IMPROVING CHANNEL ESTIMATION IN VEHICULAR COMMUNICATION NETWORKS WITH DIVERSE MOBILITY PATTERNS USING BI-LSTM MODEL

Accurate and low-latency channel estimation is essential for reliable vehicle-to-vehicle (V2V) communication in high-mobility environments such as intelligent transportation systems (ITS). Conventional techniques such as Least Squares and Minimum Mean Square Error perform poorly in dynamic wireless environments. This research introduces a deep learning (DL)-based channel estimation model employing a Bi-directional Long Short-Term Memory (Bi-LSTM) network, and evaluates its performance against traditional methods as well as a range of machine learning (ML) and DL models, including Support Vector Machine (SVM), Random Forest (RF), Extreme Gradient Boosting (XGBoost), Recurrent Neural Network (RNN), and LSTM. Using the CN+ vehicular dataset, the models were trained on features like velocity, distance, signal strength, Doppler shift, and path delay. Results show that ML models, particularly RF and XGBoost, achieve high accuracy, with RF reaching 99.94%. Among DL models, Bi-LSTM performs best with 98.58% accuracy, and outperforms other models under high-speed conditions, due to its ability to capture temporal dependencies and track rapid channel variations. Thus, AI-based approaches can enhance channel estimation for safer and smarter vehicular communication systems.