Automated Synthesis of Controllers for Search and Rescue from Temporal Logic Specifications (1304.6898v1)

Abstract: In this thesis, the synthesis of correct-by-construction controllers for robots assisting in Search and Rescue (SAR) is considered. In recent years, the development of robots assisting in disaster mitigation in urban environments has been actively encouraged, since robots can be deployed in dangerous and hazardous areas where human SAR operations would not be possible. In order to meet the reliability requirements in SAR, the specifications of the robots are stated in Linear Temporal Logic and synthesized into finite state machines that can be executed as controllers. The resulting controllers are purely discrete and maintain an ongoing interaction with their environment by changing their internal state according to the inputs they receive from sensors or other robots. Since SAR robots have to cooperate in order to complete the required tasks, the synthesis of controllers that together achieve a common goal is considered. This distributed synthesis problem is provably undecidable, hence it cannot be solved in full generality, but a set of design principles is introduced in order to develop specialized synthesizable specifications. In particular, communication and cooperation are resolved by introducing a verified standardized communication protocol and preempting negotiations between robots. The robots move on a graph on which we consider the search for stationary and moving targets. Searching for moving targets is cast into a game of cops and robbers, and specifications implementing a winning strategy are developed so that the number of robots required is minimized. The viability of the methods is demonstrated by synthesizing controllers for robots performing search and rescue for stationary targets and searching for moving targets. It is shown that the controllers are guaranteed to achieve the common goal of finding and rescuing the targets.