Qazi Hamza Jan

• Navigating driverless minibuses through pedestrian-dense areas presents a significant challenge, exacerbated by the necessity for these vehicles to transport passengers efficiently to their destinations. For such vehicles to maintain a pace equal to or faster than walking speed, it is important to minimize unnecessary braking and avoid unwarranted stops. Achieving such navigation proficiency in pedestrian zones requires driverless minibuses to possess advanced interactive systems capable of signaling their intentions to pedestrians preemptively, thereby reducing potential misunderstandings. This thesis focuses on the development and implementation of strategies for the efficient and safe operation of driverless minibuses in pedestrian environments. Although these vehicles are equipped with comprehensive safety systems designed to initiate emergency stops to avert collisions, such interventions can disrupt their smooth flow. It is essential for such vehicles to interpret pedestrian intentions and their awareness levels accurately to prevent the undue activation of emergency stop mechanisms and pauses. Moreover, it is vital to alert pedestrians who may be oblivious to the vehicle's proximity. To this end, the thesis explores various interaction strategies, including auditory signals, visual cues, and car control adjustments, to facilitate clear communication with pedestrians. Smart interaction strategy is proposed to enhance interactions between driverless minibuses and pedestrians, incorporating four pivotal elements: pedestrian behavior identification, interaction zone establishment, decision-making processes, and the activation of interaction modules. These elements are foundational for achieving a sophisticated interaction level. Identifying pedestrian behavior is crucial for detecting their potential unawareness of the vehicle's proximity. Interaction zones are crafted to calculate a risk value based on pedestrian behavior and proximity, guiding the decision-making process within the navigable space. Following this assessment, interaction modules are deployed to actively engage with pedestrians, embodying the full spectrum of the interaction strategy developed in this study. The integration of these modules into a standard navigation system enables effective operation in high pedestrian traffic conditions. The smart interaction strategy was evaluated using a practical system referred to as "Autobus", which is analogous to a driverless minibus. Initial tests of these components were conducted separately to verify their efficacy, followed by their amalgamation into the system to realize the intended outcomes. Experimental findings confirm that the implemented system enables Autobus to communicate effectively with pedestrians across various scenarios, significantly reducing the frequency of unnecessary stops.