Research On Terminal Risk Probability Prediction Method Based On Fog Computing In Internet Of Vehicles

As the number of vehicles in China has reached 327 million so far,so many vehicles make traffic safety issues more and more serious and pose a growing threat to human life.In order to improve the road safety situation,it is necessary to predict the vehicle accident risk for introducing the intelligent transportation system and the vehicle safety assisted driving technology.However,the current vehicle accident risk prediction method has the problem that the service response delay is too high and the prediction accuracy is low.Therefore,this paper proposes a terminal risk probability prediction method based on fog computing in Internet of Vehicles.The specific research contents are as follows:First of all,considering the problem that the offloading ability of the edge computing network is easy to reduce the service quality of the hotspot area due to the limited computing resources,this paper proposes a computing resource-efficient task offloading algorithm with load balancing based on fog computing.The algorithm firstly models the task assignment in fog computing network as an optimization problem of service response time and the computing resource occupation of fog computing data center meeting the limit of task deadline and network resource.Then,when a vehicle terminal proposes an offloading task request,the algorithm estimates its computing resource requirement.For its computing resource requirement,the scheduler preferentially offloads the task to the data center with the most computational resources which are free in the fog computing network.If the computing resources occupation of all fog computing data center is too high,the task will be offloaded to the cloud computing data center.Finally,simulation experiments show that the computing resource-efficient task offloading algorithm with load balancing can ensure that the proportion of fail responses to request tasks in the computing system is only 2.74% and the delay of response to request tasks is 0.47 seconds faster than the average task response time of classical fog computing method.In addition,considering the problem that the efficiency and accuracy of big data real-time processing in multiple formats generated by vehicle terminals cannot meet the requirements of vehicle safety assisted driving,this paper proposes the trichotomy Adaboost with Synthetic Minority Oversampling Technique and One-Hot encoding(AdaBoost-SO)algorithm to attain vehicle accident risk prediction model.In our work,predicting accident risk is mainly based on use of Big Data mining and analysis of real-life accidents data.Firstly,the experimental dataset is reconstructed by Synthetic Minority Oversampling Technique.We complement missing data and encode each sample feature to One-Hot code.Secondly,the trichotomy AdaBoost algorithm is respectively used to train a series of weak classifiers from the experimental dataset and then combine them into a strong classifier to get the prediction model.Finally,extensive simulation results illustrate that using the prediction model by trichotomy AdaBoost-SO algorithm can take the area under curve of 0.77 and real-time into account.