The traffic condition likelihood extraction using incomplete observation in distributed traffic loop detectors

Here, for extracting a Gaussian Mixture Model of traffic flow in a certain freeway, some distributed EM estimation, including the consensus and particle filter, has been presented, these approaches converge at a linear rate to a stationary point of the log likelihood function, which is usually a local maximum, more rapidly than standard EM. With reasonable assumptions, it was shown that DEMs communication requirements are quite modest. These distributed filtering algorithms only need information exchanges between neighbor sensor nodes. The global information can be diffused over the entire network through the local information exchanges. These are scalable because the adding of more nodes does not affect the algorithms´ performance. It is also robust as it can still produce the right results even if failures of some nodes occur. Currently, the number of Gaussian components is given. We can also use a distributed algorithm to estimate this number. A well-fitted approach to the estimate of this number is the one proposed in [12].The simulation tests justify the performance of distributed MLE in Intelligent Transportation Systems.