Speed control for unmanned mining trucks based on frequency-domain energy shaping and dynamic co-optimization

Speed-command-induced torsional vibration is a primary cause of drivetrain fatigue failure in unmanned mining trucks. We propose a frequency-domain energy shaping and co-optimization framework that links the speed command spectrum to system modal characteristics, suppressing vibration at its source. A three-metric system---excitation energy, residual vibration, and acceleration duration---jointly optimizes the speed profile and controller parameters. NSGA-III generates a Pareto-optimal set quantifying the vibration-suppression-mobility trade-off. Hardware-in-the-loop experiments show that the method reduces torsional vibration by 91.5% and 80.4% compared to conventional trapezoidal and active-damping strategies, respectively, while mitigating gear fatigue damage. Robustness tests confirm strong adaptability with online computation below 0.06 ms, providing a transferable framework for command-induced vibration suppression in electromechanical systems.