**2.3.1 Sensor / actuator design approach**

Previously we have developed a haptic tweezer for force feedback microgripping. Different from devices like the PHANTOM of the previously tested commercial force feedback joysticks, where the position of the inpout device is sensed and then a torque/force is applied with a motor, forces on the microgripper are sensed with strain gauges and the tweezer arms are actively moved with a microservo. The closed loop control does not only allow a precise control of the forces excerted by the haptic tweezer to the fingers, but also the behavior of the device can be programmed. In the easiest form, the "virtual" spring constants of the tweezer can be set.

The main idea for the R&D of our force feedback joystick was to use the same sensing/actuation principle we used for the haptic tweezer for a device with more DOFs. A first 1DOF demonstrator was built consisting of a high torque servo and a 12 cm PEEK handle mounted on the servo actuator. The diameter of the handle was 12mm, but reduced to a square profile of 5x5 mm² in a 15mm long section close to the *axis of rotation of the handle*  in the high strain area. Here 2 strain gauges were positioned in this area.

A microcontroller (Atmel ATmega32L) was used to read out the strain gauges and control the servo. A closed loop control was implemented allowing to


144 Haptics Rendering and Applications

Fig. 9. Multifunctional catheter tip for impedance and optical spectroscopy with integrated

In the first instance, the steerable catheter described in 2.2. was actuated with a commercial available low cost force feedback joystick. This kind of joysticks is typically used for playing computer games like flight simulators to achieve a more realistic gaming experience. The position of the joystick is sensed with potentiometers and haptic force feedback is generated by DC motors. An interface was developed under National Instruments Labview to read in the joystick coordinates and to apply forces with the DC motors. First tests revealed that the precision of the force feedback motors is rather insufficient and cannot be precisely controlled and the maximum torque is rather low. A precise force feedback comparable to operator experiences when using an expensive haptic input device like the Phantom is not

Previously we have developed a haptic tweezer for force feedback microgripping. Different from devices like the PHANTOM of the previously tested commercial force feedback joysticks, where the position of the inpout device is sensed and then a torque/force is applied with a motor, forces on the microgripper are sensed with strain gauges and the

**2.3 Research and development of a haptic force feedback joystick** 

For the control of the catheter a device is necessary that allows:



suction fixature

possible with this kind of joystick.



**2.3.1 Sensor / actuator design approach** 

Fig. 10. Torque on the handle of the actuator

Simple scenarios were successfully tested, including the simple spring model (the torque on the handle increases with the angle), zero force actuation (a minimum force is required to move the handle) and mixtures of both. The latter was especially important as it reflects the requirements for the catheter control. When the catheter tip is freely moving without touching, zero actuation force is used. Alternatively, one might consider to program a low

Sensorized Tools for Haptic Force Feedback in Computer Assisted Surgery 147

a b

Fig.13a shows the electronic control board which is fixed in a cavity under the base plate. This cavity was milled out of the bottom of the base plate. The control system purpose is to control the servo motors to give a resistive force to the user which is determined by the hand movements of the user and the control algorithms used by the microcontroller to vary the user experience. The strain gauges are in a Wheatstone bridge configuration and the differential voltages are fed to the voltage amplifiers (2 x AD623, #3). The amplified voltages (which represent the resistive force felt by the user) are processed by the ATMEL ATmega32 microcontroller (#2) which through a specific algorithm then send digital commands to the servo motors (#4) to continually adjust this force. The force sensed by the user and the position of the control stick in both X & Y axes can be communicated to a PC by the microcontroller. The microcontroller transmits the data to the RS232 chip (#1) which communicates to the external device connected to #6 with the RS232 protocol. The board is

a b

Fig.13b shows a PC user interface developed using Labview which was used to test the performance of the input device in registering both control stick position (#1) and force (#2) and also using the PC interface to change the stiffness (resistance) of the control stick sensed

Fig. 12. a. Fabricated joystick prototype; b. Strain gauges and connector on control stick

powered with a regulated 6V supply connected at #5.

Fig. 13. a. Electronic control board; b. Graphical User Interface

spring constant to move the joystick back to central position when released. When at a certain angle the catheter touches for instance a wall, a force is sensed in the bowden wires. This force signal is then used to generate a force feedback signal. Figure 10 shows the torques applied by the servo motor in these 3 scenarios. In the simple spring model the torque is directly dependent on the angle (0.5 N cm/°). In the zero actuation force model practically no force is required to move the handle. If the handle is released, it stays at the last position. In the mixed model, a low spring constant was programmed for 0-25°, above that a spring constant of 10Ncm/° was implemented to simulate touching. If the handle is released, it returns to the central position.
