Assessment of an Adaptive Cruise Control Based on Model Predictive Control for Buses under Passenger Load Distribution Variation

Car and pedestrian impacts are one of the most frequent accidents on the roads due to driver distraction or misjudgment of traffic. In urban environments, these accidents normally occur at relatively low speeds where the impacted car is already at standstill, or the pedestrian suddenly crosses the path. Although current Adaptive Cruise Control (ACC) technology is available for different vehicle types, it mostly addresses use cases that may have limited applicability for urban passenger buses (e.g., high speeds on divided highways). In this work, the design and verification of a novel ACC based on Model Predictive Control (MPC) for a 12m urban bus is presented. As in transit applications, the influence of the passenger bus load distribution is considered to evaluate the ACC capability to offer both safety and comfort. Moreover, performance limitations such as maximum deceleration, sensor ranging, and actuation time-lag are modeled to resemble real systems. The approach is verified in a virtual environment considering three challenging driving scenarios under 30km/h: Car-to-Car Rear (CCR) stationary, CCR braking, and car-to-pedestrian. Results demonstrate the robustness of the verified system, assuring both safe and comfortable behavior regardless of the load distribution and performance limitations.