A novel framework for adaptive stress testing of autonomous vehicles in multi-lane roads

TL;DR

Quantitative and qualitative analyses of the experimental results demonstrate that the proposed novel AST framework outperforms the state-of-the-art AST scheme in identifying corner cases with complex driving maneuvers.

Abstract

Stress testing is an approach for evaluating the reliability of systems under extreme conditions which help reveal vulnerable scenarios that standard testing may overlook. Identifying such scenarios is of great importance in autonomous vehicles (AV) and other safety-critical systems. Since failure events are rare, naive random search approaches require a large number of vehicle operation hours to identify potential system failures. Adaptive Stress Testing (AST) is a method addressing this constraint by effectively exploring the failure trajectories of AV using a Markov decision process and employs reinforcement learning techniques to identify driving scenarios with high probability of failures. However, existing AST frameworks are able to handle only simple scenarios, such as one vehicle moving longitudinally on a single lane road which is not realistic and has a limited applicability. In this paper, we propose a novel AST framework to systematically explore corner cases of intelligent driving models that can result in safety concerns involving both longitudinal and lateral vehicle's movements. Specially, we develop a new reward function for Deep Reinforcement Learning to guide the AST in identifying crash scenarios based on the collision probability estimate between the AV under test (i.e., the ego vehicle) and the trajectory of other vehicles on the multi-lane roads. To demonstrate the effectiveness of our framework, we tested it with a complex driving model vehicle that can be controlled in both longitudinal and lateral directions. Quantitative and qualitative analyses of our experimental results demonstrate that our framework outperforms the state-of-the-art AST scheme in identifying corner cases with complex driving maneuvers.

Figures (6)

• Figure 1: Our proposed Adaptive Stress Testing framework for finding crashed scenario in the multi-lane roads. AV, FL and LV abbreviations refer to the vehicle (i.e., ego vehicle), following vehicle, and leading vehicle, respectively.
• Figure 2: Illustration of interaction between vehicles with two components: other vehicle safety and ego vehicle collision.
• Figure 3: Lane change scenario with our uIDM which is similar to MOBIL mobil2007. The following vehicles in the left, ego, and right lanes each have their own accelerations, denoted as $a_{n,l},a_{o},a_{n,r}$ respectively. The ego vehicle, which is currently accelerating at a rate of $a_e$, has the option to switch to either the left lane or the right lane, with new accelerations of $\widetilde{a}_{e,l}$ and $\widetilde{a}_{e,r}$, respectively. After the ego vehicle changes lanes, the following vehicle can move with the new accelerations $\widetilde{a}_{n,l},\widetilde{a}_o,\widetilde{a}_{n,r}$, respectively.
• Figure 4: Comparison of our proposed framework with AST4AV ast4av2018 in percentage of crashed scenario on velocity range.
• Figure 5: Comparison of our proposed framework with AST4AV ast4av2018 in percentage of crashed scenario on lane index.