Online Trajectory Generation for Mobile Robot Arm Using Many-Objective Optimization*

Both mobile robots and robotic arms provide valuable capabilities in a variety of industries, from the ability of robotic arms to perform manipulation tasks accurately and repeatably in manufacturing environments, to the autonomous navigation capabilities of mobile robots deployed in myriad dynamic and changing environments. Combining the navigation capabilities of mobile robots with the manipulation abilities of robotic arms further expands the application areas of robotics, and integrated control of such a mobile robotic arm improves the manipulation capabilities beyond that of a static robotic arm by increasing the XY working field of the arm and allowing for smoother, more energy-efficient trajectories. We propose a many-objective optimization-based method of online trajectory generation for integrated control of a mobile robot arm. We focus on trajectory generation to a specified end-effector configuration in an obstacle-free environment, with the main contribution being a method of generating trajectories for a differential-drive-based mobile robot arm by coordinating wheel joint and arm joint motions. Both offline and real-time simulated trials demonstrated a $100\%$ success rate in reaching the desired end-effector configuration. Furthermore, the mobile robot arm was able to reach the desired configuration in <30 s in $73 \%$ of the trials, and <60 s in $94 \%$ of the trials, demonstrating the efficiency of the trajectories generated in real-time using this method.