An improved hybrid channel Estimation technique for vehicle to Vehicle communication

Abstract

The Vehicle to Vehicle (V2V) is a wireless technique which offers connectivity between vehicles moving in a highway. It also offers infotainment and safety services like music, streaming videos, internet access, collision warning, speed warning etc. The high speed vehicular movements in V2V scenario have turned the channel estimation into a challenging task. The dynamic nature of the V2V has contributed towards Doppler shift that seriously affected the estimation of the channel. Hence, there is a need for robust models to estimate the channel. The Pilot Symbol Aided Channel Estimation (PSACE) and Decision Directed Channel Estimation (DDCE) are widely used in the estimation of channel in V2V communication. However, the PSACE faces the problem of pilot overhead since the technique demands huge number of pilots for estimation process. The DDCE which depends on data symbols and it suffers from error propagation in dynamic environment. To crack these challenges, a hybrid method is proposed which combines the properties of both the PSACE and DDCE. The proposed strategy employed a bio-inspired approach as Improved Crow Search Algorithm (ICSA) and a Machine Learning (ML) technique as Weight Regulated Convolution Neural Network (WCNN) to explore the issues caused by PSACE and DDCE. newlineThe ICSA is devised to optimize the huge number of pilot s thereby reducing the issue of pilot overhead. The optimization is achieved through the effective fitness function which has mitigated the Doppler Effect as well as the regulation of search pattern. The channel estimation is successful since the regulation of search pattern has avoided the premature convergence which in turn has guaranteed a better performance in the estimation of channel through PSACE. The channel estimation with ICSA is compared with various conventional bio inspired models through simulations newline