Optimization and Design of Hardware System for Campus Intelligent Vehicles

With the rapid advancement of smart campus construction, the demand for efficient and adaptive campus mobility solutions has grown significantly. Existing campus intelligent vehicles, however, suffer from redundant sensors, mismatched computing power, high energy consumption, and poor stability—failing to meet the unique needs of campus scenarios characterized by low speed, dense pedestrians, and dynamic environments (e.g., tree shade, temporary obstacles). To address this gap, this study optimizes the hardware system for campus-specific use cases by building on prior system design frameworks. First, we clarify the core connotations of the campus intelligent vehicle hardware system. Then, we select key components tailored to campus requirements: a 16-line LiDAR (HESAI XT16) for short-range 360° perception, an RTK-GPS+IMU (OXTS RT3002) integrated with the UKF algorithm for high-precision positioning, and an Intel i7-8700 (IPC) for efficient data fusion. To resolve issues in previous hardware setups, we conduct phased integration and debugging, including single-component testing, multi-sensor calibration and vehicle- level road tests. Experimental data from campus road validation show that the optimised system achieves a positioning error rate of only 0.82% and an average takeover distance of 13.34 km. These results demonstrate that the system effectively balances performance, cost, and power consumption, confirming its feasibility and adaptability for campus intelligent vehicle applications.