Research On Ambulance Dynamic Route Planning For MEC Collaborative Service

The rapid development of 5G communication,MEC,IoT and other technologies has brought new opportunities to pre-hospital intelligent first aid.The deployment of a large number of intelligent diagnosis,high-definition video and other equipment has put forward more stringent requirements for medical network bandwidth,delay,storage,etc.With the development of edge computing,the tasks of edge nodes and cloud computing "upload and download"to network edge are considered as key technologies for the future development of the Internet of Things and 5G.The deployment of medical network in suburban areas will have limited network resources,sparse and uneven distribution of communication resources,and differences in computing resources,which will easily make scheduling difficult.How to realize efficient and reliable task offloading in resource-constrained environment is an urgent problem to be solved in suburban medical network.Therefore,this paper studies the computing offloading scheme and the reasonable allocation of limited MEC resources,and proposes a two-stage computing offloading strategy combining task grouping and resource allocation.After solving the problems on the medical network,it is of great significance to improve the survival probability of patients by providing timely and effective treatment for patients on the ambulance in pre-hospital emergency.Therefore,aiming at the problems of empty network resources,unbalanced deployment of network resources,and the road conditions,network status,patient’s condition and other information on the route of ambulances in the urban areas,we can make real-time and dynamic decisions on the route of ambulances,and solve the problems of dynamic real-time path selection and real-time and efficient data transmission between hospitals and ambulances.The main research contents are as follows:(1)Medical emergency network is facing the challenge of resource constraints such as energy,computing power and capacity in the deployment of resource-constrained Internet of Things in suburban areas.To ensure the quality of service,a two-stage computing offloading strategy combining with task clustering and resource allocation(TSCOS)was proposed.Jointly optimizing offloading decision and resource allocation scheme to reduce delay and improve resource utilization rate.Dentally,a grouping model based on task preference was designed to improve the accuracy and matching time during offloading phase,through pre-matching the offloaded tasks with the candidate servers.Then,an improved Gale Shapley method was proposed to realize the load balance of edge network by quickly calculating the optimal solution set of many-to-many game matches.Compared with distance grouping model,the task preference grouping model can improve the utilization of computing resources.To evaluate the performance of the proposed mechanism,the TSCOS was compared with random walk strategy(RS),greedy strategy(GE)and dynamic resource scheduling strategy(FFS+IPFS).Comparative experiments showed that the edge server load balancing variance reduced by 2-4 times,and the task completion success rate is improved 5%-15%.Under the same energy consumption,the energy-efficient is improved 5%-10%.Moreover,the average acceptance rate of offloading tasks and the average success rate are 92%and 96%,respectively.(2)With the development of 5G and real-time transmission technology,doctors can carry out remote consultations for patients in ambulance equipped with intelligent devices under the support of 5 G technology.In order to ensure the continuous operation of remote consultation and the rapid arrival of ambulances,the time-consuming and network quality of route must be both considered in path planning.In this paper,we proposed a dynamic path planning model based on network and time-consuming constraints(DPMNTC)for ambulances.To obtain the optimal path,firstly,use Time Series Decomposition Long Short Term Memory(TSD-LSTM)network to predict the time of the passing path and establish network quality function.Secondly,combine with the above model and the dynamic ant colony optimization algorithm(DACO)based on pheromone gain and dynamic vitalization factor proposed in this paper,to realize the path planning of joint the time-consuming and the network coverage on the path under dynamic road condition.Finally,the simulation results show that,compared with heuristic algorithm,DACO algorithm can find a more accurate solution with the fastest convergence speed.In addition,compared with the other two path planning,the DPMNTC can satisfy the needs of pre-hospital emergency care in ambulance and maintain real-time data transmission between hospital and ambulance,while ensuring the ambulance can arrive the hospital within a time threshold.