Non-Circular Robot Path Planning With 3D Orientation Costmaps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional path planning algorithms for non-circular robots in environments with obstacles are computationally intensive, leading to increased planning time and resource usage, making them unsuitable for real-time navigation and requiring significant computational effort, especially when considering the robot's orientation and shape.
Innovation Solution
A method utilizing a 3D costmap that accounts for the robot's orientation and shape by iteratively selecting consecutive positions to minimize cumulative cost values, including a stress value that reflects changes in movement direction, allowing for real-time collision-free path planning optimized for non-circular robots.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional path planning algorithms are used for non-circular robots, then collision-free paths can be planned, but the computational effort and planning time are significantly increased
Solution Approach 1:
The patent segments the continuous configuration space into discrete 3D costmap cells, where each cell represents a specific robot pose (position and orientation). This discretization allows the use of efficient graph search algorithms while maintaining accuracy for non-circular robots, resolving the contradiction between collision-free planning reliability and planning time.
Solution Approach 2:
The patent extends the traditional 2D grid map to a 3D costmap by adding the orientation dimension as the third axis. Each cell in the 3D costmap stores cost information for a specific robot orientation at that position. This dimensional extension enables efficient representation of orientation-dependent collision costs without requiring computationally intensive continuous optimization, thus reducing planning time while ensuring collision-free paths.
2Measurement precision
If conventional path planning algorithms account for robot orientation and shape, then accurate collision-free paths are achieved, but computational resources are significantly consumed
Solution Approach 1:
The patent performs preliminary computation of the 3D costmap before path planning, pre-calculating collision costs for all possible robot poses in the environment. This pre-computation stores orientation-dependent collision information in an efficient data structure, allowing the actual path planning to query this pre-computed information without重新calculating collisions, thus maintaining high precision while reducing real-time computational resource consumption.
Solution Approach 2:
The patent creates a computational model (3D costmap) that copies and represents the physical environment's collision characteristics in a simplified digital format. This virtual model allows the robot to plan paths by querying pre-computed collision costs rather than performing real-time geometric calculations, maintaining measurement precision while significantly reducing computational resource usage during actual navigation.
3Reliability
If the minimum required distance to obstacles depends on robot orientation, then accurate collision avoidance is achieved, but the path planning complexity increases due to the piano mover's problem
Solution Approach 1:
The patent resolves the piano mover's problem by adding the orientation dimension to create a 3D costmap. Instead of complex continuous optimization to find valid orientations at each position, the system discretizes orientations into fixed angular steps and pre-computes collision costs for each orientation cell. This transforms the complex continuous problem into a simpler discrete search problem, maintaining collision avoidance accuracy while reducing planning complexity.
Solution Approach 2:
The patent changes the parameter representation from continuous position and orientation to discrete 3D costmap cells with quantized orientations. By parameterizing the configuration space in this discrete manner, the system avoids the computational complexity of continuous optimization while maintaining sufficient accuracy for collision avoidance, effectively resolving the piano mover's problem.
Data Source
Figure 1~2
Figure 3A~3B
Figure 4
AI summary
A technique of path planning for a non-circular robot in an environment comprising obstacle. A 3D costmap which comprises a 2D map of a plane and a cost value per position on the 2D map, indicative of a cost of positioning the robot with a predetermined orientation in the presence of obstacles in a vicinity of the position is received. A start position and an end position of the robot comprising a start and an end orientation of the robot are received. Consecutive positions for a path are iteratively selected from the start position to the end position. A cumulative cost value over the consecutive positions, which comprises a sum of the cost values, is minimized. The cost value further comprises a stress value indicative of a change in a movement direction along the planned path.