Vehicle longitudinal control method and device, electronic equipment and storage medium
By acquiring vehicle status information and traffic environment information, determining the nominal feasible speed range and constructing a longitudinal risk profile, and generating continuous longitudinal control commands, the problem of insufficient generalization and safety of longitudinal control in multiple traffic scenarios is solved, and stable and safe vehicle operation is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CRSC URBAN RAIL TRANSIT TECH CO LTD
- Filing Date
- 2026-05-19
- Publication Date
- 2026-06-16
AI Technical Summary
Existing technologies suffer from insufficient generalization and safety in longitudinal control of multiple traffic scenarios. Reliance on multiple model switching leads to strategy complexity and discontinuity, while end-to-end learning results in weak constraint guarantees and difficulty in suppressing boundary risks.
By acquiring information on the vehicle's longitudinal operating status and traffic environment status, the nominal feasible speed range is determined, and a longitudinal risk characterization is constructed to generate longitudinal control commands. Traffic constraints and vehicle dynamics constraints are integrated to form a unified control framework. The longitudinal control model is trained using reinforcement learning to quantify the risk and generate continuous control commands.
It achieves stable, safe and efficient longitudinal vehicle movement control in complex multi-traffic scenarios, enhances the interpretability and verifiability of control decisions, detects and suppresses high-risk boundary conditions in advance, and solves the problems of discontinuous control and low training efficiency in traditional methods.
Smart Images

Figure 1 
Figure 2