ERNSS-MCC: Efficient relay node selection scheme for mission critical communication using machine learning in VANET

In recent days, vehicular ad hoc networks (VANET) is advanced in communication paradigm, safety, and intelligent-based technique for message dissemination with the help of an intelligent transportation system (ITS). The VANET's major challenge is retaining network stability while providing communication in emergency and disaster situations where the vehicular nodes are highly dynamic. Mission-critical communication (MCC), which uses relay-assisted VANET to address this issue, is a key component of the ITS. Nevertheless, one of the most pervasive issues in VANET is selecting the relay vehicle in the unpredictable vehicular environment with a large node density in MCZ. The proposed temporary database learning and control (t-DLC) unit is attached to a road-side unit (RSU) and is triggered during an emergency to minimize communication overhead in the medium. The proposed system uses GPS to initially identify the location coordinates of each vehicle in the MCZ. Moreover, clusters are formed in MCZ using a modified K-means algorithm based on zone patterns by hard and soft cluster formation. For efficient relay selection, an adaptive neuro-fuzzy inference system (ANFIS) is used which takes fuzzy parameters as input. The vehicle with minimum root means square error (RMSE) is chosen as a relay vehicle. Experimental results prove that the proposed methodology significantly improves the packet delivery ratio (PDR) and throughput while minimizing delay and computational complexity compared with the existing systems.