Dynamic Prediction Of The Road Status Based On Artificial Immune Network

The data of vehicles in the cities can be gathered due to the developing and popularization of GPS sensors and all kinds of wireless communicating network, while these data covers much information. The data itself of moving objects is very huge and the traditional data processing method will bring its localization as the dynamic updating of time, so it is too hard to dispose the dynamic data. How to deal with the data to make it forecasted and monitored will become the emphases of this article.The theory of artificial immune is an intelligent computing method come into being the elicitation of innate immune system. It has strong capability of learning and memory, which can exact and analyze different characteristic data. So the it gives a new idea of dealing with existing problems. This article will focus on using the theory of artificial immune to study the disposing problems of roaming object. The mainly tasks of this paper are as follows;(1) Raising a technique of predicting dynamic moving vehicles through artificial immune theory. The technique will carry out a model of the moving object from the aspect of moving tendency to recognize and predict some trails of unknown moving pattern when there's no function of the moving object.(2) It will bring problems such as sensitivity to the beginning input sets or trapping in partial best answer by using of analyzing the traditional K-means algorithm. In order to put forward immune K-means algorithm, we have tried making use of the global search and the immunological memory character of artificial immune theory into optimizing the K-means algorithm. The experimental results show that this algorithm can adapt dynamic data better than the traditional one.(3) Traditional grid clustering is sensitive to the distribution of data form and this paper brings out a new way of partitioning grid to optimize grid clustering by analyzing this problem. Not only can this method play a high efficiency of grid clustering, but also it solves the problem of sensitive to data distribution to a degree.