An Approach to Predicting the Location of Moving Objects During On-Road Navigation
Craig I. Schlenoff, Rajmohan Madhavan, Stephen B. Balakirsky
For an autonomous vehicle to navigate in real-time within a dynamic environment, it must be able to respond to moving objects. In particular, it must be able to predict, with appropriate levels of confidence, where those objects are expected to be at times in the future. It must then capture this information internally in its world model in a format amenable for planners that intend to use it.In this paper, we provide an overview of a framework to address the challenges involved in predicting and representing the future location of moving objects. This framework uses a multi-representational approach to model information about moving objects, thus allowing for planners that require different forms of knowledge representation. We then describe a probabilistic, logic-based algorithm to predict the future location of vehicles in an on-road environment. Included in this discussion are the factors that affect the probabilities associated with various actions that the moving object may take.
18th International Joint Conference on Artificial Intelligence
, Madhavan, R.
and Balakirsky, S.
An Approach to Predicting the Location of Moving Objects During On-Road Navigation, 18th International Joint Conference on Artificial Intelligence, Acapulco, -1, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=824436
(Accessed February 22, 2024)