Latency-Reduced Equalizer with Model-Based Channel Estimation for Vehicle-to-Vehicle Communications

Equalization of fast time-varying channels is impacted by short coherence time and deep signal fading, especially in vehicle-to-vehicle communications due to the high mobility of user terminals. Furthermore, the stringent latency requirement of safety applications, such as collision avoidance, cannot tolerate high-complexity operations and long processing time. In order to achieve both requirements of equalization accuracy and short latency for fast varying channels, we propose a new model-based time-domain equalizer. In this equalizer, estimation and equalization are performed in two parallel parts to shorten the processing time. In main path, data symbols pass through an equalizer preset with the up-to-date channel impulse response. Since the channel variation model remains invariable for sufficiently long time, the current channel is estimated in parallel path from a number of past channel impulse responses to improve the accuracy. However, the channel variation during the long processing time of estimation leads to equalization error. Therefore, a predictor is used to update the channel response related to the processing delay, and perform the channel estimation beyond the channel coherence time. Thus, high accuracy and delay-free equalization can be achieved through this parallel structure.