Improving Manual Tracking of Systems With Oscillatory Dynamics

This paper examines the manual control of systems with oscillatory dynamics. Tracking performance is improved by using input shaping to suppress command-induced oscillation. An operator study tested tracking behavior using controlled elements with both low-frequency (1.25 rad/s) and high-frequency (5 rad/s) oscillatory modes. After each experimental trial, measures of tracking performance and subjective task difficulty were recorded, and frequency-domain control characteristics were computed. Results showed that the high-frequency oscillatory mode did not greatly decrease the tracking performance from the nonoscillatory case; thus, input shaping did not produce a significant improvement in the tracking performance. However, input shaping did cause a decrease in the average subjective task difficulty and made the system closely resemble McRuer's “crossover model.” For the low-frequency case, the addition of input shaping significantly improved the tracking performance and reduced the tracking difficulty. These results demonstrate that input shaping can greatly improve the continuous tracking ability of human-machine systems that have oscillatory modes.