Embedded Files
Problem
Find the model parameters to satisfy the end rotation and translation of a 3-section constant-curvature robot.
Overview
The solver is proposed by bridging both the theoretical reduction and numerical correction.
Evaluation
Over 2,000 random experiments with or without obstacles.
Success Rate
Full percentage in both experiments.
No dependency on initial values.
Ability to obtain multiple solutions.
Fewer steps of numerical iterations (20% with 0 steps, > 85% within 2 steps).
Less overall runtime (15% in free space, 25%with obstacles, regardless of the success rate).
Efficiency
Multiple Solutions
The circular traversal ensures the resolution completeness. The equilibria of error function are candidate solutions.
A few steps of numerical iterations are applied to reach the final solutions (logarithmic scale for the y-axis).
Finally, there are 4 multiple solutions in this example.
Path Planning
ctm_mpt_star_demo2.mp4A Pentagram-Shaped Path.
Achieved by greedy algorithm.
The Optimal RSS-Shaped Path.
Achieved by dynamic programming.
ctm_mpt_rss_demo2.mp4Google Sites
Report abuse