Building a Search Tree for a Pilot System of a Rescue Search Robot in a Discretized Random Step Environment

Rescue robotics applies search and rescue robots to expand rescue capabilities while increasing safety. Mobile robots working at a disaster site are monitored remotely by operators who may not be able to see the site well and select work paths appropriately. Our goal is to provide a “pilot system” that can propose options for traversing 3D debris environments. This requires a special debris path search algorithm and an appropriately defined search tree ensuring smooth exploration. To make a path search feasible in huge real state space we discretize search space and robot movement before a search. In this paper we present path quality estimation and search tree branching function F , which defines search tree building process online through node opening and branching. Well-defined function F removes unsuitable search directions from the search tree and enables dynamic path planning accounting for debris. Exhaustive simulation was used to structure and analyze data. Experiments confirmed the feasibility of our approach.