Timed trajectory generation for a toy-like wheeled robot

Abstract

In this work, we address temporal stabilization of generated movements in autonomous robotics. We focus on generating movement for a mobile robot, that must reach a target location within a constant time. Target location is online calculated by using the robot visual system, such that action is steered by the sensory information. This is a very critical issue in several robotic tasks including: catching, hitting, and humanrobot scenarios. Robot velocity is controlled through an Hopf oscillator, adapted according to temporal feedback. Timing of the velocity profile is modulated according to an adaptive mechanism that enables setting different times for acceleration and deceleration. Results on a DRK8000 mobile robot confirm the system's reliability with low-level sensors.