Real-time estimation of neighbouring node density in chaotic VANET environments

Submission note: A thesis submitted in total fulfilment of the requirements for the degree of Doctor of Philosophy to the School of Engineering and Mathematical Sciences, College of Science, Health and Engineering, La Trobe University, Bundoora.

Node density estimation can have a strong influence on improving network throughput. It can be used as a parameter to predict the congestion in the network. Accurately estimating node density in vehicular ad hoc networks (VANETs) is a challenging and crucial task. Various approaches exist, mainly based on the reception of beacon messages, yet none are able to deliver an accurate estimation in all situations, especially in highly saturated networks and in the presence of channel congestion. In this thesis, we take a step towards improving the awareness of the communicating vehicles about their surrounding node density especially in the presence of collision and detect/foresee possible channel congestion and the so-called broadcast storm problem based on the number of neighbouring nodes. We begin by analysing the Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) protocol with exponential back-off, as the most popular media access control (MAC) layer protocol in the IEEE 802.11x family. We calculate the exact probability of collision-free transmission when a different number of nodes contend for the channel, both mathematically and through a Markov model. The results are used to define channel congestion and the broadcast storm problem based on the number of nodes that contend for the channel. We then calculate the relationship between the number of nodes that contend for the shared wireless channel and the expected number of successful nodes that gain access to the channel and transmit simultaneously. The calculation is first done mathematically and then via a discrete event simulator and the results are presented in the form of an equation. We also evaluate the relationship between the number of simultaneously transmitting signals (via different nodes) and the received signal strength (RSS). For this evaluation, we first implement stochastic channel models for different (highway and urban) environments and develop realistic platforms to simulate vehicle-to-vehicle (V2V) communication for each environment. We then use the simulation results to evaluate the relationship between the average value of the multiple consecutive sampled RSSs and the number of simultaneously transmitting signals separately for highway and urban environments. The results are also presented in the form of an equation. We then create separate look-up tables showing the relationship between the three parameters of RSS, the number of simultaneously transmitting signals and the number of active nodes in the network that contend for the channel, for highways and urban environments. After finding the relationship between the three mentioned parameters, we propose a novel adaptive framework that allows individual nodes to estimate the node density of their surrounding network. Our proposed solution is based on sampling real-time RSSs and using the averaged values in the relevant look-up tables, based on the current environment of the estimating node (vehicle). Unlike other node density estimation methods, our proposed solution takes advantage of the chaotic channel condition and estimates node density by retrieving information from collisions. Our proposal adapts to different environments and can address rapid changes in the configuration of the network and the properties of the channel in VANETs. Our experimental results indicate that the method yields reliable estimation in an acceptable time frame. Finally, we present how the estimated node density can be used to detect/predict possible channel congestion and the broadcast storm problem to be reported to higher layers with a request for an adaptive collision avoidance strategy.