An information-theoretic path planning for exploration mission in unknown environment is proposed. Instead of using traditional grid-based occupancy maps, the Gaussian process (GP) is used to build an environmental map. The GP map performs inference directly on the collected sensor datasets which allows it to infer the probability of collision for any query point in continuous 3D space, removing the need for maintaining a full discrete map. This GP map is fused with the Rapidly exploring random tree (RRT) path planner to plan a safe path and acquire information about the unknown environments. Using mutual information as an information measure, the most informative path is chosen as the path for exploration. Simulation results show that GP map combined with RRT planner achieves exploration task successfully in unknown complex environment and present their potential implementation in unmanned aerial vehicle (UAV) missions.
Information-Theoretic Exploration of Unmanned Aerial Vehicle in Unknown Cluttered Environment
Handbook of Unmanned Aerial Vehicles ; Chapter : 81 ; 1655-1675
2014-08-09
21 pages
Article/Chapter (Book)
Electronic Resource
English
A localization system of unmanned aerial vehicle under cluttered indoor environments
European Patent Office | 2023
|A localization system of unmanned aerial vehicle under cluttered indoor environments
European Patent Office | 2024
Intelligent camera for environment exploration unmanned aerial vehicle
European Patent Office | 2021
|