Research On Key Technologies Of Virtual Vehicle Behavior For Environmental Sensing In Internet Of Vehicles

With the development of network technology, Internet of Vehicles(IoV) will provide the user with more diversified and elaborated services.However, the service platform of IoV is hardly to sense precise data from road-network environment, which results in a reduced ability of service delivery. In order to sense the global road-network environment, we implement an accurate sense of road-network environment in an information space based on virtual vehicle (W), and mainly study technologies such as the optimal path learning, the VV coalition discovery,the coordination of VV coalitions and so on. Study yields the following results:1. We propose a W learning method based on ensemble neural networks to acquire related optimal route knowledge, which can be used to discover vehicles to sense route-network-environment information. The method first selects a fixed number of neural networks based on data characteristics. Then, before the new data arrive, the proposed partition method divides the ensemble network into three parts,retrainsthe three parts, and reintergrats them into a new ensemble neural network.Experimental results based on real data demonstrate that the proposed method performs better than other comparative methods.2. We propose a W coalition discovery method based on coalition game theory for gathering the regional road-network environment information. We first identify the objective of data gathering from road-network environment and define the payment sharing scheme. Then, we design a game strategy between VVs. Finally, using a coalition formation algorithm, VVs are able to form coalitions through interaction and accomplish data gathering tasks cooperatively. The simulation experiment results show that the proposed method can effectively discover VV coalitions.3. We propose a VV coordination method based on coordination control theory for gathering the global road-network environment information.Firstly, we identify multiple VV coalitions based on the spatial clustering algorithm. Then, we achieve the efficient collaboration among VVs by controlling the relationship between VVs based on coordination control theory. And finally, we conduct theory analysis of the coordinated control algorithm. The simulation experiment results show that the proposed method can realize the effective coordination among VV coalitions.