Sensor Fusion for Accurate Object Detection Underneath Stationary Vehicles

Object detection under a vehicle's undercarriage is receiving a lot of attention nowadays. This paper proposes a system to detect objects under a stationary vehicle, whether the object is on the ground or attached to the undercarriage. Sensor fusion is used to increase system accuracy. Ultrasonic sensors as well as a LiDAR are used in the proposed system. Fusion is performed at the individual sensor level first then data from both types of sensors is fused together to accurately detect and locate a potential object. Experiments were successfully conducted using a HC-SR04 ultrasonic sensor and a TFmini LiDAR. A flexible system is then proposed that can be tuned to the characteristics of other types of ultrasonic sensors and LiDARs.