Risk-based merging decisions for autonomous vehicles.

Area of Science:

• Robotics and Control Systems
• Transportation Engineering
• Traffic Safety

Background:

• Autonomous vehicles (AVs) face challenges in safe freeway merging within mixed traffic environments, posing risks to safety and efficiency.
• Existing merging strategies may not adequately address the complexities introduced by non-AVs in the traffic stream.

Purpose of the Study:

• To quantify freeway merging conflict risk for AVs in mixed traffic.
• To develop and evaluate an AV merging decision strategy based on conflict risk assessment.

Main Methods:

• Developed a risk-based merging decision strategy for AVs.
• Evaluated the strategy under various traffic conditions: uncongested, near-congested, and congested.
• Compared the strategy against base models using metrics like Time-to-Collision (TTC), Time Exposed Time-to-Collision (TET), and Time Integrated Time-to-Collision (TIT).

Main Results:

• The risk-based strategy resulted in less abrupt deceleration of upstream vehicles in the target lane.
• It ensured minimum safe gaps with downstream vehicles.
• Demonstrated lower conflict risk (TET, TIT) and minimal impact on average traffic speed compared to base models.

Conclusions:

• The developed risk-based merging strategy offers significant safety benefits for AVs.
• It improves the overall safety of AV merging operations in mixed traffic streams.
• Findings can inform the design of AV controllers for enhanced safety.