Smooth online time-varying input shaping with fractional delay FIR filtering

Abstract

Time-Varying Input Shaping has been proven effective in suppressing unwanted mechanical vibrations in industrial robots. However, no appropriate method exists for robots with online trajectory generation and position control with feed-forward torque. Based on detailed analysis, a new method is proposed, which employ Discrete Time Time-Varying Input Shaping Technology with an alternative discrete time convolver named the Output Side Algorithm FIR filter combined with numerical differentiation. Second, the proposed method is extended to be seen as a Fractional Delay FIR filter, which can be extended to higher order for additional effectiveness, and hereby forming a second new method. Numerical simulations are conducted on a simple time-varying dynamic system with parameters based on a UR10e robot. The simulations show very good vibration suppression with the two new methods, which have great potential application to the vibration control of practical robotic systems.