Performance evaluation of vertical handover in Internet of Vehicles

Internet of Vehicles (IoV) is developed by integrating the intelligent transportation system (ITS) and the Internet of Things (IoT). The goal of IoV is to allow vehicles to communicate with other vehicles, humans, pedestrians, roadside units, and other infrastructures. Two potential technologies of V2X communication are dedicated short-range communication (DSRC) and cellular network technologies. Each of these has its benefits and limitations. DSRC has low latency but it limits coverage area and lacks spectrum availability. Whereas 4G LTE offers high bandwidth, wider cell coverage range, but the drawback is its high transmission time intervals. 5G offers enormous benefits to the present wireless communication technology by providing higher data rates and very low latencies for transmissions but is prone to blockages because of its inability to penetrate through the objects. Hence, considering the above issues, single technology will not fully accommodate the V2X requirements which subsequently jeopardize the effectiveness of safety applications. Therefore, for efficient V2X communication, it is required to interwork with DSRC and cellular network technologies. One open research challenge that has gained the attention of the research community over the past few years is the appropriate selection of networks for handover in a heterogeneous IoV environment. Existing solutions have addressed the issues related to handover and network selection but they have failed to address the need for handover while selecting the network. Previous studies have only mentioned that the network is being selected directly for handover or it was connected to the available radio access. Due to this, the occurrence of handover had to take place frequently. Hence, in this research, the integration of DSRC, LTE, and mmWave 5G is incorporated with handover decision, network selection, and routing algorithms. The handover decision is to ensure whether there is a need for vertical handover by using a dynamic Q-learning algorithm. Then, the network selection is based on a fuzzy-convolution neural network that creates fuzzy rules from signal strength, distance, vehicle density, data type, and line of sight. V2V chain routing is proposed to select V2V pairs using a jellyfish optimization algorithm that takes into account the channel, vehicle characteristics, and transmission metrics. This system is developed in an OMNeT++ simulator and the performances are evaluated in terms of mean handover, handover failure, mean throughput, delay, and packet loss.