Control and Vibration of a 5-DOF Long-reach Robotic Arm

Abstract

Deploying autonomous in-field robotic phenotyping platforms consisting of robot manipulators and unmanned ground vehicles, increases crops monitoring (phenotyping) efficiency; such systems can collect data quickly and accurately, and without the need for experienced operators. Objectives of this research are to simulate motion control of a long-reach 5-DOF hybrid robot manipulator and to assess its operational vibration. The manipulator has recently been developed in the Robotics lab at the University of Saskatchewan. The manipulator holds several sensors attached to its tip and places them at specific positions relative to targeted crops to obtain best possible phenotypic data. Vibration from the vehicle can transfer to the manipulator’s tip and can cause inaccuracies in collected data. It is of high importance to assess the vibration behavior of the manipulator for its intended application. Correct motion control of the manipulator tip also directly impacts data collection accuracy. Finite element modal analysis is carried out for two different configurations of the manipulator. Modal analysis on one of the operating configurations is completed using experimental modal analysis. In experiments, acceleration data are obtained by utilizing sensors and are post-processed using Fast-Fourier Transform analysis. The dominant natural frequencies obtained numerically are verified using experimental modal analysis. A harmonic response simulation is conducted for the manipulator with base excitations induced from the moving vehicle; numerical results obtained give a good indication of a safe range for the excitation frequency for operation of this manipulator. A simulation model for optimizing PID gains and testing trajectory tracking performance of the manipulator is presented. Trajectory tracking simulations are carried out by applying a fifth-order polynomial trajectory planning in joint space. Satisfactory results are obtained using optimal PID and PID-CTC controllers. It is shown that optimized PID-CTC controller can lead the manipulator on smooth and stable trajectories with virtually no steady-state error for the intended farm application. An open-loop motion control software is developed and tested on the fabricated manipulator; performance of the developed software is verified for the intended farming application. Overall, in this research, a thorough vibration analysis, motion control planning, simulations, and experiments, are accomplished for a developed manipulator.