Angle-Weighted Time-to-Collision Algorithm for Enhanced Obstacle Avoidance in Autonomous Driving
2025-01-7022
01/31/2025
- Features
- Event
- Content
- To tackle the challenge of accurately predicting collision times for autonomous vehicles navigating complex dynamic obstacles, this paper proposes an innovative Angle-Weighted Time-to-Collision (AW-TTC) algorithm. Traditional TTC algorithms are known for their computational simplicity and strong real-time performance, making them widely applicable across various driving scenarios. However, they often struggle with predictive accuracy when encountering obstacles moving at angles, which can delay vehicle response and compromise overall safety. To address this limitation, this study introduces a modification to the traditional TTC algorithm by incorporating an angle-based weighting factor, improving collision time prediction accuracy. A Hardware-in-the-Loop (HIL) experimental setup was developed, utilizing a Vehicle Control Unit (VCU) and the SCANeR simulation platform to simulate dynamic obstacles in complex traffic scenarios. The AW-TTC algorithm’s performance was then evaluated, particularly in predicting collision times for obstacles with motion angles. Experimental results demonstrate that the AW-TTC algorithm significantly enhances predictive accuracy, achieving an average improvement of 0.8 seconds in collision time estimates compared to traditional TTC algorithms. Furthermore, it enables an increase in maximum vehicle speed by 4 km/h while maintaining safety standards. In summary, the AW-TTC algorithm not only improves collision time prediction accuracy in complex traffic scenarios but also enhances the system’s ability to manage angled obstacles effectively. These results highlight its potential as a reliable solution for obstacle avoidance in future autonomous driving systems, ensuring both safety and performance.
- Pages
- 9
- Citation
- Zhu, S., Tang, Y., Li, B., Wang, X. et al., "Angle-Weighted Time-to-Collision Algorithm for Enhanced Obstacle Avoidance in Autonomous Driving," SAE Technical Paper 2025-01-7022, 2025, https://doi.org/10.4271/2025-01-7022.