Collision Avoidance/Mitigation System: Motion Planning of Autonomous Vehicle via Predictive Occupancy Map

Lee, Kibeom; Kum, Dongsuk · 2019 · DOAJ

DOI: 10.1109/ACCESS.2019.2912067

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the limitation of current autonomous vehicle systems, which often lack defensive driving capabilities to avoid collisions caused by human error. Despite advancements in automation, vehicles like Waymo’s have experienced collisions where human drivers would have performed evasive maneuvers. To resolve this, the authors propose a Collision Avoidance/Mitigation System (CAMS) that overrides normal driving algorithms when imminent collision risks are detected. The system is designed to rapidly evaluate risks from all surrounding vehicles and maneuver the ego-vehicle into a safer region, specifically targeting difficult-to-predict scenarios such as side and rear-end collisions. The methodology centers on a Predictive Occupancy Map (POM), a risk assessment module that computes potential risks based on the relative position, velocity, and acceleration of surrounding vehicles. The system uses a Constant Acceleration (CA) model for short-term trajectory prediction, suitable for immediate collision avoidance. To handle sudden acceleration changes, an Advanced Time-to-Occupancy (ATTO) metric is introduced, which adjusts for acceleration effects. The POM integrates these vehicle risks with environmental risks, including drivable boundaries and traffic lane penalties, creating a comprehensive spatial risk map. Motion planning involves selecting the safest trajectory from 12 pre-determined local candidates spaced 30 degrees apart. This selection is constrained by vehicle stability limits defined by a g-g diagram, ensuring that lateral and longitudinal acceleration profiles remain within physical friction and engine limits. The performance of the CAMS was validated through simulations involving side and rear-end collision scenarios. The results demonstrate that the system can detect collision risks 1.4 seconds before a crash occurs. By evaluating the maximum, mean, and minimum risk values along each candidate trajectory, the algorithm successfully identified and executed a safe avoidance maneuver. The simulations confirmed that the POM effectively distinguishes between high-risk trajectories (e.g., those leading to collisions with approaching obstacles) and low-risk alternatives, allowing the vehicle to navigate away from danger while maintaining stability. The significance of this work lies in its ability to provide a robust, computationally efficient framework for emergency collision avoidance in multi-vehicle environments. Unlike previous methods that focused on single obstacles or normal driving conditions, this approach integrates spatial and temporal risk assessment to handle unexpected maneuvers by other drivers. The proposed system offers a practical solution for enhancing the safety of autonomous vehicles by enabling defensive driving behaviors, thereby addressing a critical gap in current autonomous driving technologies.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• telematics crash prediction