An Approach to Improving Attitude Estimation Using Sensor Fusion for Robot Navigation

Abstract

Recently, many low-cost micro electro mechanical systems (MEMS) IMUs have emerged for only several hundred US dollars. In comparison to high-end IMUs, an entire Inertial Navigation System (INS) can be implemented with smaller size/volume, lower weight and costs. On the other hand, they have a relatively lower accuracy due to their larger systematic errors, such as bias, scale factor and drift, which highly depend on disturbance and temperature. Consequently, the original signal output of low-cost IMU must be processed to reconstruct smooth attitude estimation. For the application of mobile robot navigation, the algorithms need to be run on embedded processors with low memory and processing resources. This paper analyses various error sources in the attitude measurement for Mobile Robot using low-cost MEMS-IMU and discusses how to improve measurement accuracy by minimizing the errors and optimizing fusion algorithms. It presents the following aspects: investigating a cheap open source MEMS-IMU, enhancing ADC resolution by oversampling and averaging, filtering the noise caused by vibration, improving attitude estimation using sensor fusing.