A Cloud-Interfaced Social Ski Driver Optimized routing and Charge scheduling for EV using deep RNN

The rejuvenation in the field of Electric Vehicles (EVs) has earned enormous interest among various researchers as the EVs are entitled to a remarkable mode of transportation in coming days. EVs are vital to conserve the conventional fuel, but EVs poses restrictions due to inadequate batteries leading to short distance travel and when battery charge nears at critical it becomes imperative to be aware of nearby charging stations (CS) and also there exist lesser charging strategies. Hence, drawing out two challenges, 1) Optimised routing to CS and 2) Charge scheduling mechanism at CS. Hence, A cloud-interfaced Vehicular Adhoc Network (VANET) model which enables communication between the vehicle traffic server, Road Side Units (RSUs) and various EVs on roads is used along with a meta-heuristic Social Ski Driver (SSD) algorithm as a new optimal routing for EVs to CS and further, charge scheduling is employed using dissimilar power configurations at CS for emergency and normal EVs. Hence, an optimal routing decision for EVs are made based on developed fitness function considering battery power, distance and traffic density computation using deep Recurrent Neural Network (RNN). First, cloud-interfaced routing decision is computed using SSD routing algorithm, Secondly charge scheduling mechanism is employed for EVs at CS using multi-objectives such as priority, delay and response time. The performance of the proposed Optimized Social Ski driver (O-SSD) is evaluated and compared with Particle Swarm Optimization (PSO). The devised scheme outperforms in all measures like traffic density, delay, success rate of allocation during (On-peak and Off-peak hours), total trip time and minimum battery power consumption with the values of 8.3 per lane, 7.80ms, (78% and 96%), 24.85min and 12486 joules respectively.