Non-Orthogonal Data Exchange in Vehicular Networks

Abstract of Research: Non-orthogonal Data Exchange in Vehicular Networks Aim This project aims to study non-orthogonal data exchange among vehicles. Specifically, the use of physical-layer network coding and multiuser detection techniques for data communications in vehicular ad-hoc networks (VANETS) will be investigated.

Project Background With the increasing popularity of inter-vehicle communications and the development of new

protocol standards, the roads Will likely be filled with high-speed wireless users in the near future. In the US, there have been initiatives to require newly manufactured vehicles to be equipped with wireless communication units. As the vehicular network gets crowded, the radio bandwidth and other signaling dimensions need to be allocated among multiple users more judiciously so that real-time data can be successfully delivered. However, due to the distributed and intermittently connected nature of vehicular networks, orthogonal domains (e.g., time, frequency) may not exist within short contact time between vehicular nodes. For time-critical applications, random access without any scheduling appears to be more suitable, but packet collisions that lead to wasted bandwidth are inevitable. Physical-layer network coding (PNC) and related multiuser techniques can exploit the interfered or overlapped signals based on receivers’ partial knowledge on the mixture. Hence, multiple users can access the same channel simultaneously and precious contact intervals can be fi.1lly utilized.

Prior studies on such techniques, however, did not take VANET scenarios and high mobility into account. VANET scenarios are governed by the road topologies, vehicular traffic and mobility patterns. On the other hand, high-speed vehicular motion can distort the mixed signals simultaneously transmitted by multiple users and introduce further difficulties in signal recovery. These challenges need to be overcome before the potentials of PNC and other related techniques can be realized in VANETS, and this project aims to solve the most fundamental issues related to these challenges.

The background and research experience of the PI and Co-I’s put them in a good position to carry out the proposed studies. The PI has solid research experience, both theoretically and practically, in VANETS and intelligent transportation systems. One of the Co-I’s and his group pioneered the concept of PNC, and they developed the first working prototype of a PNC system. Another Co-I is in charge of a campus-wide VANET testbed in Singapore.

Brief Proiect Descrigtian

Our preliminary study suggests that PNC systems could work well in highly-mobile environments in principle, but there are still many aspects that require further attention, and results based on analysis and simulation need to be fiirther validated by experiments.

To account for these outstanding issues, both theoretically and practically, we propose a three-year fiill-scale research plan that is structured into three inter-related topics: (1) we will investigate network-coded multiuser data exchange in random access vehicular networks based on PNC and related techniques; (2) we will propose relevant channel estimation and signal decoding schemes to combat the performance degradation in PNC due to vehicular motion; and (3) we will study experimentally the impact of various practical factors in vehicular environments and evaluate the performance of the proposed schemes for gaining a good balance between bit-error performance and timeliness.

Significance of the Profect

This research project, if successful, will be the first to demonstrate the use of non-orthogonal data communications in vehicular networks with PNC and related techniques, not only in theory but also in practice. The research outcomes are expected to open up a Whole new dimension in real-time vehicular networking for enhancing safety and efficiency in the future.