Protocol design for performance improvement in VANETs

Abstract

Nowadays, road accidents and traffic congestion pose serious societal and economic problems to both developed as well as developing countries like India. Vehicular Ad hoc Networks (VANET), a new type of Mobile Ad hoc Networks (MANET) can address these problems and can also be used for non-safety infotainment applications. VANET offers Vehicle-to- Vehicle (V2V) and Vehicle-to-Roadside (V2R) wireless communication network. In a VANET, a vehicle uses advanced sensors for gathering information and uses wireless medium for transmitting information to other vehicles. The dynamic nature of VANET results in large variations in node density and frequent fragmentations of disconnected nodes causing unique challenges in routing, bandwidth management and link management. Thus, the thesis deals with enhancing the performance of VANET by formulating an effective routing strategy, clustering techniques, message dissemination techniques and traffic flow management schemes. In the first part of the thesis, a Multipath Route Restoration Protocol (MRRP) is proposed for VANET scenario to formulate an effective routing strategy that focuses on restoring the original optimal path after congestion subsides below a threshold level. Also, this work inherits a route discovery model to handle congestion occurring in the existing optimal path. From our research work, it has been observed that certain improvement has been observed in terms of end-to end delay reduction and improved overall performance of the network. In the second part of the thesis, a Travel-time based Clustering Algorithm (TTCA), which is a TDMA-based cluster scheduling scheme for VANET based on a vehicle‟s driving time is proposed. As per the existing European Union regulations, a vehicle cannot drive continuously for more than 9 hours. Taking account of this rule, the proposed algorithm considers a vehicle‟s on-road time, along with traditional cluster formation parameters to form a more stable cluster. In the third part of the thesis, a Topology-based Emergency Message Dissemination (T-EMD) scheme is proposed and is based on road map complexity and message categorization. T-EMD approach significantly reduces Broadcast Storm Problem (BSP) by enhancing on the existing TLC framework by adaptively determining the quantity and type of message that needs to be sent, based on the topology of the accident location. In the fourth part of the thesis, an Adaptive Signal Controller for Urban Traffic (ASCUT) controller based on Kerner‟s traffic flow theory is proposed. It uses the disseminated broadcast messages from vehicles and determines the optimal green time relative to incoming traffic volume. Kerner‟s theory has classified a congested traffic situation into synchronized flow and wide moving jam. The proposed controller not only reduces waiting time of vehicle platoons at traffic intersections, but also to streamline traffic and improve cluster management by sending warning messages to improve the traffic flow. Experimental evaluation of the proposed clustering techniques has been done using network simulators such as NS-2 simulator and Omnet++, traffic simulators such as SUMO and the simulation results demonstrate improvement in terms of packet delivery ratio, cluster stability, delay and congestion clearance time. The proposed techniques of this research work can be applied to overcome challenges that occur in traffic management systems for both urban and rural areas. Furthermore, this research work can be extended to develop an analytical model to verify the improvements in waiting time of nodes at intersections and queue lengths of clusters. The parameters for TTCA approach can be optimized using multi-objective evolutionary algorithms that will lead to improved cluster stability.