Motion planning for multiple autonomous vehicles

• Main Author: Kala, Rahul
• Published: University of Reading 2013
• Subjects: 629.2220285
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.602709

Unorganized traffic is characterized by the vehicles not adhering to the marked lanes, and is a reality in various countries like India. Such traffic can lead to higher traffic efficiency when the vehicles are diverse in sizes and speed capabilities - a factor currently absent in most countries. Autonomous driving technology can lead to application specialized vehicles which may introduce diversity into the presently organized traffic, necessitating a shift to unorganized traffic. Technology also has the potential to make the entire transportation system more efficient. Given a map produced by vision systems, the outcome of motion planning is then responsible for all vehicle decisions, which includes a mechanism to avoid static obstacles and other vehicles. This " planning needs to be cooperatively performed between vehicles. This is handled by the coordination mechanism. Communication, if available (in an autonomous vehicles only scenario), enables the deployment of a common planning algorithm distributed across the vehicles thereby resulting in better coordination. The literature on autonomous vehicles largely assumes lanes, while the literature on mobile robots lacks the notions of overtaking and vehicle following. This thesis takes popular algorithms from the mobile robotics literature, models them as per traffic scenarios, embeds in them the notions of overtaking and vehicle following as heuristics for faster and better planning, and analyzes all the algorithms hence produced. The study is broadened to the intelligent management of transportation systems with intelligent agents performing in diverse traffic. The thesis also considers a typical city road infrastructure for which an efficient routing strategy is devised which can lead to congestion avoidance. Cities may witness a large number of vehicles and a limited road infrastructure and hence the problem of vehicle scheduling is studied. The transportation system is made cooperative to enable all vehicles to reach their destination on time.