Research On Intelligent Vehicle Fault Detection Mechanism

Vehicles play a vital role in people’s lives,which in high-intensity overload operation may cause unexpected problems.Once vehicle faults occur,they will cause vehicle damage and affect driving.In addition,they affect the normal maintenance of traffic condition,disturb the city’s traffic order,cause traffic jam and accident.Real-time fault detection of the vehicle can be effective avoids potential dangers.Therefore,how to efficiently collect vehicle fault information,so as to guide vehicle safety maintenance and road traffic safety work is very important.This thesis mainly pursues research from two aspects: Big Data processing and Vehicular cloud computing(VCC)resource allocation.On the one hand,how to apply limited local computing power and storage capacity for real-time and efficient fault detection is an urgent problem.On the other hand,other delay-insensitive requests are offloaded to the cloud platform after maximizing the utilization of local resources.After waiting a while time,the vehicles can download the processed test results provided by cloud platform.The two methods are combined online and offline to maximize the fault detection rate.Aiming at the online intelligent detection algorithm based on the measured data,we use the vehicle fault data collected by the on-board sensors to conduct in-depth research,and divide them into four types which are braking system faults,instrument communication faults,lithium battery system faults and other faults.A two-step fault detection method is proposed to monitor the state of the vehicles.First,we apply ReliefF based feature selection algorithm to determine the best number of features.Then,combining the best number of features with BP neural network,a sliding window-based method is proposed to evaluate the vehicle status and give early warning in real time.Finally,a comprehensive simulation is conducted,which indicates that our proposed detection method can meet the requirements of real-time detection.Aiming at the problem of vehicle fault detection under the VCC,we design the maximum system reward based vehicle fault detection mechanism(Mobile Edge Caching based Resource Scheduling,MECRS).The faults,which cannot be processed locally by the vehicle,can be uploaded to the cloud to process using the powerful computing power.The requests of the cloud platform are divided into four categories: vehicle fault requests,network operation and maintenance requests,emergency requests,and large-scale file requests.We establish the target function based on the maximum system reward and solve it by using the joint optimization algorithm.The simulation results show that this mechanism can improve the service ratio of the vehicle fault business on the premise of ensuring the smooth service of other requests.The mechanism proposed in this thesis can ensure the timely treatment of vehicle faults by combining online and offline methods.At the same time,the research of this article can continue to improve.In terms of online intelligent detection algorithm,multi-fault early warning function can be considered.In terms of offline cloud resource allocation algorithm,more factors can be considered when unloading the fault requests.In addition,the measured data can be imported for simulation,which will be a very meaningful work.