Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Precursor systems analyses of automated highway systems. Knowledge based systems and learning methods for AHS. Volume 10. Final report, September 1993-February 1995

Technical Report ·
OSTI ID:91591
Managing each AHS vehicle and the AHS system as a whole is an extremely complex yndertaking. The authors have investigated and now report on Artificial Intelligence (AI) approaches that can help. In particular, we focus on AI technologies known as Knowledge Based Systems (KBSs) and Learning Methods (LMs). Our primary purpose is to identify opportunities: we identify several problems in AHS and AI technologies that can solve them. Our secondary purpose is to examine in some detail a subset of these opportunities: we examine how KBSs and LMs can help in controlling the high level movements--e.g., keep in lane, change lanes, speed up, slow down--of an automated vehicle. This detailed examination includes the implementation of a prototype system having three primary components. The Tufts Automated Highway System Kit(TAHSK) discrete time micro-level traffic simulator is a generic AHS simulator. TAHSK interfaces with the Knowledge Based Controller (KBCon) knowledge based high level controller, which controls the high level actions of individual AHS vehicles. Finally, TAHSK also interfaces with a Reinforcement learning (RL) module that was used to explore the possibilities of RL techniques in an AHS environment.
Research Organization:
Tufts Univ., Medford, MA (United States)
OSTI ID:
91591
Report Number(s):
PB--95-253639/XAB; CNN: Contract DTFH61-93-C-00196
Country of Publication:
United States
Language:
English