Search Tree Path Planning for Robot Balance Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing path planning and control methods face high computation costs and constraints when transitioning a machine from an initial state to a target state, particularly in scenarios like a humanoid robot losing balance or an airplane stalling, due to the limitations of PID control and search tree-based approaches.
Innovation Solution
A path planning method and system that utilize a search tree with pre-defined sections in state space, allowing for efficient determination and control of paths from various initial states to a target state by dividing the state space into sections and using a path generation unit to identify routes on the search tree, reducing computation cost and constraints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If PID control scheme or potential scheme is used to control a machine from initial state to target state, then the control method is simple to implement, but the machine may not work properly depending on the state where the machine lost balance and the target state is difficult to reach within a short time
Solution Approach 1:
The state space is divided into multiple sections, and a search tree is constructed with branches corresponding to different sections. This segmentation allows the system to handle different initial states more effectively while maintaining computational efficiency, resolving the contradiction between simple implementation and reliable control across various balance-loss scenarios.
Solution Approach 2:
The search tree is constructed in advance with branches representing different state transitions. By pre-computing the search tree structure, the system eliminates the need for complex real-time evaluation functions while ensuring reliable path finding from any initial state to the target state, thus improving both implementation simplicity and control reliability.
2Adaptability or versatility
If search tree-based path planning is used to determine path from initial state to target state, then the path can be determined in state space, but the computation cost for searching all derived branches becomes huge
Solution Approach 1:
The state space is divided into multiple sections, and the search tree is constructed with branches corresponding to these sections. This segmentation reduces the computational burden by organizing the search space more efficiently, allowing the system to maintain adaptability to various initial states while significantly reducing computation time compared to searching all branches of a traditional search tree.
Solution Approach 2:
The search tree is constructed in advance with all possible branches representing state transitions. By pre-computing and storing the search tree structure, the system eliminates the need for expensive real-time search operations, thus reducing computation time while maintaining the ability to handle various initial states through the pre-built tree structure.
3Loss of time
If evaluation function is used to select branches in search tree, then computation cost may be reduced, but the evaluation function needs to be determined for each object based on expertise of a designer and search may not be sufficiently performed
Solution Approach 1:
The search tree structure is designed to be universal and applicable to different objects and scenarios without requiring object-specific evaluation functions. The tree branches represent general state transitions that can be used across various balance-loss scenarios, eliminating the need for designer expertise in creating evaluation functions while maintaining sufficient search capability for different objects.
Solution Approach 2:
The search tree is constructed in advance with all possible state transitions represented as branches. This pre-computation eliminates the need for real-time evaluation functions, reducing computation time while maintaining generality. The pre-built tree structure can be applied to different objects without requiring object-specific customization, thus achieving both computational efficiency and broad applicability.
Data Source
AI summary
A path planning system for bringing state of an object into a target state includes a search tree production unit for producing in advance, in a state space with said target state defined as a root, a search tree having a branch at each one of a plurality of sections of the state space, said state space being divided into the plurality of sections in advance. The system also includes a search tree memory unit for storing the search tree, and a path generation unit for determining, a route on the search tree from the branch corresponding to the current state to the root. The path planning/control system further includes a path control unit for controlling the path of the object to bring the state of the object into the target state in accordance with the route on the search tree determined by the path planning system.