A Fuzzy Rule-Based Fog-Cloud for Control the Traffic Light Duration Based On-road Density

Abstract

A traffic light control system is important to reduce traffic jams. Several methods have been proposed to control traffic lights. However, most of them are inaccurate because do not use data on traffic density status. This study proposes an automatic traffic light control system by instilling artificial intelligence and Radio Frequency Identification (RFID) technology which is used to determine the best duration of traffic lights on an intersection. RFID is used to calculate the average speed of vehicles and the percentage of road occupancy in each lane. The average speed value and the percentage of road occupancy are used as inputs for the fuzzy rule-based algorithm. The outputs of the fuzzy rule-based are the status of traffic jams, road occupancy rate on each lane, the average speed of vehicles on each lane, and real time duration of traffic lights. The fuzzy computing process is carried out locally on the fog server via a Wi-Fi gateway to reduce cloud load. We evaluate the rule-based algorithm on an intersection with 4 lanes. The results show that the average speed of lane 1 is middle 0.922, lane 2 middle 0.699, lane 3 middle 0.599 and lane 4 middle 0.621. for fuzzification value of road density obtained lane 1 high 0.409, lane 2 low 0.475, lane 3 mid 0.951 and lane 4 mid 0.858. The conditions of traffic jams using the rule-based are as follows: "Heavy-Clock" for lane 1, "Light" for lane 2, "Light-Heavy" for line 3, and "Light-Heavy" for line 4. The system built-in using RFID technology can calculate average speeds and road occupancy rates accurately.