Design and calibration of an inertial sensor system for precise vehicle navigation

Abstract

accumulate with time. The errors associated with absolute sensors on the other hand are fixed. However, the update rates are generally low [2].Abstract-This paper presents a developed inertial navigation multi sensor node, along with an efficient calibration technique to improve the accuracy of the measurement acquired from a set of inertial sensors and magnetic encoder. The implemented node consists of a high resolution magnetic compass sensor, a distance sensor works as magnetic encoder, an accelerometer and a gyroscope to detect the linear and angular acceleration. The node is installed on an autonomous mobile vehicle as part of an inertial system to find the real-time position and direction of the vehicle. RF wireless transceivers and a microcontroller are installed to process and to send the measured data wireless to a central processing unit for more processing and evaluation. The whole hardware and software design of the system is presented. Accuracy and calibrations issues are also discussed. Errors caused by bias, scale factors and nonlinearities in the sensor readings which cause accumulations in navigation errors with time are considered.An extensive body of research on inertial navigation systems and theirs applications has been reported in the literature. The work of Beauregard [4] describes an approach for using shoe mounted sensors and inertial mechanization equation to directly estimate the displacement of the feet between footfalls. A pedestrian tracking framework based on particle filters is proposed in [3]. This framework is supposed to extend the typical WLAN -based indoor positioning systems by integrating a low cost MEMS accelerometer and map information. Different of these research contributions our approach concentrates more on the practical design concepts and inertial measurements integration aspects in order to get optimal navigation accuracy.The main contribution in this work has been on the development of a accurate inertial navigation system based on thorough characterization of the errors which inversely affect the navigation accuracy. Our goal is to achieve a minimum error of position and direction over a given travelled distance. A well designed recursive filter will be still needed to minimize the continuously accumulated heading error Index Terms-Indoor precise navigation, Inertial sensors Wireless sensor node, Sensor fusion.

Intelligent Real-Time MEMS Sensor Fusion and Calibration

“…In inertial navigation it is required to measure vectors of linear acceleration and angular velocity in all three axes (the 3-axial MEMS vibrational gyroscope and accelerometer are usually used for this purpose [4]). Each axis has its own gain and the offset calibration functions (independently from each other) according to (33).…”

“…Zhang et al proposed a method of estimation of the calibration constants for the 3-axial inertial sensor (gyroscope, accelerometer) [2]. Gyroscope bias is determined directly in a steady state and accelerometer bias is computed after multiple steps when the acceleration sensor is oriented vertically along each of its axes one by one or the sensor has to be exposed to precisely known stimuli [3] [4]. All these methods are working in the offline mode.…”

Decentralized Data Fusion Algorithm Using Factor Analysis Model

“…Support vector machine (SVM) was used for training process. Qasem et al [15] developed inertial navigation multi sensor node to improve the accuracy of the measurement acquired from a set of inertial sensors and magnetic encoders. Koo et al [16] presented a real-time heading estimation algorithm using IMU and strap down magnetometer without any other external heading reference system.…”

Control of Mobile Robot Using Web Interface

“…In recent years the area of mobile robots has become very active in both research and development. Are being developed a new, technically more sophisticated robot design, increases in which the right to sophisticated control systems which are already established, for example, the fuzzy logic or neural networks [5], [6].…”