Information-Driven Path Planning for UAV With Limited Autonomy in Large-Scale Field Monitoring

This article presents a novel information-based mission planner for drone tasked to monitor spatially distributed dynamical phenomenon. For the sake of simplicity, area be monitored is discretized. The insight behind proposed approach that, due spatiotemporal dependencies observed phenomenon, one does not need collect data on entire area, which main limiting factors in unmanned aerial vehicle (UAV) applications their limited autonomy. In fact, unmeasured states can estimated using an estimator, such as Kalman filter. this context, planning problem becomes generating flight plan that maximizes quality state estimation while satisfying constraints (e.g., time). first result formulation special orienteering where cost function measure estimation. results mixed-integer semidefinite formulation, optimally solved small instances problem. larger instances, heuristic proposed, provides suboptimal results. Simulations numerically demonstrate capabilities and efficiency path-planning strategy. We believe has potential increase dramatically monitor, thus increasing number monitoring with drones become economically convenient. Note Practitioners —This was motivated by performing large-scale field activities (UAV), at moment very time-consuming limits definitive adoption UAVs kind activities. caused autonomy commercial lack systematic ways missions so maximize amount information collected. work starts from observation many applications, phenomena wants observe have dynamics statistical properties. Accordingly, are directly measured characterizable error. article, we develop theoretical foundations path define designing optimal optimization based knowledge presented improve effectiveness applications.