Abstract:

We address motion planning for a mobile agent to acquire information from multiple monitored tar- gets using sensors with limited capabilities. To represent sensor limitations, such as range, field of view, and allowed acquisition directions, we introduce a nonnegative metric that quantifies the rate of information acquisition and is positive only when these limitations are satisfied. To en- able optimization-based trajectory generation, we impose temporal logic specifications to ensure that the information acquisition metric and its gradient are nonzero over some time interval. This enables the application of continuous- time successive convexification to solve the motion planning problem. We demonstrate the proposed approach in a case study of a quadrotor that must acquire information on multiple targets with sensor range, field of view, and acquisition direction constraints.