About 60 percent of the Japanese causes of death are the neoplasm (cancer), the cardiac disease (cardiopathy), and the brain blood vessel disease (apoplexy), which are the three major lifestyle-related illnesses.(right fig)
Among these diseases, the main causes of a cardiac disease and a brain blood vessel disease are arteriosclerosis. Although arteriosclerosis advances with aging, it is influenced by the custom and home environment of everyday life. Therefore, the age of the onset of arteriosclerosis vary depending on individuals.
　It is very important to detect of the arteriosclerosis at an early stage when more option of medical treatment is available.
　There are some diagnostic method for blood vessels, such as X ray, CT, an ultrasound (B-mode and Doppler), etc.
　There is a newly proposed diagnostic method called the wave intensity（WI）measurement. It diagnoses the advance degree of arteriosclerosis by computing the index which shows the contraction power of the heart which measures the blood pressure and the blood flow velocity of carotid artery using ultrasound diagnostic equipment.
　Our research project is related to the measurement of this WI, by using the robot system.
　On the other hand, abdominal ultrasound is effective in early detection of the neoplasm (cancer) with which about 30 percent of the Japanese have passed away. Therefore we are also doing research and development of a robotic-assisted system providing support to doctors while performing the abdominal ultrasound.（WTA-2）

　Wave Intensity is an hemodynamic index which originally Paker and Jones proposed as an index which can compare the influence of the forward and backward waveform of the blood pressure. Although WI which came to be called such from it being the same unit as the intensity of a sound wave attracted attention only from the physical meaning, superiority or inferiority of the influence of an advance wave and a reflective wave, which the mark of the positive/negative has it becomes clear that the size of WI has an important physiological meaning by the research of Professor Sugawara of the Himeji Dokkyo University, and it attracts attention as a new indicator of blood-flow dynamics.
　WI is calculated by the product of the time differentiation value of the blood pressure P and the blood flow velocity U.

Approximate adjustment
　(One-touch switching of fixation and release,Passive)
　 ・Positioning arm
　 ・Probe holding device
Fine adjustment
　(Master slave operation,Active)
　 ・Probe holding manipulator（Slave）
　 ・Probe controller（Master）
　 ・Motor driver

　The probe holding manipulator has 6-DOF, linear parallel link mechanism, and a probe holding device is attached in the end. It is similar to the Stewart platform. The different point is that the actuator part is not inside the link. By this configuration, the link diameter can be smaller, which reduces the interference between links.
　For the practical use in the medical treatment site, we are required to optimize the size of the whole manipulator and improve the positioning accuracy of the manipulator. We have applied the genetic algorithm for optimization of link position. Each components are eliminated in order to reduce the backlash for the improvement of positioning accuracy.
　The overall size is 96[mm] in width, 74[mm] in height, 384〜443[mm] in length, and weight is approximately 2.9[kg].

Actuator part

　The actuator part is composed of the DC motor and the ball screw, and the ball spline. The link part is composed a linear actuator (1-DOF), universal joint (2-DOF), and spherical rolling joint (3-DOF). The thickness of the link is 4[mm], and a movable angle of universal joint and spherical rolling joint are 45 [deg].

　The objective of the probe holding device is to fix the ultrasound probe to the end-effector. 1-DOF rotational joint is required for adjusting initial angle between probe and manipulator. The probe holding device is composed of the gear and the board spring, and can be fixed by the board spring engages in the gear. The movable angle of X-axis rotation is ±45[deg] with step of 10[deg] which is the distance of each teeth of the gear.
　Also the probe holding device has Z-axis rotation 90[deg] by ratchet mechanism for scanning the cross section image of carotid artery.

1. The longitudinal section of the carotid artery is clearly observed
2. The centre of the image is 20[mm] from the bifurcation of the carotid artery
3. Inclination of the artery in the image is about 5 [deg.]
4. Intima (most inner layer of the artery walls) is clearly observed

Algorithm

1. scan pixels along with the vertical lines set at regular interval, then extract the dark gap by applying intensity threshold. (Fig (a))
2. connect those dark gap horizontally taking into account the horizontal continuity. (Fig.(b))

　This is required for confirming that the probe is at the exact center of the artery. The artery wall has 3 layers, adventitia, media intima (from outside to inside). In the ultrasound image, the adventitia and intima are shown as white lines, and media is a dark line. However, intima is observed only when the ultrasound beam hits the artery walls in right angle. That is why the intima observation is used for the confirmation of the right positioning of the probe. Diameter of the artery is tracked during the WI measurement, therefore slight deviation from the center of the artery causes an inaccuracy of the measurement. The algorithm of the intima detection is developed based on the intensity gradient of the pixels around the detected artery walls.

We have designed the sequential pattern of the probe movement as follows;

1. Defining the points on the patient neck
　Probe is moved downwards until it detects the surface of the patient skin, then adjust the orientation (as below 2.), and records the coordinate in 6 axes. After that, probe is moved upward, forward, downward for 3 times. Total of 4 points are recorded.

2. Adjustment of the probe orientation
　If the black zone at the edge of the image (because the probe does not fully contact with skin) is detected, the probe is rotated towards the black zone to fit the probe.

3. Scanning
　The recorded 4 points are interpolated and probe scans the patient neck following the interpolated line. The scanning stops when the carotid artery is detected in the image.

4. The vein is often detected as the carotid artery because the appearance in the ultrasound image is similar to the carotid artery. 　The robot presses the patient neck by probe slightly, then observes the change of the diameter. If it is the vein, the diameter significantly changes because of the lower inner pressure than the carotid artery. (see the figure below)

5. Detection of the bifurcation
　Only this sequence is done by operator. Ultrasound The operator commands the moving direction of the probe. During the movement if the carotid artery is missing, the robot repeats the scanning in small area and keep the probe on the carotid artery.

6. Inclination of the artery in the image
　The angle of the artery in the image can be measured from the detected contour of the artery walls. The probe is rotated until the angle of the artery reaches 5 [deg.]

7. The probe scans the small area so that the robot guides the probe towards the position where the clearest image of the intima is observed.

8. Detection of the missing intima
　If the decrease of the intima score is detected (caused by the movement of the patient, for example), the sequence #7 is repeated.

The following video shows the full sequence of the automated positioning.

　The objective of positioning accuracy experiment is to verify the positioning accuracy of new manipulator by using optical 3D tracker "Optotrack".
　Round trip command with a small displacement in 6 axes were given to the end-effector at 3 point in the workspace. The maximum error from the desired trajectory was measured. The round trip command in translational X,Y and Z-axes are ±5[mm]. The ones in rotational X-axis was ±7.5[deg], Y-axis ±5[deg], and Z-axis is ±10[deg] from start position.
　Measured absolute deviation from the target trajectory are shown in right figure. "New model" represent WTA-2R, "Old model" represents our previous model "WTA-2".

　The objective is operability experiment is to measure the time consumption of the probe positioning and to obtain the sonographers' opinions. 5 sonographers tried to position the probe on the carotid artery with this robot (WTA-2R) and the old model (WTA-2). The time consumption for the positioning both
　(1)Rough positioning by using passive arm
　(2)Precise positioning by using manipulation
were recorded. After the positioning task, the Sonographers filled in the questionnaires about operability etc. The result is as the following i), ii).