Research Based On Of MCT Prediction Algorithms In VC

With the development of cloud computing, people propose the new way to use it-add loud computing into the traffic system. Since the Auto PC was installed in the car, the development of VANET becomes inevitable. But because the car moves so fast, the car cloud computing can’t allot the tasks like the cloud computing or mobile cloud computing. In the environment of cloud computing, comparing with the mode online, batch processing can allot more reasonably the tasks. But in the environment of the car cloud computing, to wait for the allocation of batch processing wastes the precious time for connection among the cars. But, in the online mode, the algorithm of minimal time with a good effect for finishing the allocation can’t find the best resource to allot the tasks in the traffic flow.In this situation, the algorithm for predicting the allocation in online mode is born at the right time. According to the driving habit, it counts the turning possibility of cars at every intersection to forecast the driving direction and speed, even the communication capability of resource agent. If the prediction is precise enough, the allocation of online mode can allot reasonably the resource agent like the allocation in cloud computing.The principal works of this thesis:●Basing on the model of Manhattan, this thesis will present dynamically the turning possibility of cars at every intersection according to their driving habit.●After a simulation experiment who runs for enough time, the turning possibility of cars will come out. According to this possibility, we can forecast the their driving ways.●Estimate the connection time entre the car terminal and resource agent according to the driving ways.●Finish the allocation of tasks in the minimal time and observe the effect of allocation in this algorithm by simulation experiment.