Design and Control of a Compact Series Elastic Actuator Module for Robots in MRI Scanners

Robotic assistance has broadened the capabilities of magnetic resonance imaging (MRI)-guided medical interventions, yet force-controlled actuators tailored for MRI environments remain limited. In this study, we present a novel MRI-compatible rotary series elastic actuator (SEA) module that employs velocity-sourced ultrasonic motors for force-controlled operation within MRI scanners. Unlike prior MRI-compatible SEA designs, our module uses a transmission force sensing SEA architecture, with four off-the-shelf compression springs placed between the gearbox and motor housings. To enable precise torque control, we develop a controller based on a disturbance observer, specifically designed for velocity-sourced motors. This controller improves torque regulation, even under varying external impedance, enhancing the actuator’s suitability for MRI-guided medical interventions. Experimental validation confirms effective torque control in both 3 Tesla MRI and non-MRI settings, achieving a 5% settling time of 0.05 seconds and steady-state error within 2.5% of the actuator’s maximum output torque. Notably, the controller maintains consistent performance across both low and high impedance conditions.

💡 Research Summary

This paper presents a compact, MRI‑compatible rotary series elastic actuator (SEA) module and a robust torque‑control strategy tailored for use inside magnetic resonance imaging (MRI) scanners. The authors first identify the safety and compatibility challenges that conventional electromagnetic, hydraulic, and pneumatic actuators face in the high‑field, confined environment of an MRI bore. While non‑magnetic ultrasonic motors (USMs) have been used for MRI‑compatible actuation, they are typically employed as pure velocity sources, making precise force or torque regulation difficult, especially when external impedance varies between free‑space motion and contact with a patient.

To overcome these limitations, the authors adopt a transmission‑force‑sensing (TFSEA) architecture. Instead of placing the elastic element between the output gear and the load (the traditional direct‑force‑sensing SEA), they locate four off‑the‑shelf 316 stainless‑steel compression springs between the gearbox housing and the stationary ground. This configuration fixes the springs to the non‑moving side, allowing the motor, gearbox, and encoder electronics to be tightly integrated into a cylindrical envelope of only 80 mm diameter and 66 mm length. The springs are arranged in parallel via two sliders, providing a nearly constant torsional stiffness of ≈10.5 N·m·rad⁻¹ over a ±12° rotation range and delivering up to 3.2 N·m at the output.

The actuation core is a velocity‑sourced USM (WLG‑75‑R) that delivers a rated speed of 4.8 rad s⁻¹ and a nominal torque of 2.5 N·m (peak output 12 W). Because USMs behave as high‑impedance velocity sources, conventional current‑based torque controllers are ineffective. The authors therefore design a disturbance observer (DOB)–based torque controller specifically for velocity‑sourced motors. The DOB estimates the mismatch between the actual plant input and a reference model, then injects a compensating signal that renders the closed‑loop dynamics independent of external load impedance. An internal inverse‑dynamics block converts torque commands into the required velocity references for the USM. The controller parameters are tuned once and remain unchanged across all test conditions.

Experimental validation is performed both inside a 3 Tesla MRI scanner and in a standard laboratory setting. The SEA is subjected to low‑impedance (free‑space) and high‑impedance (rigid contact) scenarios to emulate the transition from non‑contact positioning to patient interaction in MRI‑guided brain‑stimulation applications. Results show a 5 % settling time of 0.05 s and a steady‑state error within 2.5 % of the maximum output torque, regardless of the impedance condition. Moreover, the MRI images exhibit no measurable degradation, confirming the non‑magnetic nature of the components and the suitability of the design for in‑bore deployment.

In summary, the paper delivers three major contributions: (1) a novel, fully in‑bore MRI‑compatible rotary SEA that integrates a USM, planetary gearbox, and a compact spring assembly using the TFSEA concept; (2) a DOB‑based torque controller adapted for velocity‑sourced actuators, offering robustness to large variations in external load impedance; and (3) experimental evidence that the system maintains consistent performance in both MRI and non‑MRI environments without retuning. This work paves the way for multi‑degree‑of‑freedom MRI‑compatible robots, such as those required for transcranial magnetic stimulation or other precision interventions, by providing a safe, compact, and precisely controllable actuation platform.