Effect of Wheel Odometer on Low-Cost Visual-Inertial Navigation System for Ground Vehicles

This paper deals with a specialized visual-inertial navigation system (VINS) for ground vehicles equipped with a monocular camera and an inertial measurement unit (IMU) aided by the embedded wheel odometer. In particular, this work is scoped to the systems with a low-cost IMU of which accelerometers are considerably unreliable and the global navigation satellite system (GNSS)-denied environments such as the urban canyon or the inside of tunnels. Since the accelerometer biases are fluctuating in low-grade cases, a general VINS cannot estimate them properly, which results in poor navigation performance. Meanwhile, the wheel odometers are embedded in most cars and provide speed information, allowing the accelerometer biases to be calibrated. This paper analyzes the effect of wheel odometer on scale accuracy of low-cost VINS solution through simulation using the KITTI benchmark dataset. The real-world experiment also verifies the preceding analysis and shows a remarkable improvement in navigation performance.