Design of a 6-DoF Parallel Robotic Platform for MRI Applications

Abstract

In this work, the design, analysis, and characterization of a parallel robotic motion generation platform with 6-degrees of freedom (DoF) for magnetic resonance imaging (MRI) applications are presented. The motivation for the development of this robot is the need for a robotic platform able to produce accurate 6-DoF motion inside the MRI bore to serve as the ground truth for motion modeling; other applications include manipulation of interventional tools such as biopsy and ablation needles and ultrasound probes for therapy and neuromodulation under MRI guidance. The robot is comprised of six pneumatic cylinder actuators controlled via a robust sliding mode controller. Tracking experiments of the pneumatic actuator indicates that the system is able to achieve an average error of 0.69 ± 0.14 mm and 0.67 ± 0.40 mm for step signal tracking and sinusoidal signal tracking, respectively. To demonstrate the feasibility and potential of using the proposed robot for minimally invasive procedures, a phantom experiment was performed in the benchtop environment, which showed a mean positional error of 1.20 ± 0.43 mm and a mean orientational error of 1.09 ± 0.57°, respectively. Experiments conducted in a 3T whole body human MRI scanner indicate that the robot is MRI compatible and capable of achieving positional error of 1.68 ± 0.31 mm and orientational error of 1.51 ± 0.32° inside the scanner, respectively. This study demonstrates the potential of this device to enable accurate 6-DoF motions in the MRI environment.

Keywords: MRI; motion generation; parallel robot.

Fig. 1.

CAD rendering of the MR-conditional parallel robotic platform in the 6-UPS Stewart–Gough platform configuration: (a) robot in the homed configuration; coordinate frames are assigned to the center of the fixed base and moving platform at the points $O_B$ and $O_P$, respectively, (b) robot in the fully extended configuration, (c) (top) fixed base geometry, (bottom) moving platform geometry.

Fig. 2.

Vector diagram for the $i$th limb used to solving the inverse kinematics problem.

Fig. 3.

Workspace of the parallel robot using the constant orientation representation: (a) view of the workspace in from $Z-Y$ plane, (b) view of the workspace from the $Z-X$ plane, (c) view of the workspace from the $X-Y$ plane.

Fig. 4.

Pneumatic circuit diagram of the double-acting cylinder.

Fig. 5.

(a) Comparison of set point tracking of the pneumatic cylinder under various loading conditions (steady-state error is given in brackets). (b) Results of the dynamic tracking experiment of the pneumatic cylinder in response to a sinusoidal input.

Fig. 6.

(a) CAD rendering of the proposed robot mounted to the patient in the MRI for abdomen interventions. (b) Experimental setup of the tissue-mimic phantom targeting experiment. (c) Phantom targeting experiment result. The red markers indicate the location of the desired target and the blue markers indicate the measured position of the needle tip.

Fig. 7.

(a) 3D gradient echo sequence images of a phantom bottle taken in a 3T Philips MRI scanner; (b) normalized SNR of the MR images. The maximum observed SNR variation was 4.7%.

Fig. 8.

(a) Experimental setup of the robot placed in a 3T Philips MRI scanner with a phantom box placed on the moving platform. The phantom box is filled with ultrasound coupling gel; (b) top view of the phantom motion experiment; the red markers indicate the desired phantom position, the blue markers indicate the measured position of the phantom, and the green markers indicate the position of the moving platform calculated offline using the forward kinematics and the recorded encoder data.