TCP Throughput for Vehicle-to-Vehicle Communications

It is well established that the relative movement of vehicles significantly affects the performance of the networking protocols in vehicle-to-vehicle communication. As TCP is the widely used end-to-end transport protocol, it is interesting to study its performance in such a highly mobile environment. In this paper, we consider a uni-directional motion of vehicles the velocities of which are gamma distributed. We present an analytical framework that forms the vehicle-to-vehicle network. With the help of a modified Bessel function of the second kind, we evaluate the cumulative distribution function of inter-vehicle distance. We also calculate the probability for catch-up and the expected time for it. The effect of round trip time and retransmission rate are studied which are then used to calculate the TCP throughput. Numerical results are presented that show high traffic density and diversity in velocity of the vehicles increase TCP throughput in a vehicle-to-vehicle network