This paper addresses the problem of traffic rule compliance in autonomous vehicles. While existing methods replan routes from scratch when traffic rules are violated, this paper proposes a route correction technique to save computational time. Combining Satisfaction Module Theory (SMT) and set-based reachability analysis, we determine whether and how the initial route can be corrected. Experiments in a high-fidelity simulator and real-world environments demonstrate the benefits of the proposed method in a variety of scenarios. Even in complex environments, the proposed method efficiently and reliably corrects violating routes, enabling autonomous vehicles to safely resume driving in real time.
