Advanced Longitudinal Control and Collision Avoidance for High-Risk Edge Cases in Autonomous Driving
Dianwei Chen, Yaobang Gong, Xianfeng Yang
TL;DR
The paper tackles safety gaps in ADAS/ADS by modeling longitudinal control that accounts for both leading and following vehicles to prevent chain-reaction pileups at high speed. It introduces a DRL-based policy (DDPG) with a novel distance- and collision-aware reward function, plus a trajectory-calibration preprocessing framework to improve training realism and stability. The methodology casts the problem as an MDP with explicit state space and a two-neighbor penalty structure, deriving an equilibrium point $x^* = \frac{\beta (D-c_{mr}+C_2)+c_{fm}-C_1}{1+\beta}$ that maximizes reward while ensuring safe spacing; convexity and stability analyses support convergence toward this optimum. In CARLA simulations across edge-case scenarios, the approach raises no-collision success from $36.77\%$ (baseline) to $99\%$, demonstrating substantial practical impact for safer ADAS/ADS deployment.
Abstract
Advanced Driver Assistance Systems (ADAS) and Advanced Driving Systems (ADS) are key to improving road safety, yet most existing implementations focus primarily on the vehicle ahead, neglecting the behavior of following vehicles. This shortfall often leads to chain reaction collisions in high speed, densely spaced traffic particularly when a middle vehicle suddenly brakes and trailing vehicles cannot respond in time. To address this critical gap, we propose a novel longitudinal control and collision avoidance algorithm that integrates adaptive cruising with emergency braking. Leveraging deep reinforcement learning, our method simultaneously accounts for both leading and following vehicles. Through a data preprocessing framework that calibrates real-world sensor data, we enhance the robustness and reliability of the training process, ensuring the learned policy can handle diverse driving conditions. In simulated high risk scenarios (e.g., emergency braking in dense traffic), the algorithm effectively prevents potential pile up collisions, even in situations involving heavy duty vehicles. Furthermore, in typical highway scenarios where three vehicles decelerate, the proposed DRL approach achieves a 99% success rate far surpassing the standard Federal Highway Administration speed concepts guide, which reaches only 36.77% success under the same conditions.