Agustín Torres Quintanilla
Faculty Advisor: Maxim Likhachev
The problem of coordinating multiple robotic agents for searching or exploration is relevant for many valuable applications, including emergency response and surveillance. In particular, autonomous unmanned helicopters could be employed for a variety of tasks such as search-and-rescue operations in hazardous environments, border control, surveillance of forests for fire emergencies and gathering of massive statistical data.
This project will address the problem of coordinating multiple unmanned helicopters for search-and-rescue operations. We add the extra characteristic that each helicopter carries one or several ground robots that can be dropped for sensing the target locations and identifying possible survivors. The goal of the project is to provide an application of the PPCP algorithm (Probabilistic Planning with Clear Preferences) to solve this problem. This algorithm is used to generate paths for multiple unmanned helicopters in order to locate a static object (for instance, a survivor) in partially known environments in the least possible amount of time. The PPCP algorithm is only applicable if the problem exhibits clear preferences, which we show is the case.
We pretend to test the performance of the algorithm by writing an implementation in C/C++ using the OGRE 3D rendering engine and running a simulation of the helicopters performing an efficient search for a person.