In this journal publication we presented the successful development of the new actuation approach by constructing and testing a 2nd generation prototype in the MRI bore. The evaluation in the context of position control and telerobotic applications showed that the parallel approach eliminated the deleterious effects of the USM velocity dead-zone and velocity reversal time delay while maintaining the high-bandwidth capabilities of the underlying USM.
USM Limitation – Velocity Dead Zone
Ultrasonic motors can provide a linear, one-to-one relationship between desired and actual output velocity if a suitable velocity control approach is implemented. However, due to the USM’s physical limitations, it generally cannot be driven below a minimum threshold velocity, typically 10% of the maximum output velocity. This creates a nonlinear velocity-dead zone about zero velocity shown in the figure. In addition, there is commonly a time delay associated with velocity reversals. These characteristics make high-bandwidth control using linear design techniques challenging.
Proposed Actuation Method
To overcome the limitations of USMs, a new actuation concept is proposed that combines the output motion of two, parallel USMs through the use of a differential mechanism. The differential mechanism acts as a motion summer, where the output is a linear combination of the two parallel USM’s motion. Using this approach, the velocity of each USM can be constrained to the linear region of operation, avoiding the dead band non-linearity and delay associated with velocity reversals.
Parallel Actuation Testbed
To assess MR-compatibility of the current prototype and to confirm our design assumptions as well as help inform future design changes, a set of simple imaging tests were performed. Specifically, MR-images were acquired using a T1-weighted spoiled gradient echo imaging pulse sequence on a 1.5 T scanner (SignaHDx, General Electric, Milwaukee, WI). Images of a spherical phantom where taken under various conditions.
Overall, the imaging sequences confirmed the MR-compatibility of the USMs used in the prototype, assuming proper shielding of the power electronics was in place. In addition, the imaging tests also confirmed that the electrically conductive structural and drive train materials used in the prototype resulted in image distortion, particularly when the device was placed next to the phantom. Elimination of electrically conductive materials would likely improve the MR-compatibility of the design, particularly in regards to its effect on the homogeneity of the local magnetic field. Structural elements could be replaced with a high strength plastic while the drive train gear components could be replaced with ceramic gears. Plastic gearing could also be used but would likely reduce the drive train stiffness and adversely affect performance.
Figure shows an overview of parallel actuator position control structure. Controller includes proportional control in the outer position loop with inner PI control loop on USM motor velocity