engleski

Intelligent Traffic Control in Urban Areas

Control systems responding to the current traffic situation by adapting its parameters enable significant benefits. However, numerous limitations exist such as the need for accurate traffic models, the uncertainty in predicting future traffic flows, the difficulty in arrival time estimation, and the lack of a self-adjusting mechanism. Difficulties in optimising the signal control strategy and the importance of finding a solution to this problem resulted in different approaches. Methods from artificial intelligence have emerged as one possible solution. Such methods have the ability to accumulate and use knowledge, set a problem, learn, process, conclude, solve the problem and exchange knowledge. In this paper, the application of the Q-learning algorithm to control an independent intersection and an on-ramp on an urban motorway is described. Implemented approach enables an on-line adaptation of the control parameters to the current traffic situation reducing the delay in traffic.

urban traffic control; artificial intelligence; Q learning; ramp metering; independent intersections

nije evidentirano