Trajectory and Motion Prediction of Autonomous Vehicles Driving Assistant System using Distributed Discriminator Based Bi-Directional Long Short Term Memory Model

The trajectory and motion prediction of autonomous vehicles stands as a pivotal element in guaranteeing the safety and effectiveness of self-driving systems. Yet, current models frequently encounter challenges in providing precise forecasts of the behaviors exhibited by nearby vehicles and objects. Traditional approaches often grapple with the intricate and ever-changing dynamics of real-world traffic situations, resulting in predictions that fall short in accuracy and, consequently, may introduce safety concerns. Moreover, they face challenges in handling large-scale datasets and real-time processing. To overcome these limitations, this research endeavors to develop a distributed discriminator-based Bi-directional Long Short Term Memory (Distributed discriminator-based Bi-LSTM) model for trajectory and motion prediction in order to predict the capabilities of autonomous vehicles and contribute to safer and more reliable self-driving systems. The Keystone for achieving precise predictions resides in the Distributed discriminator-based Bi-LSTM model, a sophisticated network that adeptly incorporates both historical and forthcoming trajectory data for distinguishing different objects and discerning intricate motion patterns. Harnessing the power of real-time tracking data, it endows autonomous vehicles with the capacity to make well-informed decisions and chart optimal routes, thereby significantly elevating their predictive prowess when navigating intricate and multifaceted road scenarios. For instance, in the case of D1, the model achieves outstandingly low error levels with MAE at 2.02, RMSE at 2.47 and MSE at 6.09. Similarly, in the K-fold 10 scenario, the Distributed Discriminator-based Bi-LSTM surpasses the existing models with notably reduced errors, registering MAE at 2.59, RMSE at 3.29 and MSE at 10.83. These results underscore the significant efficiency enhancement offered by the proposed model.