On-road autonomous navigation of autonomous vehicles requires real--time motion planning in the presence of moving objects. Based on sensed data of moving objects in the environment and the situation it finds itself in, an autonomous vehicle has to plan a path by evaluating the future location of objects of interest. In this context, an object of interest is a moving or stationary object in the environment that has a reasonable probability of intersecting the path of the autonomous vehicle within a predetermined time frame. In this paper we investigate how situation awareness may be used to identify and select appropriate objects of interest in order to plan collision free paths. We include the concept of Dangerous Zone around a moving vehicle. Each object in the Dangerous Zone does not represent the same level of danger, and thus we use an approach based on the Fuzzy Space concept to affect a degree of risk to each object within the Dangerous Zone. Furthermore, using the high-fidelity physics-based framework for the Unified System for Automation and Robot Simulation (USARSim), we show the impact of the appropriate selection of objects of interest in term of driving by simulating different scenarios with the PRIDE (PRediction In Dynamic Environments) algorithms. PRIDE is a multi-resolutional, hierarchical framework that predicts the future location of moving objects for the purposes of path planning and collision avoidance for an autonomous vehicle.
Proceedings Title: Proceedings of the 2008 Performance Metrics for Intelligent Systems (PerMIS) Workshop
Conference Dates: August 19-21, 2008
Conference Location: Gaithersburg, MD
Pub Type: Conferences
4D/RCS, autonomous vehicle, Dangerous Zone, fuzzy logic, Fuzzy Space, long-term prediction, moving object prediction, object of interest, PRIDE, situation awareness