Research On A Cross-layer Routing Protocol In VANET Under Urban Transportation

In recent years,there is a diversified tendency on Internet of vehicles in the field of Intelligent Transport System (ITS) · It is of great significance in acquiring the information of vehicle basic states, charging and the accidents by communication between vehicles to vehicles. Researching on routing protocols of Vehicular Ad-hoc Network (VANET) has becoming an important part in the technology of Internet of vehicles. Due to frequent changes in network topology and the limited application circumstances in cities and towns,nodes in VANET can be more complicated than in MANET since the layout of streets and obstacles on the roadside. Therefore, researchers focus on the relatively reliable and efficient routing protocols.On basis of several classic routing protocols, such as AODV, DSDV, DSR, GPSR, there exists some issues of low PDR and large average end to end delay raised by single-layer construction and single-factor. According to these issues, a multi-factor cross-layer position-based routing protocol for VANET (MCLPR) is proposed to improve reliability and efficiency of the network, which obtains the information of MAC layer and some factors to achieve the promotion and optimization on algorithm,weighted by a variety of information above. MCLPR has the following characteristics: (1) Information of nodes, roads, link quality,SNIR and frame error rate (FER) are taken into consideration. (2) Algorithm for vehicles selection at intersections (AVSI) and Algorithm for vehicles selection of non-intersections(AVSNI) are proposed for strategies of vehicles selection whether they are at intersections or not. In AVSI , the position and speed of vehicles, density of intersections, link quality, the information of MAC layer are considered and weighted to acquire a best selection of next hop on the complex circumstances. In AVSNI,the mobility of vehicles has certain predictability,only position and direction are considered to select the next hop. (3) Since the dynamic vehicular mobility model, store-and-forward technology, position correcting mechanism and dynamic HELLO STREAM are added to MCLPR, which ensures the reliability of data transformation and reduces negative impact on the performance of the protocol.During the simulation,a special real map is obtained by a series of operations,which is regarded as the abstraction of the mobility named Manhattan Grid with a tool of Bonnmotion,then import the mobility to the NS3 simulator under MCLPR, AODV, DSDV, DSR and GPSR.Packet deliver radio (PDR) and average end to end delay are focused among these protocols by changing the density of network,the speed of vehicles and the radius of transmission.After simulation it can be concluded that MCLPR shows better performance than others on PDR under the same condition in a certain area. In larger density of network, MCLPR show lower end to end delay than other protocols. MCLPR shows the better PDR and the lower end to end delay when the radius of transmission become larger.