T. Okada, A.M. Mahmoud, and W.T. Botelho (Japan)
Legged robot, Skid-steering, Walking, Trajectory estimation, Coefficient of friction, Load distribution.
This paper proposes an iterative algorithmic method for estimating a walking trajectory of a legged mobile robot. There are two assumptions: First; the ground is an uneven but smooth and sloped with a certain angle with differ ent ground properties. Second; the robot has no steering mechanism for moving by sliding and skidding effects as a legged skid-steering mobile robot (LSSMR). In the discus sion, we analyze a leg motion combined with sliding, slip ping, skidding and sluing for generalizing a mathematical formulation of an energy cost. Then, we minimize the cost of motion to find an actual movement utilizing the princi ple of virtual work. The minimization makes it possible to simulate the walking for visualizing trajectories while the robot turns or spins on different grounds. Finally, we verify simulation results by demonstrating the hybrid robot PEOPLER-II as a legged-type. Experiments show that the method is valid and effective for depicting walking trajec tories. Also, it is widely applicable to multi-legged SSMRs even when the robot is limping in a broken leg.
Important Links:
Go Back
