Automobile system for predicting the trajectory of surrounding vehicles

The improvement of autonomous driving systems and advanced driver assistance systems (ADAS) heavily relies on accurate vehicle trajectory prediction. This research focuses on developing a robust method for predicting the trajectories of surrounding vehicles by leveraging machine learning techniques and neural networks. The study integrates data from onboard sensors, vehicle-to-vehicle (V2V) communication, cameras, LiDAR, and Differential GPS (DGPS) to enhance the accuracy and reliability of trajectory forecasts. Traditional approaches, primarily based on physical models, fall short in complex driving scenarios due to their dependency on uniform motion parameters. The investigated approach addresses these limitations by employing separate algorithms to predict longitudinal and lateral positions, thereby improving safety and reducing collision risks. The implemented algorithms were initially tested on simulated data, confirming their functionality. Future steps involve collecting and preparing real-world data to evaluate the algorithms under diverse and complex road conditions. This paper lays the groundwork for future developments in collision avoidance systems and highlights the potential benefits of integrating advanced perception systems in enhancing environmental perception and data quality. The proposed method shows promise in mitigating traffic accidents and optimizing traffic flow, underscoring its importance for future automotive innovations.