On the Development of Autonomous Vehicle Safety Distance by an RSS Model Based on a Variable Focus Function Camera

Area of Science:

• Autonomous Driving Systems
• Robotics and Control Theory
• Computer Vision and Sensor Fusion

Background:

• Increasing adoption of autonomous driving features like adaptive cruise control (ACC) necessitates robust safety and reliability measures.
• Autonomous vehicles rely on sensor fusion (lidar, radar, cameras) to detect obstacles, but ensuring safety remains a critical challenge.
• The Responsibility-Sensitive Safety (RSS) model offers a mathematical framework for ensuring autonomous vehicle safety and clarifying accident liability.

Purpose of the Study:

• To investigate the application of the RSS model to a variable focus camera system.
• To adapt the RSS model's variables for a single camera sensor capable of replacing lidar and radar.
• To derive safe distances for autonomous vehicles equipped with variable focus cameras based on the RSS model.

Main Methods:

• Defined and determined RSS model variables specifically for variable focus camera functionality.
• Calculated safe distances for various vehicle velocities using the determined RSS variables.
• Analyzed data acquisition times and camera focal length adjustment times to validate response times.

Main Results:

• Successfully adapted the RSS model for variable focus camera integration in autonomous driving.
• Derived specific safe distances correlated with vehicle velocity under the RSS framework.
• Confirmed the validity of the calculated response times, considering sensor and camera operational latencies.

Conclusions:

• The integration of the RSS model with variable focus cameras is a viable approach to enhance autonomous driving safety.
• The derived safe distances provide a quantifiable measure for safe operation across different speeds.
• The study validates the system's responsiveness, crucial for real-world deployment of autonomous vehicles.