Robot Control System for Moving Part State Combinations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robot control methods for unstructured environments face challenges in obtaining appropriate control laws for the moving part when there are many combinations of arm and moving part states, particularly due to the consideration of inertia forces, making it difficult to perform target motions such as gripping, carrying, or processing objects efficiently.
Innovation Solution
A robot control system that includes a motion planning part, an arm control part, and a movement control part, which generates arm and movement control instructions to ensure the moving part operates within an allowable range, using model predictive control and optimal control techniques to account for arm state transitions and inertia forces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If independent arm controller and movement controller are used, then it is possible to perform new target motions by changing only arm controller control law, but it becomes difficult to obtain appropriate movement controller control law when there are many combinations of arm state and moving part state
Solution Approach 1:
The control system is divided into independent arm controller and movement controller, each handling specific control tasks. The arm controller manages arm motion while the movement controller handles moving part positioning, allowing separate optimization and simplifying the overall control law acquisition process despite multiple state combinations
Solution Approach 2:
The movement controller pre-calculates and stores control laws for various arm states in advance. By preparing control laws beforehand for different arm position scenarios, the system avoids real-time complexity when dealing with multiple state combinations during actual operation
2Reliability
If control law for moving part is learned under specific conditions, then stable control is achieved under those conditions, but it becomes difficult to obtain appropriate control law when many combinations of arm state and moving part state must be considered
Solution Approach 1:
The movement controller is designed to handle multiple arm states using a unified control framework. By creating a universal control law that works across different arm positions and configurations, the system maintains stable control performance while adapting to various state combinations without requiring separate control laws for each scenario
Solution Approach 2:
The control system uses simulation environments to create virtual copies of the physical robot for learning and testing control laws. By training the movement controller in simulated scenarios with various state combinations, the system achieves stable control performance that can be transferred to the real robot without requiring extensive real-world experimentation
3Extent of automation
If deep reinforcement learning is used to obtain control law for moving part, then control instruction can be generated, but it requires considering too many state combinations including inertia force effects
Solution Approach 1:
The control system dynamically adapts the movement controller's behavior based on the current arm state. Rather than pre-computing control laws for all possible state combinations, the movement controller uses the current arm position and velocity information to dynamically determine appropriate control actions, reducing the computational burden while maintaining accuracy
Solution Approach 2:
The system introduces an intermediary dynamic model that bridges the arm controller and movement controller. This intermediary layer processes arm state information and translates it into appropriate movement control commands, avoiding the need to directly handle all possible state combinations in the deep reinforcement learning process
Data Source
AI summary
[Problem]To provide a robot control system that can easily obtain an appropriate moving part control law even when there are many combinations of arm states and moving part states to achieve a target motion.[Solution]A robot control system includes: a robot including an arm and a moving part; and a control device that controls the robot. The control device includes; a motion planning part that outputs a target motion associated with a motion of the arm and an allowable range for the motion of the moving part corresponding to the target motion; an arm control part that outputs an arm control instruction associated with the target motion and an arm state transition; and a movement control part that generates a movement control instruction associated with a motion of the moving part to fall within the allowable range using the arm state transition. The robot has the arm controlled by the arm control instruction and has the moving part controlled by the movement control instruction.