Research On Routing Protocols Based On OLSR In UAV Scene

With its high flexibility and convenience,drones have frequently appeared in human social activities in recent years.Compared with a single drone that does not communicate between nodes,the clustering of drones must be the general trend.Considering factors such as terrain and climate,it may not be possible to establish ground control stations in some areas,so the centerless UAV ad hoc network is of great research value.Due to the existence of the routing protocol in the UAV ad hoc network,the UAV nodes can carry out data interaction and complete tasks together,which greatly improves work efficiency.As a typical routing protocol in ad hoc networks,OLSR has the excellent characteristics of low latency and simple networking,which is more suitable for the high latency requirements of UAV scene.Based on the research status of the OLSR protocol in the existing literature,two improved OLSR routing protocols are proposed for addressing the needs of the UAV ad hoc network in this thesis.The main innovations accomplished include the following two aspects:1.Aiming at the problems that the MPR selection mechanism of the OLSR protocol is unreasonable because of lacking of the consideration of the mobility of the UAV node and network update untimely or excessive control overhead due to a fixed sending period of TC message in the existing literature,an improved OLSR routing protocol(Dynamic Sensor Optimized Link State Routing Based on Q-learning Algorithm,DSQ-OLSR)is proposed,which has the function of dynamically sensing network topology and based on the Q-learning algorithm.Firstly,based on the reasonable evaluation of link performance,the protocol defines two variables: link stability and link lifetime,which are used to optimize the selection of MPR nodes,and combines the connectivity of one-hop neighbors to propose a three-index-based MPR selection mechanism.This mechanism improves the rationality of MPR selection and the stability of topology;Secondly,the proposed protocol traverses the one hop neighbor table and takes the minimum link lifetime as the sending interval of TC messages to realize the timely perception and diffusion of network topology;Finally,the protocol takes the nodes in the network as the state,and takes the data transmission to the next hop node as the action,and then initializes the reward function with the obtained network topology.Based on the above operation,the proposed protocol optimizes the routing calculation process of OLSR protocol through the iterative process of Q-learning algorithm.Simulation results show that dsq-olsr protocol improves the success rate and throughput.Simulation results show that DSQ-OLSR protocol has improved performances such as success rate and throughput.2.Aiming at the problem of single routing index and single reachable route of OLSR protocol in the existing literature,a load-aware multi-index and multi-path routing protocol(Optimized Link State Routing Protocol based on Multi-indicator and MultiPath,MIMP-OLSR)is proposed.Firstly,to increase the rationality of routing,the protocol fully considers the node load and other levels,and uses the number of hops,the blocking degree of MAC layer,the change rate of one-hop neighbor and the number of MPR-S nodes as routing metrics when calculating routing.Secondly,in order to enable the nodes in the network to obtain the above routing metric information and then execute the routing algorithm in a distributed manner,a metric notification mechanism is proposed,which modifies the HELLO and TC message packet formats and use the interactive process of control messages to realize the network-wide notification of indicator information;Finally,based on the above routing metric,a multi-path routing scheme is proposed,which can calculate two(if any)routes actually used for data interaction,and solve the problem of single reachable route in OLSR protocol.The simulation results show that MIMP-OLSR protocol improves the performance of packet delivery success rate,throughput and end-to-end delay.