Controlling Time of Traffic Lights Automatically and Adaptively

Controlling of traffic lights at road intersection will be optimum, if the adjustment is appropriate to level of vehicles queue. The densest queue of vehicles requires green time longer than less dense queue of vehicles. A new algorithm referred to Webster method is proposed to count the green, red, and humber time at traffic light automatically and adaptively, by using to predict number of vehicles from previous research. The algorithm that was resulted by composite between Radial Basis Function Network (RBFN) methods with K-Means cluster was used to predict number of vehicles. The resulting of simulation shows that system could predict number of vehicles in real time by total error 0,053977. Webster method is generally used to calculate green, red, and humber time, but it is constant, while level of dense of vehicles queue is not constant. This paper uses Webster method to calculate time of traffic lights, and uses actual vehicles flow that is predicted by composite as main parameter to count green time. The humber time is determined by speed of vehicles and distance between stopping line with target line. The speed is obtained by implementation of XOR and subtraction method into speed algorithm. So, green and humber time is always change appropriate with levels of vehicles density. Effectiveness of new algorithm and system is simulated by software. The system can determine maximum green time if the condition of vehicles flow is dense, and determine minimum green time if the condition of vehicles flow is quiet