A Unified Boundary-Aware Prediction Model for Autonomous Driving

Automated driving prediction plays a crucial role in ensuring the safety, reliability, and efficiency of autonomous vehicles. Different from the previous research, this paper presents the whole-scenario prediction framework for autonomous driving. This method is based on a unified boundary for drivable space. By studying the temporal changes in boundaries, predictions can be made about the future drivable space. Through the application of a Backpropagation (BP) neural network, the relationship between the current boundary and the future boundary can be derived. The verification results show that this model has a comprehensive prediction for the drivable space boundary, which allows the system to make precise decisions and plans, ensuring the safety and efficiency of autonomous driving.