Mobile robot indoor dual Kalman filter localisation based on inertial measurement and stereo vision

Abstract

This study presents a novel navigation method designed to support a real-time, efficient, accurate indoor localisation for mobile robot system. It is applicable for inertial measurement units (IMU) consisting of gyroscopes, accelerometers, and magnetic besides stereo vision (SV). The current indoor mobile robot localisation technology adopts traditional active sensing devices such as laser, and ultrasonic method which belongs to the signal of localisation and navigation method which has low efficiency complex structure, and poor anti-interference ability. Through dual Kalman filter (DKF) algorithm, the accumulated error of gyroscope can be reduced, while combining with SV, mobile robot binocular SV orientation of inertial location can be realised under the DKF mechanism, which is introduced. First, high precision posture information of mobile robot can be obtained using fusing Kalman filter algorithm of accelerometer, gyroscope and magnetometer data. Second, inertial measurement precision can be optimised using Kalman filtering algorithm combined with machine vision localisation algorithm. The results indicate that the method achieves the levels of accuracy location comparable with that of the IMU/SV fusion algorithm; <0.0066 static RMS error, <0.0056 dynamic RMS error. The mobile robot using DKF algorithm of inertial navigation and SV indoor localisation is feasible.

A Review of Technologies and Techniques for Indoor Navigation Systems for the Visually Impaired

“…Cheng et al presented a study on a hybrid navigation method that combines computer vision with a set of inertial sensors to offer greater precision and efficiency in locating targets indoors [ 114 ]. The visual information is formed by a combination of two cameras (stereovision), which allows the calculation of a depth map and the distance for markers (reference points) or obstacles, in addition to reducing the errors accumulated by the gyroscope sensor.…”

A Systematic Review on Fusion Techniques and Approaches Used in Applications

“…Further, authors in [50] determined the localization of mobile robots in a complex environment by combining stereo vision localization with inertial measurement. They proposed a fusion technique that uses dual Kalman filter (DKF).…”

A Hybrid Indoor Positioning System Using a Linear Weighted Policy Learner and Iterative PDR

“…The linear Kalman filter was constructed to meet an intermediate profile between the Kalman filter, developed for linear problems, and the Extended Kalman filter, used in nonlinear problems [6], [31].…”

“…In the mapping process, each marker was considered a node in a bidirectionally-connected graph, registered in an adjacency matrix due to its low degree of complexity [31].…”

“…Due to a large amount of data presented as candidates for location indication, only k RPs are chosen according to the first k minimum distances identified by Equation (31).…”