A safety-prioritized trajectory planning method based on multimodal predictive risk field for ramp merging

Area of Science:

• Traffic safety engineering
• Intelligent transportation systems
• Robotics and control systems

Background:

• Ramp merging zones pose significant safety risks due to complex vehicle interactions and speed differentials.
• Traditional trajectory planning for connected and automated vehicles (CAVs) struggles with the dynamic uncertainties of real-world traffic.
• Existing methods often rely on deterministic predictions, limiting adaptability in complex merging scenarios.

Purpose of the Study:

• To develop an advanced trajectory planning framework for CAVs specifically addressing safety challenges in ramp merging scenarios.
• To enhance the adaptability and safety of CAVs in dynamic and uncertain traffic environments.
• To reduce collision risks at ramp merging areas through improved trajectory planning.

Main Methods:

• Constructed a multimodal predictive risk field (MPRF) to model uncertainties in vehicle behaviors.
• Developed a multimodal trajectory prediction model using a modified interactive multiple model (IMM) Kalman filter with a priority mechanism.
• Created a dynamic risk field calibrated with real-world aerial trajectory data processed via YOLO-based computer vision.
• Designed separate risk matrices for path and speed planning to independently optimize trajectory components.

Main Results:

• The proposed MPRF framework effectively captures uncertainties in vehicle behaviors.
• The IMM Kalman filter integration improved the accuracy of lateral and longitudinal movement predictions.
• The calibrated dynamic risk field accurately reflects diverse driving behaviors.
• Separating path and speed planning enhanced the handling of multimodal prediction uncertainties and dynamic-static coupling.

Conclusions:

• The novel trajectory planning framework significantly improves safety performance in ramp merging scenarios.
• The method maintains traffic efficiency while enhancing safety.
• The MPRF approach offers a robust solution for safe and efficient autonomous driving in complex traffic situations.