CN-122009180-A - Four-wheel drive robot control method and device apparatus, medium, and program product

CN122009180ACN 122009180 ACN122009180 ACN 122009180ACN-122009180-A

Abstract

The embodiment of the application provides a four-wheel drive robot control method apparatus, device, medium, and program product. The method comprises the steps of obtaining multi-sensor data collected in the running process of the four-wheel drive robot, predicting road condition information in a preset time interval in the future through a preset first model based on the multi-sensor data, generating power distribution data of four-wheel drive wheels through a preset second model based on the road condition information, wherein the power distribution data comprise one or more of a torque distribution proportion, a differential mechanism mode and a braking torque adjusting instruction, and controlling the four-wheel drive robot to run according to the power distribution data. The self-adaptive power distribution and high-stability steering of the four-wheel drive robot under the complex terrain are realized, and the adaptability, the reliability and the task completion efficiency of the four-wheel drive robot in the complex terrain are improved.

Inventors

TANG XIAOLIANG
YANG FANG
YANG SIFENG
LIU ZIWEI
HAN LEI
ZHOU YABING
TANG LI

Assignees

广东电网有限责任公司清远供电局

Dates

Publication Date: 20260512
Application Date: 20260312

Claims (12)

1. A four-wheel drive robot control method, characterized by comprising: acquiring multi-sensor data acquired in the running process of the four-wheel drive robot, wherein the multi-sensor data comprises one or more of running state data, environment sensing data and hardware feedback data; Predicting road condition information in a preset time interval in the future through a preset first model based on the multi-sensor data; generating power distribution data of four-wheel drive wheels through a preset second model based on the road condition information, wherein the power distribution data comprises one or more of a torque distribution proportion, a differential mechanism mode and a braking torque adjustment instruction; and controlling the four-wheel drive robot to run according to the power distribution data.
2. The method of claim 1, wherein the traffic information includes a road surface condition and at least one associated parameter of the road surface condition, the method further comprising: acquiring a first training set, wherein the first training set comprises a plurality of groups of training data pairs, the training data pairs comprise preset multi-sensor data and label information, and the label information comprises road condition information corresponding to the multi-sensor data; performing iterative training operation on a preset first model through the first training set until the preset convergence condition is met, so as to obtain a trained first model; the first model is formed by fusing a model predictive control algorithm and a reinforcement learning algorithm, and is coupled to a whole vehicle controller of the four-wheel drive robot.
3. The method according to claim 1, wherein the generating power distribution data of the four-wheel drive wheels by a preset second model based on the road condition information includes: Inputting the road condition information into the second model, wherein the second model is used for determining the power distribution data based on the road condition information through at least one performance constraint associated with the four-wheel drive robot and a preset road condition database, and the road condition database comprises power distribution data corresponding to at least one road condition information; or alternatively; the second model comprises a generating model obtained through pre-training, and the generating of the power distribution data of the four-wheel driving wheels through the preset second model based on the road condition information comprises the following steps: And inputting the road condition information into the second model, wherein the second model is used for generating power distribution data matched with the road condition information based on the road condition information.
4. The method of claim 1, wherein said controlling said four-wheel drive robot to travel in accordance with said power distribution data comprises: according to the torque distribution proportion, the transmission ratio of the central differential is adjusted, and the differential is switched to a preset differential mode through the wheel edge differential; and controlling the electromagnetic brake to apply differential braking torque to four wheels according to the braking torque adjusting instruction.
5. The method of claim 4, wherein the controlling the electromagnetic brake to apply differential braking torque to four wheels comprises: In the steering process, applying braking moment to the external wheels according to the detection result of the roll angle of the vehicle body so as to counteract the roll trend; In the gradient change process, according to the longitudinal pitch angle detection result, braking torque is applied to the front wheels or the rear wheels so as to maintain the balance of the vehicle body.
6. The method of claim 4, wherein the differential modes include a limited slip differential mode and an electronically controlled differential lock mode; The switching to a preset differential mode through the wheel side differential includes: in response to receiving a switching instruction of a limited slip differential mode, controlling the wheel-side differential to enter passive adjustment, and triggering the friction plate group to lock by the wheel rotation speed difference; And in response to receiving a switching instruction of the electric control differential lock mode, controlling the wheel speed sensor to acquire the rotating speed in real time, and controlling the electromagnetic clutch to completely lock the wheel edge differential mechanism through the whole vehicle controller.
7. The method of claim 4, wherein said adjusting the gear ratio of the central differential according to the torque split ratio comprises: And controlling the planetary gear mechanism and the multi-plate clutch assembly in the integrated transfer case according to the torque distribution proportion, and adjusting the transmission ratio of the central differential mechanism in real time to uniformly distribute the torque to the front/rear drive axle.
8. The method according to any one of claims 1-7, further comprising: acquiring operation data in the operation process of the four-wheel drive robot in real time; In response to determining that the four-wheel drive robot fails based on the operational data, the four-wheel drive robot is controlled to operate using a redundancy mechanism that matches the failure.
9. The method of claim 8, wherein said employing a redundancy mechanism matched to said fault to control operation of said four-wheel drive robot comprises: in response to determining that the hardware of the four-wheel drive robot fails based on the operational data, controlling the four-wheel drive robot to operate using at least one level of hardware redundancy mechanism; And in response to determining that the software of the four-wheel drive robot fails based on the operation data, controlling the four-wheel drive robot to operate by adopting a preset software redundancy mechanism.
10. The method of claim 9, wherein the employing at least one level of hardware redundancy mechanism to control the four-wheel drive robot operation comprises: In response to determining that the motor rotation speed associated with the four-wheel drive robot fails based on the operation data, controlling the four-wheel drive robot to operate by adopting a first hardware redundancy mechanism, wherein the first hardware redundancy mechanism is used for correcting the power distribution data through the second model when the motor rotation speed is within a preset first interval, and triggering a magnetic brake braking operation to reduce the motor rotation speed when the motor rotation speed is within a preset second interval, and the first interval is smaller than the second interval; In response to determining that the single-wheel motor and/or the differential of the four-wheel drive robot is/are faulty based on the operation data, controlling the four-wheel drive robot to operate by adopting a second hardware redundancy mechanism, wherein the second hardware redundancy mechanism is used for cutting off power output of the faulty wheel, distributing torque of the faulty wheel to the other three wheels through a central differential and an inter-axle differential lock, readjusting the torque proportion of a front drive axle and a rear drive axle, and controlling the second model to regenerate power distribution data after power is reconstructed; and in response to determining that the hydraulic brake of the four-wheel drive robot fails and/or the single wheel brake fails based on the operation data, controlling the four-wheel drive robot to operate by adopting a third hardware redundancy mechanism, wherein the third hardware redundancy mechanism is used for triggering the electromagnetic brake to break off the power and automatically brake when the hydraulic brake fails, and/or controlling the rest three wheels to increase the braking moment when the single wheel brake fails so as to realize four-wheel braking moment redistribution.
11. The method of claim 9, wherein the employing a preset software redundancy mechanism to control the four-wheel drive robot operation comprises: And correcting the power distribution data generated by the second model based on the operation data, improving the data feedback frequency of hardware equipment in the four-wheel drive robot, and controlling the four-wheel drive robot to operate according to preset running parameters.
12. The four-wheel drive robot is characterized by comprising a whole vehicle controller, a left drive axle assembly, a right drive axle assembly, a permanent magnet synchronous motor, a motor controller, a two-stage speed reduction system, an electromagnetic brake, a wheel encoder and a multi-sensor fusion module; the multi-sensor fusion module and the wheel encoder are used for acquiring multi-sensor data in the running process of the four-wheel drive robot; The whole vehicle controller is used for calling a preset first model and a preset second model based on the multi-sensor data to generate power distribution data; The motor controller is used for controlling the permanent magnet synchronous motor to output corresponding power based on the power distribution data; the secondary speed reduction system is used for reducing speed and increasing torque of the power output by the permanent magnet synchronous motor so as to match the torque requirement of the four-wheel drive robot and transmitting the torque to a drive axle; The left drive axle assembly/right drive axle assembly is used for transmitting the power distribution proportion in the power distribution data to the wheel sides and realizing the dynamic torque distribution among the wheels through the wheel side differential mechanism according to the differential strategy in the power distribution data; The electromagnetic brake is used for outputting a designated braking torque when the four-wheel drive robot is unstable in posture and assisting differential steering through differential distribution of braking force.

Description

Four-wheel drive robot control method and device apparatus, medium, and program product Technical Field The present application relates to the field of artificial intelligence and robot engineering, and in particular, to a four-wheel drive robot control method, apparatus, device, medium, and program product. Background With the gradual development of science and technology, robots are widely applied to operation scenes of complex unstructured terrains, for example, robots need to cope with soft soil, broken stone roads, ice and snow pavements, steep slopes, shoals and other changeable terrains, and meanwhile, the robots need to have high stability, strong traction force and quick response capability. The conventional two-wheel driving and crawler-type driving robots are easy to cause task interruption due to the problems of uneven power distribution, steering clamping stagnation, single-point faults and the like in complex terrains. In addition, when the robot executes tasks, the robot needs to sense the terrain change in real time, forecast the dynamic interference and quickly adjust the power distribution strategy, the existing control method is mostly based on a fixed strategy library, and the dynamic forecast and self-adaptive adjustment of complex terrains are difficult to realize, so that the power distribution strategy is lagged or is not matched with the actual requirements, and the problems of high energy consumption, insufficient stability and the like are aggravated. Disclosure of Invention The four-wheel drive robot control method, device, equipment, medium and program product provided by the embodiment of the application are used for achieving the effects of improving the adaptability, reliability and task completion efficiency of the four-wheel drive robot in complex terrains. In a first aspect, an embodiment of the present application provides a four-wheel drive robot control method, including: acquiring multi-sensor data acquired in the running process of the four-wheel drive robot, wherein the multi-sensor data comprises one or more of running state data, environment sensing data and hardware feedback data; Predicting road condition information in a preset time interval in the future through a preset first model based on the multi-sensor data; generating power distribution data of four-wheel drive wheels through a preset second model based on the road condition information, wherein the power distribution data comprises one or more of a torque distribution proportion, a differential mechanism mode and a braking torque adjustment instruction; and controlling the four-wheel drive robot to run according to the power distribution data. In a second aspect, an embodiment of the present application provides a four-wheel drive robot, including a vehicle controller, a left drive axle assembly, a right drive axle assembly, a permanent magnet synchronous motor, a motor controller, a secondary speed reduction system, an electromagnetic brake, a wheel encoder, and a multi-sensor fusion module; the multi-sensor fusion module and the wheel encoder are used for acquiring multi-sensor data in the running process of the four-wheel drive robot; The whole vehicle controller is used for calling a preset first model and a preset second model based on the multi-sensor data to generate power distribution data; The motor controller is used for controlling the permanent magnet synchronous motor to output corresponding power based on the power distribution data; the secondary speed reduction system is used for reducing speed and increasing torque of the power output by the permanent magnet synchronous motor so as to match the torque requirement of the four-wheel drive robot and transmitting the torque to a drive axle; the left drive axle assembly/right drive axle assembly is used for transmitting the power distribution proportion in the power distribution data to the wheel sides and realizing the dynamic torque distribution among the wheels through the wheel side differential mechanism according to the differential strategy in the power distribution data; The electromagnetic brake is used for outputting a designated braking torque when the four-wheel drive robot is unstable in posture and assisting differential steering through differential distribution of braking force. In a third aspect, an embodiment of the present application provides a four-wheel drive robot control device including: The acquisition module is used for acquiring multi-sensor data acquired in the running process of the four-wheel drive robot, wherein the multi-sensor data comprises one or more of running state data, environment sensing data and hardware feedback data; The prediction module is used for predicting road condition information in a preset time interval in the future through a preset first model based on the multi-sensor data; the generation module is used for generating power distribution data of four-wheel drive wheels through a preset