Minimum-energy trajectory planning and control for TOMRs

Abstract

To maximize the total operation time of three-wheeled omni-directional mobile robots (TOMRs), min-imum-energy translational and rotational velocity trajectory planning and control algorithms are investigated using Pontryagin’s minimum principle, a numerical analysis, and a practical cost function as the total energy drawn from the batteries to motors, based on the accurate TOMR dynamic model including both actuator motor dynamics and the Coriolis force In this dissertation, the minimum-energy trajectory planning problem for TOMRs is configured as four sub-problems according to the given constraints. Also the minimum-energy trajectory problem on one of the possible transport connections in the industrial environment is formulated. First, we formulate the minimum-energy translational trajectory generation problem on a straight line path with a fixed-heading condition. Using Pontryagin’s minimum principle, we find the minimum-energy translational velocity trajectory in analytic form. To determine the energy efficiency obtainable, we conduct the simulation using the minimum-energy translational velocity trajectory, the loss-minimization velocity trajectory, and the conventional velocity trajectory with the energy-efficient trapezoidal velocity profile. The energy-efficient trapezoidal velocity profiles have the energy-efficient acceleration rate and the constant ve-locity which are obtained using ‘fmincon’ function in the MATLAB optimization toolbox. Also a trajectory following system for TOMRs is implemented using a resolved-acceleration controller with motor voltage input to validate the actual performance of the minimum-energy trajectory. Second, we investigate the minimum-energy translational and rotational trajectory planning problem on a straight line with a self-rotation. Using the optimal control theory, we find the minimum-energy rotation-al velocity trajectory in analytic form. Then the minimum-energy translational velocity trajectory is found using the ...