Towards IMU-based Full-body Motion Estimation of Rough Terrain Mobile Manipulators

Xiaolong Zhang

Research output: Book/ReportDoctoral thesisCollection of Articles

Abstract

For navigation or pose estimation, strap-down Micro-Electro-Mechanical System (MEMS) Inertial Measurement Units (IMU) are widely used in all types of mobile devices and applications, from mobile phones to cars and heavy-duty Mobile Working Machines (MWM). This thesis is a summary of work focus on the utilization of IMUs for state estimation of MWM. Inertial sensor-based technology offers an alternative to the traditional solution, since it can significantly decrease the system cost and improve its robustness.

For covering the research topic of whole-body estimation with IMUs, five publications focus on the development of novel algorithms, which use sensor fusion or rotary IMU theory to estimate or calculate the states of MWM. The test-platforms are also described in detail.

First, we used low-cost IMUs installed on the surface of a hydraulic arm to estimate the joint state. These robotic arms are installed on a floating base, and the joints of the arms rotate in a two-dimensional (2D) plane. The novel algorithm uses an Extended Kalman Filter (EKF) to fuse the output of the gyroscopes and the accelerometers, with gravity as the reference. Second, a rotary gyroscope is mounted on a grasper of a crane, and the rotary gyroscope theory is implemented to decrease the drift of the angular velocity measurement. Third, low-cost IMUs are attached to the wheels and the bogie test bed, and the realization of IMU-based wheel odometry is investigated. Additionally, the rotary gyroscope provides information about the roll and yaw attitude for the test bed. Finally, we used an industry grade IMU fuse with the output of wheel odometry to estimate the position and attitude of the base for an MWM moving on slippery ground.

One of the main aims of this research study is to estimate the states of an MWM only using IMU sensors. The research achievements indicate this approach is promising. However, the observability of IMU in the yaw direction of the navigation frame is limited so it is difficult to estimate the yaw angle of the rotation plane for the robotic arm when only using IMUs, to ensure the long-term reliable yaw angle and position of the vehicle base, external information might also be needed. When applying the rotary IMU theory, minimization of the power supply for the rotation device is still a challenge.

This research study demonstrates that IMUs can be low-cost and reliable replacements for traditional sensors in joint angle measurement and in the wheel rotation angle for vehicles, among other applications. An IMU can also provide a robust state for a vehicle base in a challenging environment. These achievements will benefit future developments of MWMs in remote control and autonomous operations.
Original languageEnglish
Place of PublicationTampere
ISBN (Electronic)978-952-03-2651-7
Publication statusPublished - 2022
Publication typeG5 Doctoral dissertation (articles)

Publication series

NameTampere University Dissertations - Tampereen yliopiston väitöskirjat
Volume705
ISSN (Print)2489-9860
ISSN (Electronic)2490-0028

Fingerprint

Dive into the research topics of 'Towards IMU-based Full-body Motion Estimation of Rough Terrain Mobile Manipulators'. Together they form a unique fingerprint.

Cite this