CN-115840368-B - Track tracking control method for tire slip estimation in orchard environment

CN115840368BCN 115840368 BCN115840368 BCN 115840368BCN-115840368-B

Abstract

The invention discloses a track tracking control method for tire slip estimation in an orchard environment, which is a method for an apple picking robot carrying four straight motors and four rotating motors to walk along an expected track given by path planning, and can carry out high-efficiency robust control on longitudinal slip and transverse sideslip existing in the orchard environment. According to the invention, the longitudinal sliding degree is described by introducing the sliding parameters, the sliding generated by the lateral direction is estimated by utilizing the sliding mode observer to obtain the actually required control quantity, and the control quantity is respectively transmitted to eight motors through the task allocation principle, so that the efficient track tracking of the picking robot when the longitudinal sliding and the sideslip exist simultaneously in the orchard environment is realized, and the track tracking stability and the track tracking precision of the system can be greatly improved. The method can realize expected track tracking by 100% when the device is normally operated.

Inventors

SUN YU
LI XINTONG
YAO WENGUANG
Li chenxing
CHEN XIANGLONG
Huang Jiangzhou
SUN YIMING

Assignees

南京理工大学

Dates

Publication Date: 20260512
Application Date: 20221222

Claims (9)

1. The track tracking control method for the tire slip estimation in the orchard environment is used for a double closed-loop control system and is characterized by comprising the following steps: Step 1, establishing a differential mobile robot kinematic model with a drive shaft not coincident with a geometric center; Step 2, establishing a track tracking error equation based on a differential mobile robot kinematic model and combining an expected track model; Step 3, designing an outer loop kinematic controller based on a Lyapunov direct method according to a track tracking error equation, and obtaining the control speed and the angular speed of the system under an ideal state through errors; Step 4, establishing a differential mobile robot dynamics model based on a Lagrangian method, designing an inner ring dynamics controller based on the dynamics model, and converting linear speed and angular speed in an ideal state into moment control on a left wheel and a right wheel; step 5, online estimating unknown items and uncertain parameters existing in the inner ring dynamics model through a neural network model, reducing system oscillation by utilizing sliding mode control, and continuously adjusting network weight through the difference between actual output and expected value; the step 5 specifically includes: Step 5.1, defining the input and output of the neural network, the number of neurons, the number of hidden neurons in the middle, a transformation function, and the center, width and initial weight parameters of the transformation function; Step 5.2, designing a sliding die surface through linear speed and angular speed output by the kinematic controller, and forcing the system to reach the sliding die surface and move along the sliding die surface; And 5.3, comparing the actual output value of the system with an expected value, and adjusting the weight of the neural network by using a gradient descent method.
2. The trajectory tracking control method according to claim 1, wherein the step 1 specifically includes: Step 1.1, determining a system state quantity by utilizing the characteristics of the differential mobile robot; and 1.2, deriving a kinematic model of the differential mobile robot when the driving shaft is not coincident with the geometric center under the global coordinate system by combining the existing kinematic model.
3. The trajectory tracking control method according to claim 2, wherein the step 2 specifically includes: Step 2.1, establishing an expected track state equation according to an expected track, wherein the equation form needs to have the same input and output form as a kinematic model equation; And 2.2, converting a kinematic model equation under a local coordinate system into an equation under a global coordinate system shown in the formula (1) by utilizing a coordinate conversion matrix, and deriving a track tracking error equation shown in the formula (2) by combining an expected track equation under the global coordinate system: (1) (2) Wherein, the In order to be in the pose, For the output of the outer loop controller, For the distance between the center of the drive shaft and the geometric center, For the position and orientation error, For the desired linear and angular speeds.
4. The trajectory tracking control method according to claim 3, wherein the step 3 specifically includes: step 3.1, obtaining a first derivative of a track tracking error equation to obtain a track tracking error model; step 3.2, constructing a Lyapunov function, satisfying all state error amounts existing in a system, and appearing in a square term form, so that the constructed Lyapunov function is positively determined; Step 3.3, solving a first derivative of the constructed Lyapunov function, and substituting a system track tracking error model to obtain a system stability judging condition; And 3.4, designing a system control input according to the Lyapunov stability judging condition so that the first derivative is negatively determined.
5. The trajectory tracking control method according to claim 4, wherein the Lyapunov function is: Wherein, the As a function of the horizontal axis error, The longitudinal error and the angular error are respectively, Is a Lyapunov function parameter.
6. The trajectory tracking control method according to claim 1, wherein the establishing a differential mobile robot dynamics model based on the lagrangian method in the step4 specifically includes: Step 4.1, establishing a kinetic equation for the differential mobile robot in an ideal state by a root motion Lagrangian method, wherein the input is the output of a motion controller; step 4.2, solving a second derivative of the kinematic model, setting robot parameters, and decomposing total disturbance existing in the system to left and right tires; and 4.3, substituting the second derivative and the mobile robot parameters into a dynamics equation to establish a dynamics model.
7. The trajectory tracking control method of claim 6, wherein the robot parameters include mass, moment of inertia, and tire radius.
8. The trajectory tracking control method according to claim 6, characterized in that the dynamics model is: Wherein, the Representing the system's inertial matrix, For the system state estimation error(s), For the total disturbance to which the system is subjected, 、 The components of the total disturbance on the left and right wheels respectively, Represented as an input transformation matrix of the system, Output torque vectors representing left and right driving wheels of the mobile robot.
9. The trajectory tracking control method according to claim 1, wherein the number of neurons is 2, the number of hidden neurons in the middle is 5, and the transformation function is: Wherein, the For the center of symmetry of the transformation function, For the width of the transformation function, As a weight vector of the weight vector, In order to be a function of the transformation, For the estimated value output by the network model, 。

Description

Track tracking control method for tire slip estimation in orchard environment Technical Field The invention relates to the field of tracking control, in particular to a track tracking control method for tire slip estimation in an orchard environment. Background The track following problem is taken as a basic requirement for normal operation of the mobile robot, and continuous development in the field has an indispensable effect on robot research. In recent years, a large number of students at home and abroad research the track tracking problem of various mobile robots, so that the track tracking technology is gradually developed and tends to be mature. The existing track tracking algorithm is mainly divided into a pure track tracking algorithm based on a kinematic model and a Stanley algorithm, and an LQR algorithm, an MPC algorithm based on a dynamic model and the like, and the different algorithms have respective advantages. The LQR algorithm is a control method aiming at state feedback, can carry out linearization processing on a nonlinear model and solve an optimal solution by designing an optimal quadratic function, but ignores nonlinear parts in a plurality of practical application scenes and has high solving process cost. There are few studies about track tracking problems in field orchard environments, but most of the existing track tracking algorithms are designed based on kinematic models, such as Pure-track (PP) and Stanley algorithms. Although the algorithm has simple thought and can realize the effect of better performance on most trajectories, the algorithm has poorer performance even has the phenomenon of tracking failure when encountering a severely constrained expected trajectory or the actual application environment is complex. The controller is designed for the above based on a kinematic model. Disclosure of Invention The invention aims to provide a track tracking control method for tire slip estimation in an orchard environment, which can efficiently and accurately realize track tracking in the orchard environment. The technical solution for realizing the purpose of the invention is as follows: The invention designs a novel control method for estimating longitudinal sideslip parameters and observing sideslip speeds of a sliding mode on line, which is a neural network self-adaptive sliding mode control method based on a dynamic model, when a kinematic controller can give expected linear speeds and angular speeds through errors between given expected tracks and actually output poses, moment output can be realized through a dynamics controller, the moment output by a left wheel and a right wheel can be accurately controlled, and then the actual linear speed of the output by the left wheel and the right wheel is controlled, and the track tracking control method for the tire slip estimation in the orchard environment is provided, and specifically comprises the following steps: step 1, establishing a differential mobile robot kinematic model with a drive shaft not coincident with a geometric center based on the differential mobile robot kinematic model in the existing ideal state; Step 2, establishing a track tracking error equation based on a differential mobile robot kinematic model and combining an expected track model; step 3, designing an outer loop kinematic controller based on a Lyapunov direct method according to a track tracking error equation, and obtaining the control speed and the angular speed of the system under an ideal state through errors; Step 4, establishing an existing differential mobile robot dynamics model based on a Lagrangian method, designing an inner ring dynamics controller based on the dynamics model, and converting linear speed and angular speed in an ideal state into torque control on left and right wheels; and 5, carrying out online estimation on unknown items and uncertain parameters existing in the inner ring dynamics model through the neural network model. Compared with the prior art, the invention has the following effects: 1. The wheel type mobile robot with the center of the driving shaft not coincident with the geometric center is used as a research object, has universality and can optimize the design of a kinematic model and a controller; 2. when the controller is designed, the slipping phenomenon caused by soft soil in the orchard environment is considered, and the response speed and the accuracy of an actual track tracking task are improved by compensating the slipping degree. Drawings FIG. 1 is a schematic diagram of a mobile robot under study according to the present invention. Fig. 2 is a schematic block diagram of a dual closed loop control system of the present invention. Fig. 3 is a flow chart of kinematic modeling and error equation derivation. Fig. 4 is a flow chart of designing a controller based on tracking error. Fig. 5 is a flow chart of neural network adaptive control. Detailed Description For a better understanding of the steps, advantage