| The Internet of vehicles has taken spatial location as the basis of environmental perception,provides information support for application scenarios such as intelligent driving,intelligent transportation and onboard information services,and realizes the development of intelligent networking of people,vehicles and things.With the construction of aerial and terristrial integrated network,the location parameter estimation technology based on wireless signal is gradually evolving towards multisource,high efficiency and intelligence.Using wireless signals to estimate the position of vehicles in the blind area of satellite coverage has become a research hotspot in academia and industry.In the process of vehicle network location estimation,there are some problems,such as serious multipath interference,dynamic changes of vehicle network,unstable number of anchor nodes and limited energy consumption and resources.Aiming at reducing the measurement error,improving the estimation quality,enhancing the system integrity and optimizing the energy consumption,this thesis has studied the measurement error correction in the non line of sight environment in the vehicle network,the selection of cooperative nodes in the vehicle ad hoc network,the vehicle/vehicle road multi-source cooperative positioning in the vehicle network and the power allocation strategy.1.Aiming at the serious non line of sight multipath interference and large measurement error in urban canyon areas,an error correction model based on deep deterministic policy gradient is proposed.Firstly,the interaction model between agent and environment is established,the action space complexity of continuous correction is reduced by using deterministic strategy,and the non line of sight error correction of the measured value of vehicle road network is realized by quadruple mutual feedback network.The simulation results show that the error of non line of sight measurement can be reduced by 41.67%after the correction of the model proposed in this paper.2.Aiming at the problems of unstable number of anchor nodes and low accuracy and integrity of position estimation of single measured value during vehicle driving,a two-layer spatio-temporal self consistent sequential confidence cooperation fusion algorithm for vehicle networking is proposed.The outer layer algorithm is used to quickly and roughly locate the fusion measurement data of vehicle and vehicle road,so as to reduce the calculation space.The spatiotemporal self consistent model is constructed through the inner layer algorithm,and the coarse positioning results are modified by sequential updating to obtain high-precision position estimation results.Simulation results show that when the number of anchor nodes is unstable,the algorithm can still maintain high positioning accuracy and integrity,and the accuracy can be improved by 17.8%compared with the traditional cooperative algorithm.3.In view of the dynamic change of vehicle ad hoc network structure,high uncertainty and difficult to accurately select the optimal cooperative node,a chain adaptive node selection method based on information entropy is proposed.The multi-resolution analysis of vehicle measurements is carried out to peel off the peak noise caused by channel disturbance caused by relative motion.Using the horse racing model,the upper limit of node selection ability is deduced theoretically.A node selection algorithm which can adaptively adjust the environmental change through historical data is proposed.Compared with the distance based node selection method,the positioning accuracy can be improved by 12%.4.Aiming at the resource limitation represented by power in vehicle network,a power allocation and dynamic control strategy based on game theory for vehicle ad hoc network and vehicle network is proposed.Firstly,the power loss fuction of vehicle to vehicle cooperative positioning based on the lower bound of square error is constructed,the power allocation strategy is abstracted as a mathematical problem of ordering the minimization of cost function under constraints,and the Nash equilibrium solution algorithm of cooperative competition game is proposed.On the other hand,the problem of energy consumption allocation in vehicle road cooperation is transformed into the control process of dynamic game,and a dynamic game algorithm of user channel power in vehicle road cooperation is proposed.Simulation results show that the proposed power allocation strategy can reduce the single point power consumption by 22.58%under the same positioning performance. |
