function relationship as the action sequence result of the fuzzy description.

[0036] Step 304: Defuzzify the action sequence results of the fuzzy description and output them to the vehicle suspension control system.

[0037] Furthermore, in step four, the control quantity is as follows:

[0038] In a trench terrain scenario, the suspension control system issues a "wheel lift" control command. Upon receiving this command, the vehicle attitude controller opens the corresponding oil release valve and oil filling valve, and the powertrain in the steerable suspension system supplies oil. The air springs gradually contract, and the vehicle body descends to its lowest point. After execution, the wheel lift locking control action is automatically performed. The vehicle attitude controller controls the wheel lift locking integrated valve block, adjusting the stiffness of the air springs to rigidity, locking the vertical movement of the tires, and keeping the tires suspended.

[0039] When facing sloping terrain, the vehicle suspension control system issues a "lower to the lowest" control command. After receiving the command, the vehicle posture controller sends the control command to the corresponding oil release valve and oil filling valve through the information bus. Each valve opens according to the command, and the air spring contracts under the action of gravity, and the entire vehicle body lowers to the lowest point.

[0040] For scenarios involving vertical walls and severe off-road terrain, the suspension control system will issue a "lower to minimum" control command. Upon receiving this command, the vehicle attitude controller will open the corresponding oil release valve and oil filling valve, causing the air springs to contract under gravity, and the entire vehicle body will lower to its lowest point. Subsequently, a "spring lock" control command will be issued, and the vehicle attitude controller will control the wheel locking integrated valve block to put the air springs in a "locked" state. Then, the vehicle suspension control system will issue a "raise to maximum" control command, causing the entire vehicle body to rise to its highest point. After completing the vehicle attitude rise, the vehicle attitude controller will issue an "unlock" command. At this time, the wheel locking integrated valve block will open, and the vehicle body will fall under gravity. The spring stiffness will become flexible, and the load distribution of the previous adjustment process will ensure that the vertical load of each tire is evenly distributed.

[0041] Beneficial effects:

[0042] 1. This invention leverages the advantages of unmanned vehicle environmental perception and autonomous control systems. By integrating environmental data and vehicle status information, it proposes an autonomous vehicle posture adjustment control system. This system enables unmanned vehicles to intelligently and autonomously adjust their posture based on terrain changes, adapting to environmental changes and overcoming road obstacles. It simplifies the operation process, improves the vehicle's passability, saves manpower and time costs, and thus improves task efficiency. It can easily cope with sudden and harsh scenarios, reduce human operation and monitoring, and expand the application scope and value of unmanned vehicles.

[0043] 2. To achieve autonomous and intelligent vehicle attitude control, real-time and accurate judgment of the unmanned vehicle's environment and ground conditions is required. Terrain input is used to adjust the vehicle's attitude, overcoming road obstacles and improving its passability. Therefore, this invention employs a terrain estimation system that fuses multi-source sensor data. In complex field environments, this system can efficiently and quickly extract terrain features and output accurate terrain results. LiDAR is used to obtain radar point clouds, and visual cameras are used to obtain image data. The terrain estimation system utilizes data provided by the environmental perception system for computational processing, accurately identifying and classifying the terrain to support subsequent vehicle attitude changes.

[0044] 3. This invention utilizes the characteristics of unmanned platform drive-by-wire systems—fast action response and multiple degrees of control—to solve typical problems such as high difficulty of human operation, cumbersome action procedures, and low task execution efficiency of unmanned platforms in complex environments by adopting an integrated process of "perception-recognition-decision-planning-control" for terrain environment. Through the mapping relationship between different surface environments and control procedures, a vehicle posture autonomous transformation control method coupled with "terrain-control" is established.

[0045] 4. The vehicle change control system of this invention combines data feedback such as vehicle status and perceived environment, and adopts multi-source information fusion and learning algorithms to judge and predict the current state of the vehicle. It can effectively remind users of dangerous scenarios and abnormal states, and improve the safety of action planning. Attached Figure Description

[0046] Figure 1 This is a system composition diagram of the present invention;

[0047] Figure 2 This is a signal flow diagram of the terrain estimation system of the present invention;

[0048] Figure 3 This is a flowchart of the autonomous vehicle posture and motion planning signal of the present invention;

[0049] Figure 4 A diagram showing the mapping relationship and signal flow between typical terrain scenarios and the drive-by-wire suspension system;

[0050] Figure 5This is a diagram showing the components of the drive-by-wire suspension system of the present invention.

[0051] Among them, 1-Environmental perception system, 101-LiDAR, 102-Vision camera sensor, 2-Terrain estimation system, 3-Vehicle posture planning system, 4-Vehicle suspension control system, 401-Suspension adjustment controller, 402-Suspension adjustment actuator, 5-Drive-by-wire suspension system, 501-Vehicle posture controller, 502-Power unit assembly, 503-a-Fuel supply valve block, 503-b-Wheel lift lock integrated valve block, 503-c-Fuel charging and discharging integrated valve block, 504-Gas spring, 505-Tire, 506-Information bus, 507-Hydraulic pipeline. Detailed Implementation

[0052] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0053] This invention provides an autonomous vehicle attitude change control system based on terrain perception, such as... Figure 1 As shown, it includes an environmental perception system 1, a terrain estimation system 2, a vehicle posture planning system 3, a vehicle suspension control system 4, and a steerable suspension system 5.

[0054] The environmental perception system 1 is used to obtain point cloud and image data of the road surface and obstacles ahead for subsequent calculation and estimation. The environmental perception system 1 includes a lidar 101 and a vision camera. The lidar 101 is used to obtain radar point cloud data, and the vision camera is used to obtain image data.

[0055] Terrain estimation system 2 uses clustering and recognition algorithms to process and calculate point cloud and image data obtained from the fusion of relevant sensors in environmental perception system 1, extracting road features and calculating the current terrain state. Terrain estimation system 2 employs a scene modeling algorithm to fuse a 3D digital high-rise map (DEM) of the scene obtained from laser information with road boundary information provided by vision, forming a probabilistic model of passable areas.

[0056] The vehicle posture planning system 3 receives the current terrain state from the terrain estimation system 2. Based on the current vehicle posture, it generates corresponding action sequences according to the mapping relationship between terrain state and vehicle posture pattern obtained by the learning algorithm, controlling the vehicle's center of gravity within its stable area. The vehicle posture planning system 3 combines vehicle state and terrain information to perform intelligent autonomous judgment. Under normal conditions, it can plan vehicle posture behavior actions and generate corresponding action sequences; when abnormal or dangerous scenarios occur, it will pause posture action planning and alert the operator. This not only completes action planning but also improves the system's safety and stability.

[0057] The vehicle suspension control system 4 receives and calculates the action sequence generated by the vehicle posture planning system 3 to obtain the control quantity required to adjust the travel and stiffness changes of each suspension.

[0058] The steerable suspension system 5 analyzes the control input obtained from the vehicle suspension control system 4 and forwards the relevant signals to the corresponding actuators to achieve the execution of actions for different vehicle posture targets, thus completing autonomous vehicle posture transformation. The vehicle suspension control system 4 includes a suspension adjustment controller 401 and a suspension adjustment actuator 402. The suspension adjustment controller 401 receives the control input, calculates the execution command of each suspension adjustment actuator 402, and transmits it to the corresponding suspension adjustment actuator 402 via a bus. The suspension adjustment actuator 402 is used to adjust the suspension travel and stiffness changes.

[0059] Among them, the environmental perception system 1, terrain estimation system 2, vehicle posture planning system, and vehicle suspension control system 4 are deployed in the perception and control computer inside the unmanned vehicle. They perform calculations and processing on relevant data and transmit and receive vehicle posture change information. The vehicle posture status and judgment output interface can be connected to the remote monitoring interface for visualization, which can further improve the safety of the unmanned autonomous control system.

[0060] The drive-by-wire suspension system 5 includes a vehicle posture controller 501, a power unit assembly 502, an oil supply valve block 503-a, a wheel locking integrated valve block 503-b, an oil filling and draining integrated valve block 503-c, an information bus 506, hydraulic lines 507, and a gas spring 504.

[0061] The oil supply valve block 503-a, the wheel locking integrated valve block 503-b, and the charging / discharging integrated valve block 503-c form a valve group. The power unit assembly 502 is connected to the various functional integrated valve blocks and the hydraulic springs 504 within the valve group via hydraulic lines 507. Different valve block switching combinations can be formed by controlling the valve group 503, realizing the communication and adjustment of the pipeline, thereby further realizing the real-time control of the hydraulic springs 504. The vehicle posture controller 501 is connected to the power unit assembly 502 and the valve group via an information bus 506 for transmitting control commands and collecting component status information. Based on the parsing of vehicle posture change commands, it performs real-time control of oil supply and valve block closure and opening, and feeds back the current status of each sub-component for overall posture judgment.

[0062] The vehicle posture controller 501 is a software control element for hydraulic systems, used for information reception, acquisition, and analysis.

[0063] The power unit assembly 502 is the power source for the entire steerable suspension system 5.

[0064] The oil supply valve block 503-a is used to control the pressure, flow and direction of the hydraulic pressure of the entire steerable suspension system 5, so as to realize the oil supply distribution control of the vehicle posture adjustment action.

[0065] The wheel-lifting locking integrated valve block 503-b is used to realize the locking and wheel-lifting functions of the vehicle suspension.

[0066] The 503-c integrated filling and discharging valve block is used to control the fuel supply for vehicle attitude adjustment.

[0067] After the vehicle posture controller 501 parses the information, it forwards the relevant signals to the power unit assembly 502, the oil supply valve block 503-a, the wheel lifting lock integrated valve block 503-b, and the oil filling and draining integrated valve block 503-c. The different actuators cooperate with each other to realize the action execution of different vehicle posture targets and complete the vehicle posture change function.

[0068] The steerable suspension system 5 adopts a highly integrated design, integrating the power source, controller, energy storage device, pipelines and valves of the required hydraulic system into one unit. It optimizes the external interface of the system and uses FlexRay bus information transmission with the vehicle suspension control system to improve information transmission efficiency, thereby greatly improving the reliability and stability of the system.

[0069] Based on the above control system, this invention provides an autonomous vehicle attitude change control method based on terrain perception, the steps of which are as follows:

[0070] Step 1: Obtain point cloud and image data of the road surface and obstacles ahead through the environmental perception system 1.

[0071] Step two involves processing and calculating the point cloud and image data obtained from environmental perception system 1 based on clustering and recognition algorithms, extracting road surface features, and calculating the current terrain state, such as... Figure 2 As shown.

[0072] Step S2-1: The terrain estimation system 2 acquires data from the lidar 101 and the visual camera sensor 102 in the environmental perception system 1 to form accumulated data.

[0073] Step S2-2: Data analysis and feature extraction are performed using the corresponding sensor feature extraction network 201. Specifically, the feature extraction of the LiDAR 101 point cloud data employs the following method: First, a transformation matrix is ​​generated by the T-net network based on the overall features of the point set. Then, features are extracted point-by-point, and the points are aligned in the feature space using this method. Finally, the aligned point-by-point feature set is used as the feature of the overall point cloud.

[0074] The following method is used to extract features from image data of the visual camera sensor 102: First, features are abstracted and globalized through convolutional neural networks and pooling operations. Then, features are further abstracted through deconvolution and unpooling. The resolution of the feature map is gradually restored to a resolution close to that of the input image. This process ensures that the low-resolution image contains global information and low-level detailed features, while the high-resolution image describes the features at each location.

[0075] Step S2-3: After obtaining the feature data, the feature data is fused. The laser point cloud is projected onto the image coordinate system and mapped onto the image feature map coordinate system. Based on the features of the pixel points in the feature map, the feature vector of the feature map at the projection position of the laser point is obtained by bilinear interpolation. This feature is the image feature of the laser point. At this time, this feature is connected with the laser point cloud feature to form the fused feature.

[0076] Step S2-4: After acquiring the above features, construct the terrain category model and scene. At this point, combine the chassis performance parameters and capability boundaries of the autonomous vehicle to classify the terrain description of the vehicle; integrate the features acquired above to construct a gridded DEM model, which includes not only the height and position information of each grid, but also the attribute information of each grid, and then construct the current driving scene to perform traversable area analysis and calculation.

[0077] This yields the current terrain status, which is then used as system input to carry out the vehicle posture planning system process 3.

[0078] Step 3: Combining the road condition information obtained in Step 2 with the current vehicle posture, the vehicle posture planning system 3 generates an action sequence based on the mapping relationship between the terrain state and the vehicle posture pattern obtained by the learning algorithm, and controls the vehicle's center of gravity within its stable area.

[0079] Specifically, such as Figure 3 As shown, in step S3-1, based on extensive experimental testing, an expert experience base for motion mode switching is established, including the mapping relationship between terrain conditions and vehicle posture. This base is continuously adjusted and improved through data accumulation and practice to enrich its data.

[0080] Step S3-2: Perform interface fuzzification. The terrain state described in natural language is fuzzified, and the input is transformed into the universe of discourse of the input to ensure that the algorithm understands the input, which is then used for fuzzy inference in step S3-3.

[0081] Step S3-3: Based on the expert experience base data, for the fuzzy quantity of the input terrain description, perform action sequence reasoning and permutation combination on the terrain state-vehicle posture action mapping relationship, and solve for the optimal solution as the action sequence result of the fuzzy description based on the objective function relationship of the task.

[0082] Step S3-4 involves deblurring the output results. Because variable fuzzing is performed in this algorithm, directly outputting the results would prevent subsequent algorithms and execution from fully understanding them. Therefore, deblurring is performed before the output results to obtain the vehicle posture action sequence, which is then output to the vehicle suspension control system 4.

[0083] Step four: The vehicle suspension control system 4 receives and calculates the action sequence, divides the action sequence control, generates a low-level control strategy, and determines the control quantities required to adjust the travel and stiffness changes of each suspension component; for example... Figure 4 As shown, the action sequence control flow for the main terrain features is as follows:

[0084] For trench terrain scenarios, the generated action sequence is as follows: the vehicle's center of gravity lowers to its lowest point → the suspension system stiffness is adjusted to rigid → the tires remain suspended. At this time, the vehicle suspension control system 4 will issue a "wheel lift" control command. After receiving this command, the vehicle posture controller 501 will open the corresponding oil release valve and oil filling valve, supplying oil to the powertrain in the drive-by-wire suspension system 5. The air spring 504 will gradually contract, and the entire vehicle body will lower to its lowest point. After execution, the wheel lift locking control action will be automatically performed. The vehicle posture controller 501 will control the wheel lift locking integrated valve block 503-b, adjusting the stiffness of the air spring 504 to rigid, locking the vertical movement of the tire 505, keeping the tire 505 suspended, and preventing the tire 505 from getting stuck during trench crossing. Through the above control process, the vehicle's ability to cope with trench environments is improved.

[0085] For sloping terrain scenarios, the generated action sequence is as follows: the vehicle's center of gravity lowers to its lowest point → the suspension system maintains a flexible stiffness. At this time, the vehicle suspension control system 4 will issue a "lower to lowest" control command. After receiving this command, the vehicle posture controller 501 will send the control command to the corresponding drain valve and fill valve via the FlexRay bus (information bus 506). Each valve will open according to the command, and the air spring 504 will gradually contract under the action of gravity, lowering the entire vehicle body to its lowest point. Through the above control process, the vehicle posture maintains a flexible state and lowers the vehicle's center of gravity, ensuring driving stability on slopes and improving the vehicle's ability to cope with climbing environments.

[0086] For scenarios involving vertical walls and severe off-road terrain, the generated action sequence is as follows: vehicle center of gravity descends to its lowest point → spring locks → vehicle center of gravity rises to its highest point → spring unlocks → suspension system stiffness becomes flexible. At this time, the suspension control system will issue a "descend to lowest" control command. After receiving this command, the vehicle posture controller 501 opens the corresponding oil release valve and oil filling valve, and the gas spring 504 gradually contracts under the action of gravity, causing the entire vehicle body to descend to its lowest point. Subsequently, a "spring locks" control command is issued, and the vehicle posture controller 501 controls the lifting wheel locking integrated valve block 503-b, so that the gas spring 504 is in a "locked" state. Then, the vehicle suspension control system 4 issues a "rise to highest" control command, and the entire vehicle body rises to reach its highest center of gravity. This "descend-rise" process mainly serves to measure the load through feedback from the hydraulic circuit, ensuring the load distribution of each gas spring 504 in the vehicle. Finally, after the vehicle's upward movement is completed, the vehicle attitude controller 501 issues an "unlock" command. At this point, the wheel locking integrated valve block 503-b opens, and the vehicle body falls under gravity. The spring stiffness becomes flexible, and the load distribution is adjusted according to the previous adjustment process to ensure that the vertical load of each tire 505 is evenly distributed. Through the above control process, the vehicle attitude is maintained in a flexible state, and the load of each tire 505 is evenly distributed, improving the vehicle's off-road driving ability and passability.

[0087] Step 5: After parsing the control quantity, the vehicle posture controller 501 forwards the relevant signals to the corresponding actuators to realize the action execution of different vehicle posture targets and complete the autonomous transformation of vehicle posture.

[0088] like Figure 5 As shown, the vehicle posture controller 501 is connected to the power unit assembly 502 and the valve group 503 via an information bus 506 for control command transmission and component status information acquisition. Based on the parsing of the vehicle posture change commands, it performs real-time control of oil supply and valve block closure and opening, and provides feedback on the current status of each sub-component for overall posture judgment. The power unit assembly 502 is connected to the functional integrated valve blocks and the hydraulic springs 504 within the valve group 503 via hydraulic lines 507. Different valve block switching combinations can be formed by controlling the valve group 503, realizing the connection and adjustment of the pipeline, thereby further realizing real-time control of the hydraulic springs 504. The lower end of the hydraulic springs 504 is rigidly connected to the vehicle suspension and tires 505. The extension and retraction of the hydraulic springs 504 drives the vertical displacement of the suspension and tires 505, realizing the vehicle posture adjustment action.

[0089] In summary, the above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An autonomous vehicle posture change control system based on terrain perception, characterized in that, This includes environmental perception systems, terrain estimation systems, vehicle posture planning systems, vehicle suspension control systems, and steerable suspension systems. The environmental perception system is used to obtain point cloud and image data of the road surface and obstacles ahead; The terrain estimation system processes the point cloud and image data based on clustering and recognition algorithms, extracts road surface features, and calculates the current terrain state. The vehicle posture planning system receives the current terrain state, generates a corresponding action sequence based on the current vehicle posture and the mapping relationship between terrain state and vehicle posture mode, and controls the vehicle's center of gravity to be within its stable area. The vehicle suspension control system receives the action sequence and calculates it to obtain the control quantities required to adjust the travel and stiffness changes of each suspension component. After parsing the control quantity, the steerable suspension system forwards the relevant signals to the corresponding actuators to realize the action execution of different vehicle posture targets and complete the autonomous transformation of vehicle posture. The drive-by-wire suspension system includes a vehicle posture controller, a power unit assembly, an oil supply valve block, a wheel lifting lock integrated valve block, an oil filling and releasing integrated valve block, an information bus, hydraulic lines, and a gas spring. The oil supply valve block, the wheel locking integrated valve block, and the oil filling and discharging integrated valve block form a valve group; the power unit assembly is connected to each valve block and the oil spring in the valve group through hydraulic pipelines; the vehicle posture controller is connected to the power unit assembly and the valve group through an information bus. The vehicle posture controller is a software control element of the hydraulic system, used for information reception, acquisition and analysis; The power unit assembly serves as the power oil source for the entire steer-by-wire suspension system. The oil supply valve block is used to control the pressure, flow and direction of the hydraulic system of the entire steerable suspension system, so as to realize the oil supply distribution control of the vehicle posture adjustment action; The wheel-lifting locking integrated valve block is used to realize the locking and wheel-lifting functions of the vehicle suspension; The integrated oil filling and discharging valve block is used to control the oil supply for vehicle attitude adjustment; The action sequence includes: The action sequence for a trench terrain scenario is: lower the vehicle's center of gravity to its lowest point → adjust the suspension system stiffness to rigid → keep the tires suspended. The action sequence for sloping terrain scenarios is: lower the vehicle's center of gravity to its lowest point → maintain the suspension system's stiffness at a flexible level; The action sequence corresponding to the scenario of facing a vertical wall and severe off-road terrain is as follows: the vehicle's center of gravity descends to the lowest point → the spring locks → the vehicle's center of gravity rises to the highest point → the spring unlocks → the suspension system stiffness becomes flexible.

2. The terrain-aware autonomous vehicle posture change control system as described in claim 1, characterized in that, The terrain estimation system employs a scene modeling algorithm, which integrates a 3D digital high-rise map of the scene obtained from laser information with road boundary information provided by vision to form a scene model, thereby creating a probability model of passable areas. This effectively determines the terrain category and provides input conditions for autonomous vehicle posture transformation.

3. The terrain-aware autonomous vehicle posture change control system as described in claim 1, characterized in that, The vehicle posture planning system combines vehicle status and terrain information to make intelligent and autonomous judgments. Under normal conditions, it plans vehicle posture behavior actions and generates corresponding action sequences. When abnormal or dangerous scenarios occur, it will pause vehicle posture action planning and prompt the operator.

4. The terrain-aware autonomous vehicle posture change control system as described in claim 1, characterized in that, The vehicle suspension control system includes a suspension adjustment controller and a suspension adjustment actuator; the suspension adjustment controller is used to receive the control quantity, calculate the execution command of each suspension adjustment actuator, and transmit it to the corresponding suspension adjustment actuator via a bus; The suspension adjustment actuator is used to adjust the suspension travel and stiffness changes.

5. A method for autonomous vehicle posture transformation control based on terrain perception, characterized in that, The method using the control system as described in any one of claims 1-4 comprises the following steps: Step 1: Obtain point cloud and image data of the road surface and obstacles ahead; Step 2: Process and calculate the point cloud and image data based on clustering and recognition algorithms, extract road surface features, and calculate the current terrain status; Step 3: Based on the mapping relationship between terrain state and vehicle posture pattern obtained by the learning algorithm, and combined with the current terrain state and current vehicle posture, generate the corresponding action sequence to control the vehicle's center of gravity within its stable region. Step 4: Based on the action sequence and calculation, obtain the control quantities required to adjust the travel and stiffness changes of each suspension component; Step 5: After parsing the information based on the control quantity, the relevant signals are forwarded to the corresponding actuators to realize the action execution of different vehicle posture targets and complete the autonomous transformation of vehicle posture.

6. The autonomous vehicle posture transformation control method based on terrain perception as described in claim 5, characterized in that, In step three, the method for generating the action sequence is as follows: Step 301: Based on experimental testing, establish an expert experience base for motion mode switching, including the mapping relationship between terrain conditions and vehicle posture movements; Step 302: The terrain state based on natural language description is fuzzified and transformed into the universe of discourse of the input; Step 303: Based on the expert experience base data, for the fuzzy quantity of the input terrain state, perform action sequence reasoning and permutation combination on the terrain state-vehicle posture action mapping relationship, and solve for the optimal solution based on the objective function relationship as the action sequence result of the fuzzy description. Step 304: Defuzzify the action sequence results of the fuzzy description and output them to the vehicle suspension control system.

7. The autonomous vehicle posture change control method based on terrain perception as described in claim 5, characterized in that, In step four, the control quantities are as follows: In a trench terrain scenario, the suspension control system issues a "wheel lift" control command. Upon receiving this command, the vehicle attitude controller opens the corresponding oil release valve and oil filling valve, and the powertrain in the steerable suspension system supplies oil. The air springs gradually contract, and the vehicle body descends to its lowest point. After execution, the wheel lift locking control action is automatically performed. The vehicle attitude controller controls the wheel lift locking integrated valve block, adjusting the stiffness of the air springs to rigidity, locking the vertical movement of the tires, and keeping the tires suspended. When facing sloping terrain, the vehicle suspension control system issues a "lower to the lowest" control command. After receiving the command, the vehicle posture controller sends the control command to the corresponding oil release valve and oil filling valve through the information bus. Each valve opens according to the command, and the air spring contracts under the action of gravity, and the entire vehicle body lowers to the lowest point. For scenarios involving vertical walls and severe off-road terrain, the suspension control system will issue a "lower to minimum" control command. Upon receiving this command, the vehicle attitude controller will open the corresponding oil release valve and oil filling valve, causing the air springs to contract under gravity, and the entire vehicle body will lower to its lowest point. Subsequently, a "spring lock" control command will be issued, and the vehicle attitude controller will control the wheel locking integrated valve block to put the air springs in a "locked" state. Then, the vehicle suspension control system will issue a "raise to maximum" control command, causing the entire vehicle body to rise to its highest point. After completing the vehicle attitude rise, the vehicle attitude controller will issue an "unlock" command. At this time, the wheel locking integrated valve block will open, and the vehicle body will fall under gravity. The spring stiffness will become flexible, and the load distribution of the previous adjustment process will ensure that the vertical load of each tire is evenly distributed.

Images