CBF-Based Collision Avoidance Approach of the Mine Inspection UAV in the Narrow GNSS-Denied Environments
Jiahao Yang, Jie Fang, Hu Cao, Yongran Zhi, Jingke Zhou, Lei Liu · 2025
This paper addresses the crucial problem of collision avoidance for unmanned aerial vehicles (UAVs) employed in mine inspection within narrow environments devoid of Global Navigation Satellite System (GNSS) signals. A novel approach based on Control Barrier Functions (CBF) is proposed. In such adverse circumstances, obtaining high-precision maps in advance is highly improbable, and precise localization is rendered extremely challenging due to the severe conditions that substantially impair sensor performance. To tackle these issues, this paper adopts velocity control, which depends exclusively on the state variables obtained from current sensor readings to govern the body-frame velocity of the UAV. Kalman filtering is applied to attenuate the white noise produced by the sensors, thereby augmenting the data reliability. Furthermore, the CBF-based mechanism is integrated to ensure efficient collision avoidance during the mine inspection process. Simulation experiments validate that the proposed method facilitates basic centralized inspection, close-range inspection at specific instants, and successfully accomplishes the avoidance of three obstacles. Comparative studies reveal that, in contrast to the traditional Artificial Potential Field (APF) method, the CBF-based approach demonstrates superior collision avoidance performance, proffering a viable solution for the safe operation of mine inspection UAVs in GNSS-denied narrow environments.