Enhancing Campus Mobility: Achievements and Challenges of Autonomous Shuttle "Snow Lion''

TL;DR

Snow Lion presents a campus autonomous shuttle designed to safely and efficiently transform on-campus mobility. It integrates multi-LiDAR perception, GNSS-aided LiDAR mapping, A* global routing, scenario-aware behavioral planning, GPMP-based motion planning, and MPC control to operate in unregulated campus environments. The study documents a real-world deployment spanning ~1.15k km and hundreds of passengers, revealing insights on localization robustness, perception noise from changing foliage, and interaction with non-compliant road users. The work contributes a practical, end-to-end shuttle system for campus mobility and offers actionable recommendations for improving reliability, safety, and user acceptance in autonomous campus fleets.

Abstract

The rapid evolution of autonomous vehicles (AVs) has significantly influenced global transportation systems. In this context, we present ``Snow Lion'', an autonomous shuttle meticulously designed to revolutionize on-campus transportation, offering a safer and more efficient mobility solution for students, faculty, and visitors. The primary objective of this research is to enhance campus mobility by providing a reliable, efficient, and eco-friendly transportation solution that seamlessly integrates with existing infrastructure and meets the diverse needs of a university setting. To achieve this goal, we delve into the intricacies of the system design, encompassing sensing, perception, localization, planning, and control aspects. We evaluate the autonomous shuttle's performance in real-world scenarios, involving a 1146-kilometer road haul and the transportation of 442 passengers over a two-month period. These experiments demonstrate the effectiveness of our system and offer valuable insights into the intricate process of integrating an autonomous vehicle within campus shuttle operations. Furthermore, a thorough analysis of the lessons derived from this experience furnishes a valuable real-world case study, accompanied by recommendations for future research and development in the field of autonomous driving.

Figures (9)

• Figure 1: This figure illustrates the operational scenario of our autonomous shuttle during its service period at The Hong Kong University of Science and Technology (Guangzhou) (referred to as HKUST (GZ)). The red lines represent the operational road map of the campus, with shuttle stations depicted as boxes (e.g., C1 and 1B). The subfigures on the right exhibit multiple images captured during the operation of the autonomous shuttle.
• Figure 2: The functions and connections of the whole system of the autonomous vehicle.
• Figure 3: The 3-D object detection module overview: Utilizing synchronized and well-calibrated LiDAR-captured point clouds, we employ an early-fusion technique to merge data from multiple calibrated LiDARs and apply VoxelNet zhou2018voxelnet for 3-D object detection using the fused results.
• Figure 4: The figure displays a satellite map of the campus on the left and a constructed point cloud map on the right.
• Figure 5: The planning pipeline of the AV, mainly consisting of four parts: global routing, behavioral planning, motion planning, and tracking controller.