Robot Motor Control Using Decoupled Joint Dynamics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for controlling motors of robots with articulated joints require significant computing capacity, making real-time control and regulation challenging, especially in dynamic environments with elastic joints and varying payloads.
Innovation Solution
A method involving a system of coupled and decoupled equations of motion is proposed to simplify motor control, allowing for efficient real-time regulation by transforming manipulated variables and state limitations, enabling precise control with reduced computational demands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a first system of coupled equations of motion is used to describe rigid-body or flexible-body dynamics of the robot manipulator, then control precision and accuracy are improved, but computing time and computational complexity increase significantly
Solution Approach 1:
The system divides the coupled equations of motion into multiple decoupled equations of motion, where each equation describes the dynamics of a single joint independently. This segmentation allows parallel computation of each joint's control parameters without requiring complex matrix operations on the full coupled system, significantly reducing computational time while maintaining control precision.
Solution Approach 2:
The invention extracts and eliminates the coupling terms between joints from the equations of motion. By taking out the interaction forces and moments between adjacent links, the system obtains simplified decoupled equations that can be solved independently for each joint, reducing the overall computational complexity while preserving the essential dynamics of each individual joint.
2Reliability
If coupled equations of motion are solved in real-time for robot control, then dynamic accuracy is improved, but device complexity and computational resources required increase
Solution Approach 1:
The control system is segmented into independent controllers for each joint, where each controller solves a simple decoupled equation rather than participating in a complex coupled system solution. This segmentation reduces the computational burden on the central processor and allows for more reliable real-time control with simpler computational resources.
Solution Approach 2:
The coupling effects between joints are extracted and compensated for through predefined transformation matrices and gravity compensation terms, allowing each joint to be controlled independently. This extraction approach maintains dynamic accuracy by accounting for coupling effects without requiring complex real-time solution of coupled differential equations.
3Adaptability or versatility
If elastic joints with defined intrinsic elasticity are used in the robot manipulator, then adaptability and compliance are improved, but control complexity and computational demands increase
Solution Approach 1:
The elastic properties of the joints are extracted as separate stiffness and damping parameters that can be independently compensated for in the control algorithm. By taking out the elastic effects from the dynamic model and representing them as distinct parameters, the system can maintain compliance while using simpler control equations that do not require solving complex flexible-body dynamics in real-time.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method and device for controlling and regulating motors MOTm of a robot, with m = 1, 2, …M, wherein the robot has robot components that are interconnected via a number N of articulated connections GELn, the joint angles of the articulated connections GELn can be adjusted by means of associated motors MOTm; Z(tk) is a state of the robot components in an interval tk; and a first system of coupled motion equations BGG is predetermined and describes rigid-body dynamics or flexible-body dynamics of the connected robot components. In the first system of motion equations BGG, um(tk) is a manipulated variable for the respective motor MOTm. For the first system of coupled motion equations BGG, restrictions of the manipulated variables um(tk) and restrictions of the states Z(tk) of the connected robot components are predetermined. The method comprises the following steps: for the first system of coupled motion equations BGG, providing (101) a second system of locally equivalent decoupled motion equations BGE that describes the rigid-body dynamics or the flexible-body dynamics of the connected robot components; providing (102) restrictions of the manipulated variables um(tk) transformed into the second system and providing restrictions of the states Z(tk) transformed into the second system; providing (103) the state Z(tk) transformed into the second system as Z*(tk); for the second system of decoupled motion equations BGE, setting (104) a target state SZ* of the robot manipulator which is to be reached starting from the state Z*(tk), and setting (104) one or more conditions BD* and/or one or more characteristics KZ* that define how to achieve the target state SZ*; in the second system of decoupled motion equations BGE, predicting (105) a state trajectory ZT*(t) and the associated manipulated variable trajectories uT*m(t) depending on the state Z*(tk) and the target state SZ* while meeting the conditions BD*, the characteristics KZ*, the transformed restrictions of the manipulated variables um(tk), and the transformed restrictions of the states Z(tk) for an interval of t = tk to t = tk+W, wherein ∆t = tk+W – tk is a predetermined prediction interval; transforming (106) the manipulated variable trajectories uT*m(t) and the state trajectories ZT*(t) into the first system of coupled motion equations BGG to generate manipulated variable trajectories uTm**(t) and state trajectories ZT**(t); from the manipulated variable trajectories uTm**(t), determining (107) manipulated variables um(tk+1) for the interval k+1 and regulating the motors MOTm by means of the manipulated variables um(tk+1); from the state trajectories ZT**(t) and/or on the basis of sensor data of a detection system of the state Z(t), determining (108) the state Z(tk+1) for the interval k+1; and for Z(tk) = Z(tk+1), performing again the method, starting with step (103), until a predetermined break-off criterion or the target state SZ* is reached.