Electromagnetic Position Estimation Using High-Magnetic-Permeability Materials: Design, Modeling, and Applications

Abstract

This doctoral dissertation proposes and demonstrates a novel electromagnetic position sensing principle based on the use of alternating magnetic fields and on the use of passive high-magnetic-permeability metal located on the moving object. Position estimation systems robust to ambient ferromagnetic disturbances are developed based on this sensing principle for a range of applications, including: (1) 1-D position estimation system for piston position measurement in linear actuators; (2) Angular position estimation system for rotational motion of mechanical links; (3) 3-D position estimation for transluminal medical robots. Furthermore, an active position sensing system based on current control is designed for achieving better estimation performance with electromagnetic sensing that also minimizes the power consumption of the electromagnet. Many existing position sensing systems require either line-of-sight access to the concerned object (e.g. optical sensing systems), intrusive installation (e.g. magnetostrictive sensors), or need physical connection to the object (e.g. potentiometers), which limit their applications or result in poor durability. The new electromagnetic position estimation system developed in this dissertation provides a non-intrusive, non-contacting, easy to install, and low-cost position sensing system. Unlike permanent magnet-based sensors, the new sensing system also does not suffer from errors due to magnetic disturbances caused by surrounding ferromagnetic objects. This is because selective sensing of the amplitude of an alternating magnetic field at a specific operating frequency eliminates the influence of ambient low-frequency magnetic disturbances. Since an electromagnet is used as the magnetic field source for the position sensing system, active real-time control of the current of the electromagnet to ensure adequate magnetic sensitivity at all positions while reducing power consumption is another aspect studied in the research. Since the current control depends on the unknown position to be estimated, the problem is a coupled active control and estimation problem. A nonlinear observer is designed to simultaneously ensure stable position estimation and optimal current profile tracking.The research presented involves analytical and numerical modeling of the magnetic sensing system and design of stable estimation algorithms for a challenging nonlinear position estimation problem. The specific technical contributions of the dissertation are: (1) Development and verification of a novel passive position sensing principle based on high-magnetic-permeability metal. (2) Hardware development for signal processing of alternating magnetic fields to achieve position estimation immune to ferromagnetic disturbance. Disturbance rejection performance is characterized for each of the three types of applications. Nonlinear estimation algorithms are implemented for position estimation. (3) Development and experimental evaluation of 1-D position estimation system for linear actuators. Analytical modeling of the sensing principle for 1-D position estimation. (4) Development, numerical modeling, and experimental evaluation of angular position estimation system for rotational motion. (5) Development, modeling, and experimental evaluation of 3-D position estimation system for localization of transluminal medical robots. (6) Development of electromagnetic position estimation system using active current control for more accurate and energy-efficient position measurement. A nonlinear observer for simultaneous active control and estimation is designed based on Lyapunov theory. Experimental evaluation is pursued for all three types of position measurement applications. The 1-D position estimation system is shown to provide an accuracy of 1% over a stroke length of 20 cm for a non-ferrous actuator with the possibility to further enlarge the stroke length by daisy-chained additional sensors. However, its accuracy deteriorates significantly if the actuator is ferrous and the mu-metal is located completely inside a ferrous body. The angular position measurement sensor is shown to provide an accuracy of 1 degree over a range of ± 60 degrees in an off-joint configuration and an accuracy of 2 degrees over a full range of 360 degrees in an on-joint configuration. Finally, a 3-D position estimation system is evaluated on a trans-esophageal endoscopic robot wherein an RMS accuracy of 5 mm and a maximum error bounded within 9 mm are achieved. The new sensing principle developed and demonstrated in this project can be the foundation that leads to practically useful position estimation systems for a wide range of industrial, transportation and biomedical applications in the future.