Unexpected hazardous road situations frequently arise on the roads, which lead to crashes. Detecting a threat in advance and generating a fallback trajectory are some of the major challenges faced by autonomous vehicles. Stochastic Model Predictive Control (SMPC) approaches have recently proven to be highly effective in controlling systems in highly uncertain environments. However, the chance-constrained formulation of the SMPC makes it computationally expensive for fast real-time implementation. This paper presents a fast, proactive decision-making approach for crash avoidance in autonomous vehicles based on SMPC. Two problems are addressed in this study: 1) Assessing threat to the autonomous vehicle from the surrounding vehicles and 2) Fast safe trajectory generation for the autonomous vehicle in case of a future predicted crash. Threat assessment is addressed using stochastic reachability analysis for the surrounding vehicles, which accounts for surrounding drivers likely future motion and intents. We have proposed a fast, proactive decision-making algorithm to generate crash avoidance trajectories based on Stochastic Model Predictive Control (SMPC). We reformulate the SMPC probabilistic constraints as deterministic constraints using convex hull formulation, allowing for faster real-time implementation. This deterministic SMPC implementation ensures in real-time that the vehicle maintains a minimum probabilistic safety. It is shown that the proposed decision-making approach is successful in avoiding crash scenarios with a medium risk factor range.

This content is only available via PDF.
You do not currently have access to this content.