Using fuzzy logic to calculate the Backoff Interval for contention-based vehicular networks

In contention-based wireless networks, packet collisions are considered the main source of data loss. Retransmission of the lost packet is done several times until an acknowledgment of successful reception (ACK) is received or the maximum number of retries is reached. The retransmission delay is drawn randomly from an interval, called the Backoff Interval. A good choice the backoff interval reduces the number of collisions and, therefore, increases the throughput and decreases the energy consumed in retransmissions. In this paper, we propose a backoff scheme based on fuzzy logic. The new scheme depends on locally measured data to estimate the backoff interval which supports the distributed nature of the vehicular networks. We present several versions of the Fuzzy Backoff scheme and compare them with other known schemes: the binary exponential backoff (BEB), and an optimal scheme which requires the knowledge of the total number of nodes in the network. We used throughput, fairness, and energy consumption as performance measures for evaluation. Results show an improvement of the fuzzy-based schemes compared to the BEB, and approaching the optimal results.