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

TL;DR

This work tackles safe, force-controlled actuation inside MRI by integrating a compact transmission force sensing SEA (TFSEA) with a velocity-sourced ultrasonic motor. A disturbance-observer (DOB) based torque controller enforces a reference model for the SEA under varying external impedance, achieving robust torque tracking with a target of $5\%$ settling time of $0.05$ s and a steady-state error within $2.5\%$ of the maximum torque. The mechanical module fits within an $80$ mm diameter and a $66$ mm length envelope and uses a planetary gearbox with $N_i^o = 3/10$ and four springs with $k_s = 4.3$ N/mm, while MR compatibility is addressed through non-magnetic materials and MR artifact testing showing acceptable $\eta_0$ perturbations; MR field deviations are characterized and mitigated by positioning strategies. Experimental validation in both 3T MRI and non-MRI environments demonstrates consistent DOB-based torque control performance across low and high external impedance, enabling MRI-guided medical interventions such as brain stimulation with a TMS coil. Overall, the compact MRI-compatible SEA with DOB-based torque control offers a practical path toward safe, precise force control in MRI-guided robotics.

Abstract

In this study, we introduce a novel MRI-compatible rotary series elastic actuator module utilizing velocity-sourced ultrasonic motors for force-controlled robots operating within MRI scanners. Unlike previous MRI-compatible SEA designs, our module incorporates a transmission force sensing series elastic actuator structure, with four off-the-shelf compression springs strategically placed between the gearbox housing and the motor housing. This design features a compact size, thus expanding possibilities for a wider range of MRI robotic applications. To achieve precise torque control, we develop a controller that incorporates a disturbance observer tailored for velocity-sourced motors. This controller enhances the robustness of torque control in our actuator module, even in the presence of varying external impedance, thereby augmenting its suitability for MRI-guided medical interventions. Experimental validation demonstrates the actuator's torque control performance in both 3 Tesla MRI and non-MRI environments, achieving a 5% settling time of 0.05 seconds and a steady-state error within 2.5% of its maximum output torque. Notably, our torque controller exhibits consistent performance across low and high external impedance scenarios, in contrast to conventional controllers for velocity-sourced series elastic actuators, which struggle with steady-state performance under low external impedance conditions.

Figures (9)

• Figure 1: Illustrated in (a) and (b), an SEA robot maneuvers a transcranial magnetic stimulator (TMS) towards a patient's head. The SEA torque controller needs to address the challenges posed by both (a) the minimal external impedance of the medical device before it makes contact with the patient and (b) the substantial impedance encountered upon contact with the human body.
• Figure 2: In (a), a conventional direct force sensing SEA positions its spring element between the output gear and the load. Contrarily, in (b), a transmission force sensing SEA situates its spring element between the gearbox housing and the ground. The transmission force sensing SEA architecture enables the construction of a more compact SEA suitable for MRI scanners. $\mathsf{f_s}$ denotes the spring force. $\mathsf{N_{i}^{o}}$, $\mathsf{N_{i}^{h}}$, and $\mathsf{N_{o}^{h}}$ denote the gear ratios between the input and the output, between the input and the gearbox housing, and between the output and the gearbox housing.
• Figure 3: The architecture of our compact SEA module comprises essential components: a spring element (a), a gearbox (b), an USM (c), and an optical spring encoder (d). Integration is facilitated by a triple ring bearing (e), allowing the entire framework to fit seamlessly into a cylindrical space, as depicted in (f), with a diameter of $80$ mm and a total length of $66$ mm.
• Figure 4: In the setup for the image artifact test, the SEA is positioned 20 cm in front of a spherical phantom inside a $3$T MRI scanner, as shown in (a). The $\mathrm{B_0}$ field deviation ($\Delta \mathrm{B_0}$) caused by the SEA is shown in the coronal plane (b) and sagittal plane (c). The dashed lines in (b) and (c) indicate the boundaries of the field of view (FOV).
• Figure 5: For the example use case, a shoulder SEA is positioned at the origin ($\mathtt{x_s} = 0, \ \mathtt{z_s} = 0$) with a shoulder-to-head distance of $60$ cm ($\mathtt{x_h} = 60 \ \mathrm{cm}, \ \mathtt{z_h} = 0$). The robot dimensions are $\mathtt{L_0} = 5$ cm, $\mathtt{L_1} = 20$ cm, and $\mathtt{L_2} = 10$ cm.