Robot Control Linearizing Differential Motion Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robot control methods face challenges such as physical constraints, singular configurations, and multi-solution switching when calculating inverse solutions, which affect the accuracy of positional calculations. Additionally, optimizing these calculations as nonlinear optimization problems is inefficient.
Innovation Solution
A robot control method that linearizes the differential motion model of a robot to transform the problem into a linear one, allowing for the determination of a predicted state and an expected state based on admittance control equations. This method then calculates compensation values for joint velocities and adjusts joint positions to improve computational accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the method of optimization problem solving is used to improve calculation accuracy, then the accuracy of positional calculation is improved, but the computational efficiency deteriorates due to the nonlinear nature of the differential motion model
Solution Approach 1:
The patent transforms the nonlinear differential motion model into a linear motion model by changing the mathematical parameters and formulation. This linearization allows the use of efficient linear optimization algorithms while maintaining sufficient accuracy for robot control, thus resolving the contradiction between calculation accuracy and computational efficiency.
Solution Approach 2:
The patent substitutes the nonlinear optimization approach with a linear optimization approach. By replacing the complex nonlinear mathematical model with a linear equivalent, the system achieves both computational efficiency and acceptable accuracy for real-time robot control applications.
2Ease of operation
If the method of inverse solution calculation is used to obtain joint angles from end trajectory, then the control process is simplified, but the accuracy of positional calculation deteriorates due to physical constraints and singular configurations
Solution Approach 1:
The patent changes the control approach from inverse kinematics (solving for joint angles from end position) to direct linear optimization (solving for optimal joint commands directly). This parameter transformation eliminates the singular configuration problems and physical constraint issues inherent in inverse solution methods while maintaining computational simplicity.
Data Source
AI summary
A robot control method, a robot, and a computer-readable storage medium are provided. The method includes: obtaining a linear motion model of a robot; determining a predicted state corresponding to each moment in a preset time period based on the linear motion model; determining an expected state corresponding to each moment in the preset time period; and determining a compensation value of a velocity of joint(s) at each moment from k-th moment to k+N−1-th moment based on the predicted state corresponding to each moment in the preset time period and the expected state corresponding to each moment in the preset time period, determining instruction parameter(s) at the k-th moment based on the compensation value of the velocity of the joint(s) at the k-th moment, and adjusting a position of each of the joint(s) of the robot according to the instruction parameter(s) at the k-th moment.