Fuzzy Logic Aggregation of Wireless Sensor Network Data for Smart Traffic Light Control

Abstract

The steady increase in the number of vehicles on the road has increased traffic congestion in most urban cities. Inadequate space and funds for the construction of new roads has prompted scholars to investigate other solutions to traffic congestion. Most popular approach in use is the traffic light system. Adopted strategies include; static and vehicle actuated (VA) lights. However, these models have limitations. The static model does not take into account the non-uniform and dynamic nature of traffic because they do not operate with real time data. VA lights were an attempt to improve the static model; they activate a change in light signal when cars are present. However, they only detect presence of cars; they do not count number of cars. To help solve these weaknesses, this research presented a novel approach to traffic routing. Our approach uses smart traffic control systems (STCS) to make traffic routing decisions. STCS use real time data and mimic human reasoning thus prove promising in vehicle traffic control. We present a smart traffic light controller using fuzzy logic and wireless sensor network (WSN). The approach is designed for an isolated four way roundabout. It employed fuzzy logic to control the lights and determine how the green light will be assigned for each approach. The WSN collected the traffic data in real time. This data is aggregated and fed into a fuzzy logic controller (FLC) in form of two inputs – traffic quantity (TQ) and waiting time (WT) for each approach. Based on the inputs, the FLC then computes an output priority degree (PD) that controls green light assignment. Using the PD, an algorithm is formulated that assigns green light to the lane with highest PD. The cycle continues until all approaches get green. Given the practical nature of the thesis, applied research was the core methodology used. This research design made it possible to gather necessary data, analyze it and develop a solution to address the weaknesses of the current traffic light controllers. To test and analyze the approach, we designed and simulated a model of a traffic light in Java. This provided a virtual representation of the proposed approach. The test bed provided 94.7% accuracy. The results further demonstrated that; a smart traffic light controller has better performance than a static one; when it comes to reducing average WT at intersections. Our approach recorded an average WT of 2.985 minutes while the fixed had an average WT of 8.955 minutes. It also showed that STCS can fairly manage traffic at four-way intersections. This is proven by the fact that our average WT was 2.985 minutes against the 6 minute maximum WT limit; and the average PD was 5.091, against a PD range of 2.500 and 7.500. Lastly the results indicate that a smart traffic light controller can effectively replace an experienced traffic officer managing traffic at a roundabout.