Abstract:When the VINS-Mono is applied to a wheeled robot, due to the low signal-to-noise ratio of the inertial measurement unit(IMU) accelerometer and the inaccurate observation scale, the localization accuracy will decrease. For this problem, an improved algorithm that fusion of monocular camera, IMU and encoder is proposed. The encoder measurement residual term is added to the objective function of VINS-Mono initialization and back-end optimization. The speed calculated by the encoder data is directly fused to enhance the observability of the scale, reduce the accumulated localization error, and improve the localization accuracy. Moreover, in order to reduce the influence of wheel slip on the localization accuracy, the slip factor is calculated by IMU gyro data to adaptively adjust the weight of the encoder measurement residual in the objective function and the threshold of its robust kernel function. Experiments on a two-wheel robot show that the improved algorithm has high robustness, and the localization accuracy of the improved algorithm is an order of magnitude higher than VINS-Mono.