Handheld Force-

Controlled Ultrasound
Probe
 Introduction
• An hand-held force controlled ultrasound probe has been developed. The
controller maintains a prescribed contact force between the probe and a patient's
body.
• The device will enhance the diagnostic capability of free-hand elastography, swept-
force compound imaging, and make it easier for a technician to acquire repeatable
(i.e. directly comparable) images over time.
• The mechanical system consists of an ultrasound probe, ballscrew-driven linear
actuator, and a force/torque sensor.
• The feedback controller commands the motor to rotate the ballscrew to translate
the ultrasound probe in order to maintain a desired contact force.
• In preliminary user studies it was found that the control system maintained a
constant contact force with 1.7 times less variation than human subjects who
watched a force gauge.
• Users without a visual force display maintained a constant force with 20 times
worse variation.
• In three mock ultrasound examinations one hour apart in which the goal was two
obtain two consistent images at the same force, an unassisted operator obtained
the second image at a 20% lower force, while the operator assisted by the
controller obtained the same force to within <2%.
• The harder the probe is pressed into the body the better the coupling
between probe and tissue and the higher the signal-to-noise ratio
(SNR) of the images.
• Typical ultrasound examinations of the carotid artery require contact
forces of up to 6.4N , and examinations of the abdomen require up to
20N.
Linear Motion Stage for Force
Control
• Six Concepts
• The same functional requirements of safety, speed, low backlash,
backdriveability, and accuracy from the Spherical Motion and Cylindrical
Motion Frames applied in the design of the Force-Controlled Stage.
• Additionally, since the device will be held in a person's hand, it must
compensate for any movement including hand tremors.
• Typical human hand tremor frequency starts at 7-12Hz and slows to 4-
6Hz after 30 minutes of physical activity .
• Thus the actuator will need to be capable of moving at well over 20Hz in
order to be faster than the fastest tremor.
Six Concepts of force controlled sensor

Concept A: "String and pinion." Cable-driven to reduce backlash. Motor capstan

drives cable, which translates the linear DOF.
Concept B: Similar to Concept A, but this configuration allows the motor to be
oriented parallel to the direction of motion, decreasing the size of the device. The
light blue pulleys allow the cable to change directions.
Concept C: Spiral pulleys. Similar to Concept B, but the green pulleys contain
a spiral shape to permit string wander. The motor capstan actuates the red string,
connected to the spiral of the pulleys. The cylindrical parts of the pulleys engage the
purple string, which move the ultrasound probe back and forth.
Concept D: Voice coil actuator. Stationary magnet and moving coil. Current
through the coil causes it to translate [1].
Concept E: Rack and pinion. The rack would need to be preloaded in order to
prevent backlash. [2]
Concept F: Low-backlash ballscrew [3].
Modeling
• The first section models the device when it is not in contact with the
environment, and can be described by a first-order system with
damping.
• Next system is modeled when it is affixed rigidly to a frame and
placed in contact with a tissue phantom. This can be described by a
second-order system.
• In the last section the system is modeled as when it is held in the
hand of the ultrasound technician and placed in contact with the
environment. This interaction is described as 6th-order model.
Position Control
• The rotating mass consists of the motor rotor, coupling, and ballscrew
shaft, which rotate together (we assume for now that the helical-beam
coupling and shaft are infinitely rigid).
• The translating mass is simply the actuator carriage. As the shaft rotates a
given angle, the carriage translates a proportional amount. Since the
screw pitch is 2mm, the carriage translates 2mm each time the screw
rotates once.
• Thus, the motion of the carriage and the screw are coupled. This means
that there is only one independent energy storage element, a translating
mass, which makes this a first-order system
• Since the only way we have of measuring position is from the encoder
output, we constuct a model in the rotational domain. This enables us to
more easily compare the model prediction with the experimental data.
• Since the rotation of the shaft is coupled to the translation of the
carriage we look for a way to represent the two inertias as one.
• A kinetic energy balance enables us to map the translational inertia of
the carriage into an effective rotational inertia by Equation:

• Where Jeff is the effective rotational moment of inertia seen at the

motor, caused by the translation of the carriage of mass m at velocity
v;
• w is the rotational speed ofthe shaft. Since the carriage translational
speed is related to the ballscrew rotational speed by the screw pitch L
(in m/rad), the effective rotational inertia of the carriage Jeff is then
• The total rotational inertia is thus:
Friction
• The rotation of the shaft is impeded by three sources of friction:
1) friction between the shaft and its ball bearings at each end,
2) friction between the shaft and the ballnut contained in the carriage,
3) linear friction between the carriage and the wall.
• We assume that all of these sources of friction can be combined into
one parameter brot.
First-order simulink model to predict position response to step change in desired position when not
in contact.
Thank You