Blended Shared Control of Zermelo's navigation problem

Many machines-from hydraulic excavators to mobile wheelchairs-are manually controlled by a human operator. In practice, the operator assumes responsibility for completing a given task at maximum utility, even though the optimal inputs may be unknown to the operator. Here we discuss a simple technique termed Blended Shared Control, whereby the human operator commands are continually merged with the commands of a robotic agent. This approach is shown to result in a lower task completion time than manual control alone when applied to a problem motivated by Zermelo's navigation problem. Experimental results are presented to compare blended shared control to other types of controllers including manual control, heads up display, and haptic feedback. Trials indicate that the shared control does in fact decrease task completion time when compared to fully manual operation.