Map updates

Abstract

Digital maps, based on which navigation systems in vehicles operate, require frequent updates to reflect changes in real world. These changes currently being issued using off-line mechanisms are cumbersome in nature as well costly. Therefore a need exists, to come up with a wireless solution which can not only perform updates without human intervention but also be cost effective. IEEE 802.11p a standard, enabling vehicles to communicate with each other to form a Vehicular Area Network (VANET), is being pushed as a technology for future safety in vehicles. The high data rates and communication range capabilities of the technology lead us to investigate it as a viable solution for providing map updates. Our thesis work is first of its kind where real world mobility traces are used in study of VANET. These traces, obtained from GPS devices used for navigation in vehicles, are collected for all vehicles moving in and around the country of Netherlands for entire 24 hours. No existing network simulators have the capabilities to work with them due to the enormity of the traces. In this article we propose our own simulator DC-PIC, which is simple and scalable in nature. It is designed using concepts of divide and conquer and particle in cell, which are widely used in the field of high performance computing. The feasibility study of application of map updates involves investigation into the degree of support required from Road Side Unit (RSU). We analyze influence of various parameters of RSU on spreading updates to vehicles and based on which we propose values for these parameters suitable for our application. The empirical analysis of mobility traces is performed to provide insight into suitable locations for potential RSUs. Further, empirical analysis is conducted to obtain insight into mobility patterns and communication capabilities of vehicles in VANETs. Finally, we propose a simple and distributive dissemination algorithm named Adaptive and Distributive Broadcasting algorithm (ADB) which provides spread of updates comparable with that of flooding algorithm but without the unnecessary overhead of excessive redundant messages and collisions.