No abstract

Traffic flow optimization within traffic networks has been approached through different kinds of methods. One of the methods is to reconfigure the traffic signal timing plan. However, dynamic characteristic of the traffic flow is not able to be resolved by the conventional traffic signal timing plan management. As a result, traffic congestion still remains as an unsolved problem. Thus, in this study, artificial intelligence algorithm has been introduced in the traffic signal timing plan to enable the traffic management systems' learning ability. Q-Learning algorithm acts as the learning mechanism for traffic light intersections to release itself from traffic congestions situation. Adjacent traffic light intersections will work independently and yet cooperate with each other to a common goal of ensuring the fluency of the traffic flows within traffic network. The experimental results show that the Q-Learning algorithm is able to learn from the dynamic traffic flow and optimized the traffic flow.

No abstract

One of the critical tasks in object tracking is the tracking of fast-moving object in random motion, especially in the field of machine vision applications. An approach towards the hybrid of particle filter (PF) and mean shift (MS) algorithm in visual tracking is proposed. In this proposed system, complete occlusion and random movement of object can be handled due to its ability in predicting the object location with adaptive motion model. In addition, the PF is capable to maintain multiple hypotheses to handle clutters in background and temporary failure. However PF requires a large number of particles to approximate the true posterior of the target dynamics. Therefore, MS algorithm is applied to the sampling process of the PF to move these particles in gradient ascent direction. Consequently a small sample size will be sufficient to represent the system dynamics accurately. The proposed approach is aimed to track the moving object in random directions under varying conditions with acceptable computational time.

Nowadays, vehicle tracking is a vital approach to assist and improve the road traffic control, surveillance and security systems by having the detail of the captured vehicle information. In past, many tracking techniques have been implemented and suffered from the well known 'occlusion' problems. Increasing the accuracy of the tracking algorithm has caused the computational cost due to the inflexibility to adapt the partial and fully occluded situations. Besides occlusion, appearance of new objects and background noises in the captured videos increase the difficulties of continuously tracking the labelled vehicles. In this paper, an adaptive particle filter approach has been proposed as the tracking algorithm to solve the vehicle occlusion problem. In order to solve the common particle filter degeneracy problem, the proposed particle filter is equipped with the adaptive resampling algorithm which is capable of dealing with various occlusion incidents. The experimental results show that enhancement of the particle filter via resampling algorithm has been robustly tracking the vehicles, and significantly improve the accuracy in tracking the occluded vehicles without compromising the processing time.