Research On Transportation-oriented Resource Scheduling Strategy In Internet-of-vehicles

When 5G communication technology is evoked to the Internet-of-Vehicles(Io V),heterogeneous wireless communication technologies become available(including Dedicated Short-Range Communication,Cellular based Vehicle-to-Everything,and Millimeter Wave).The computing-thirsty applications undergo exponential growth(cooperative driving,automatic driving,alert delivery,etc.).However,most literatures are dedicated to the wireless resources assignment for the Io V applications without leveraging the communication and computing resources to improve the transportation performance.The off-the-shelf Io V architecture also ignores the characteristics of the road transportation performances.It is therefore difficult to handle the scenarios with certain road transportation demands.The emerged 5G-based Io V technology is not only to provide the service for communication applications but also to guarantee the resource demands of road transportation performance,such as driving safety,road traffic throughput,and vehicle string stability.It is essential to joint resource optimization for the Io V and road transportation performance,which requires to clarify the mechanism of Io V resources(communication and computing resources)with road transportation performance.However,due to the constraints of on-board computing capacity and the long distance to the cloud platform,the traditional cloud computing may not meet the delay requirement of transportation systems.Fortunately,the edge computing technology that deploys computing resources close to the users,which could efficiently finish the computing offloading in real-time.This new computing pattern is designed for diverse demands.However,the inherent problems of the transportation system,such as the time-varying demands and the highly dynamic road traffic,which also arouses great challenges to apply edge computing in the transportation system.To address the above challenges,we proposes a vehicular fog structure,which provides intelligent access decisions at the edge.This dissertation delicates to the combination of artificial intelligence technology and vehicular edge computing into the joint optimization of communication and computing resource taking advantage of network slicing technology,so as to serve the diverse resource and transportation demands.The remainder of the dissertation is organized as follows: 1)Io V resources-aided road transportation performance;2)communication and computing resource optimization in the heterogeneous Io V;3)the platoon-based dynamic resource scheduling;4)the network slicing orchestration strategy for transportation system.The first part reveals the interaction between driving safety,road traffic throughput,vehicle string stability,as well as the Io V resources.Above all,investigating the vehicle following scenario,the quantitative function of safety distance with Io V resources is obtained.According to the safety distance function,we attain the relation of Io V resources with road traffic throughput.To build the connection between the vehicle string stability and Io V resources,this dissertation first reveals the function of the safety distance on the vehicle string stability through the vehicle flow mechanics.Hereafter,leveraging the safety distance,the vehicle string stability is determined by Io V resources.To meet diverse resource demands,a novel vehicle fog computing is proposed with the hierarchical computing pattern.The maldistributed application demands could be digested by using the differentiated computing resources from platoons and base stations.In the second part,to reduce the offloading failure probability and the Io V resource cost,we propose an intelligent task offloading framework in heterogeneous vehicular networks with three Vehicle-to-Everything(V2X)communication technologies.Based on stochastic network calculus,this dissertation firstly derives the delay upper bounds of different offloading technologies with certain failure probabilities.The delay upper bound and failure probability fully concern the impact of communication transmission and on-board computing capacity.Moreover,we propose a federated Q-learning method that optimally utilizes the available resources to minimize the communication/computing budgets and the offloading failure probabilities.Simulation results indicate that our proposed algorithm can significantly outperform the existing algorithms in terms of resource cost and offloading failure probability.Meanwhile,the simulation represents that the CV2 V communication achieves the best road traffic throughput compared to other V2 V communications.The third part delicates to propose the platoon-based edge computing strategy in high mobility scenario.Since the communication outage probabilities are different with channels,we select the channel with the least outage probability to form the platoon.A low complexity fair candidate selection strategy is proposed to select the candidate members and specify the amount of computing resources provided by the platoon.However,the instability of vehicles makes the platoon difficult to complete the task offloading.To address this problem,We next attain the upper bound delay of Vehicle-to-Vehicle(V2V)offloading based on network calculus theory.Hereafter,an efficient sleeping multi-armed bandit tree-based algorithm is proposed to realize the resource assignment.Compared with existing algorithms,the proposed offloading strategy improves the service reception rate and reduces the system overhead with low complexity.In the fourth part,according to the Vehicular Fog architecture,a smart slice scheduling scheme is proposed for the collaborative optimization of the Io V resource.This dissertation first explores the joint optimization of driving safety,vehicle string stability,and road traffic throughput by leveraging on the consensus Alternating Directions Method of Multipliers algorithm(ADMM).Then,the communication and computing demands of the optimal transportation system are captured.However,available fog resources as well as network traffic are all dynamic and unpredictable due to high mobility of vehicles,which results in weak resource utilization.Accordingly,an intelligent algorithm for network slices is proposed based on the Monte Carlo Tree Search(MCTS)in terms of a new metric Cross Entropy(CE),which is able to allocate the resource allocation for the match of traffic load in the time-space domain.This slice scheduling algorithm does not require any prior knowledge of the network traffic.In the high network load,resource rich vehicles can be aggregated into a platoon providing computing resources to reduce the heavy load of edge infrastructure.Simulation results indicate that the proposed algorithm outperforms several baselines in terms of throughput and delay..