FT-Control - Robotized Transcranial Magnetic Stimulation - page 82

To overcome these limitations and to allow seamless integration into the
clinical workflow, we implement a Force-Torque (FT) control for the flexible
robotized TMS system to increase usability, safety, and precision:

• Usability: We implement a hand-assisted positioning method for faster and
easier coil placement. In this way, the robot is moved in a hand-guided fashion
by grasping and acting on the coil in an intuitive fashion.
• Safety: We realize an automatic distance adjustment to place the coil gently on
the

head. Furthermore, collision detection for non-contact trajectories is

integrated.
• Precision: We combine the existing motion compensation with a contact pres-
sure control to respond to head movements and maintain coil contact during
stimulation.

In this chapter, we first describe the basic principles of force-torque measurement.
Afterwards, we present a method to calibrate the FT sensor to the robot's end
effector. Furthermore, we introduce the idea of gravity compensation, which
compensates for the tool weight, and present a tool calibration procedure. Addi-
tionally, we discuss the impact of the heavy supply cable of the TMS coil on the
force-torque measurements. Subsequently, we describe the implementation of
hand-assisted positioning and contact pressure control. Concluding, we perform an
evaluation using eight different TMS coils, demonstrating that the force-torque
control is suitable for robotized TMS and provides superior patient and user
comfort.

5.1 Basic Principles

An FT sensor with six Degrees of Freedom (DOF) allows measuring forces F in all
three spatial axes and the associated torques M around these axes. The general
relationship between forces and torques can be expressed as the impact of a force
at a certain distance s:

M ¼ F s :

ð 5 : 1 Þ

In this case, s is called the lever arm.
There are different techniques for measurement of forces (and torques). The
most common one is based on strain gauges. On principle, strain gauges constrict
or extend, depending on the load. In this way, the electrical resistance changes
almost linearly. With a calculated calibration, these voltage changes are trans-
ferred into forces and torques. This measurement technique corresponds to the
force-torque sensors utilized in this work.
Mounting an FT sensor to the robot's end effector allows detecting impacts on
the effector. However, two challenges occur for a smooth application. First, to
control the robot based on the detected forces and torques, the transform

E T FT