A comparison of stationary target search algorithms in real time situated agents with variable sensor ranges

This study compares the real time and incremental heuristic search methods used for stationary target search in real time situated agents. The agent with a sensor is modeled in such a way that it can observe only the area inside its sensor range and update its observations as it proceeds through the environment. Two well known representatives of incremental and real time approaches, namely D* Lite and LSS-LRTA* (Local Search Space-Learning Real Time A*) algorithms respectively, are modified to be used for an agent model with a a specific sensor range. In addition, LPA* (Lifelong Planning A*), one of the first incremental approaches designed for stationary agents, is improved to perform route planning for moving agents. The simulations show that the algorithms behave differently and have advantages over each other as the sensor range changes.