Multi-UAV Path Planning in a Spreading Wildfire
Rachit Aggarwal, Alexander A. Soderlund, Mrinal Kumar · AIAA Scitech 2021 Forum · 2021
View Video Presentation: https://doi.org/10.2514/6.2021-0866.vid This work addresses the problem of path planning for multiple Unmanned Aerial Vehicles (UAVs) tasked to provide situational awareness in a spreading wildfire. While the wildfire's evolution can be estimated using a wildfire forecast model to aid suppression tactics, it is important to update the estimates using in-field measurements to improve the forecast. The information fusion step utilizing the forecast and the ground-level temperature sensors to improve the estimates can often result in disagreement or conflict. By utilizing additional aerial measurements performed by a fleet of UAVs, these conflicts can be resolved. However, the extreme temperatures in the regions immediately above the wildfires pose operational threat to UAVs. The objective of the multi-UAV path planning problem is to allocate conflict locations to the vision-equipped UAVs such that it optimizes the overall energy requirements to fly to high conflict locations. Using the wildfire forecaster and the radiation model, the heat flux at the flight level is estimated. Computationally efficient discrete path planning techniques are presented that provide approximate paths which are used to formulate the assignment problem. Once the assignments are made, kinematically feasible and optimal paths are computed using Gauss Quadrature collocation method. It is shown that by employing this multi-step approach, the path planning can be performed in quick successions allowing recurring implementation in a spreading wildfire.