Use of simulator motion feedback for different classes of vehicle dynamics in manual control tasks

With the development of moving-based flight simulators in mind, a large number of researchers have considered pilot manual control behaviour in tasks where the motion of the vehicle can be felt by pilots. While it is known that the dynamics of the controlled vehicle are a key factor that determines the usefulness of motion feedback for pilots, currently available methods for analysing multimodal pilot control behaviour have not yet been systematically applied to study the usefulness of motion feedback over a wide range of different vehicle dynamics. Therefore, this paper describes a study of the effect of simulator motion feedback on task performance and pilot control behaviour for different types of vehicle dynamics using multimodal pilot identification techniques. An experiment was performed in the SIMONA Research Simulator at Delft University of Technology where participants were instructed to perform a heading tracking task both with and without motion feedback, for a wide range of different vehicle dynamics. The considered dynamics were position control (gain), velocity control (single integrator) and acceleration control (double integrator). The experiment results indicate that the performance of double integrator control tasks is significantly better with motion feedback than that without motion, but show no performance improvement for the gain and single integrator control tasks provided with motion, respectively. In addition, the analysis of pilot control behaviour clearly shows that for double integrator control tasks the motion feedback is fully utilized by the pilots, but for gain and single integrator control tasks the motion is not utilized by the pilots. The performance improvement seen for the double integrator control tasks with motion feedback is explained by the fact that pilots make explicit use of feedback on the accelerations of the motion of the vehicle. The experiment results confirm previous observations in the aspects of task performance and pilot adopted control behaviour that motion feedback is very important for controlled vehicle dynamics for which pilots need to generate lead (i.e., double integrator), but perhaps not for vehicles with more stable dynamics.