Ernur Karadogan

A stiffness discrimination experiment including analysis of palpation forces and velocities.

Introduction: The incorporation of haptics, the sense of touch, into medical simulations increases their capabilities by enabling the users to “feel” the virtual environment. We are involved with haptics-augmented virtual reality training for palpatory diagnosis. We have developed a stiffness discrimination program to train and test users in finding subtle differences in human tissue stiffness for medical diagnoses. In this article, we studied the effect of surface stiffness on the stiffness discrimination task and analyzed the palpation force and speed during haptic exploration. Methods: The ability to discriminate stiffness differences was studied by means of a psychophysical experiment with 13 second-year medical students (eight women and five men). Subjects were asked to identify the stiffer of two virtual computer-generated surfaces (top surfaces of two cylinders) by means of a PHANToM Omni (SensAble Inc.) haptic device with a modified stylus to accommodate their fingers. The modification of the stylus provided the mechanical advantage to simulate surface stiffness values that are beyond the original capability of the haptic device. An adaptive two-alternative forced-choice procedure was used on each trial. Palpation velocity and force vectors were recorded directly from the haptic device for further analyses. Weber fraction was determined by using an automated mastery algorithm. Results: Four standard stiffness values (0.25, 0.50, 1.00, and 1.25 N/mm), typical of the stiffness range of human soft tissues, were used as references. The average experimental Weber fractions observed were 0.20, 0.27, 0.26, and 0.30, respectively, with higher Weber fractions corresponding to lower stiffness discrimination ability. At 1.00 and 1.25 N/mm standard stiffness, the correlation analysis for Weber fraction and the palpation speed revealed significant differences (P < 0.05). These differences suggested that the subjects with a higher palpation velocity tended to have a higher Weber fraction. There was no significant difference between male and female subjects. There was no significant difference between subjects new to the haptic device and those who had used it previously. The average amount of force that was applied by the subjects to the standard stiffness side and the comparison stiffness side within the sessions was not significantly different. However, the subjects increased the average force they applied with increasing standard stiffness value across the sessions (P < 0.05). Conclusions: For the four standard stiffness values investigated, 0.25, 0.50, 1.00, and 1.25 N/mm, the resulting average stiffness-discrimination Weber fractions were 0.20, 0.27, 0.26, and 0.30, respectively. The average of the forces applied by the subjects was constant within a single session (with a single standard stiffness value). This average force monotonically increased as the standard stiffness value increased across the sessions. We also found positive correlation between the Weber fraction and the palpation speed in the sessions tested with 1.00 and 1.25 N/mm standard stiffness. This correlation suggested that higher speed is related to lower sensitivity in discrimination of stiffness differences for these two standard stiffness values. Our results are applicable to tasks involving stiffness discrimination between multiple objects.

Design and Analysis of a Four-Pendulum Omnidirectional Spherical Robot

This paper presents the design, analysis, and comparison of a novel four-pendulum spherical robot. The proposed mechanism rolls omnidirectionally via four tetrahedrally-located pendulums that shift the robot’s center of mass to create rolling torque. The nine dynamic equations of motion are derived via the Lagrangian and nonholonomic constraint equations, and then simulated numerically; results show successful propulsion with expected behaviors. The mechanism is then compared to existing center-of-mass designs in terms of directionality, drive torque arm, and inertia eccentricity. In these regards, the four-pendulum design is a balance of existing designs: it is omnidirectional with eccentricity and torque capability that are in the middle of the range exhibited by existing designs. In addition, the new four-pendulum mechanism has been built and tested as a successful proof-of-concept prototype.

The robotic lumbar spine: dynamics and feedback linearization control.

The robotic lumbar spine (RLS) is a 15 degree-of-freedom, fully cable-actuated robotic lumbar spine which can mimic in vivo human lumbar spine movements to provide better hands-on training for medical students. The design incorporates five active lumbar vertebrae and the sacrum, with dimensions of an average adult human spine. It is actuated by 20 cables connected to electric motors. Every vertebra is connected to the neighboring vertebrae by spherical joints. Medical schools can benefit from a tool, system, or method that will help instructors train students and assess their tactile proficiency throughout their education. The robotic lumbar spine has the potential to satisfy these needs in palpatory diagnosis. Medical students will be given the opportunity to examine their own patient that can be programmed with many dysfunctions related to the lumbar spine before they start their professional lives as doctors. The robotic lumbar spine can be used to teach and test medical students in their capacity to be able to recognize normal and abnormal movement patterns of the human lumbar spine under flexion-extension, lateral bending, and axial torsion. This paper presents the dynamics and nonlinear control of the RLS. A new approach to solve for positive and nonzero cable tensions that are also continuous in time is introduced.

Three-Dimensional Static Modeling of the Lumbar Spine

This paper presents three-dimensional static modeling of the human lumbar spine to be used in the formation of anatomically-correct movement patterns for a fully cable-actuated robotic lumbar spine which can mimic in vivo human lumbar spine movements to provide better hands-on training for medical students. The mathematical model incorporates five lumbar vertebrae between the first lumbar vertebra and the sacrum, with dimensions of an average adult human spine. The vertebrae are connected to each other by elastic elements, torsional springs and a spherical joint located at the inferoposterior corner in the mid-sagittal plane of the vertebral body. Elastic elements represent the ligaments that surround the facet joints and the torsional springs represent the collective effect of intervertebral disc which plays a major role in balancing torsional load during upper body motion and the remaining ligaments that support the spinal column. The elastic elements and torsional springs are considered to be nonlinear. The nonlinear stiffness constants for six motion types were solved using a multiobjective optimization technique. The quantitative comparison between the angles of rotations predicted by the proposed model and in the experimental data confirmed that the model yields angles of rotation close to the experimental data. The main contribution is that the new model can be used for all motions while the experimental data was only obtained at discrete measurement points.

A CABLE-ACTUATED ROBOTIC LUMBAR SPINE FOR PALPATORY TRAINING OF MEDICAL STUDENTS

This paper presents the kinematic and pseudostatic analyses of a fully cable-actuated robotic lumbar spine (RLS) which can mimic in vivo human lumbar spine movements to provide better hands-on training for medical students. The design incorporates five active lumbar vertebrae between the first lumbar vertebra and the sacrum, with dimensions of an average adult human spine. Medical schools can benefit from a tool, system, or method that will help instructors train students and assess their tactile proficiency throughout their education. The robotic lumbar spine has the potential to satisfy these needs in palpatory diagnosis. Medical students will be given the opportunity to examine their own patient that can be programmed with many dysfunctions related to the lumbar spine before they start their professional lives as doctors. The robotic lumbar spine can be used to teach and test medical students in their capacity to be able to recognize normal and abnormal movement patterns of the human lumbar spine under flexion-extension and lateral bending. This project focus is on palpation, but the spine robot could also benefit surgery training/planning and other related biomedical applications.

Haptic modules for training in palpatory diagnosis

We have developed and evaluated a novel tool based on haptics and virtual reality technology for augmenting the teaching of palpatory diagnosis. This novel tool illuminates palpatory diagnosis concepts by touch on a laptop PC using affordable haptic interfaces. There are six training modules each targeting a specific aspect of palpation. The difficulty level for all modules is adjusted automatically by measuring users performance in real-time. The haptic interface used in this study was the PHANTOM Omni® (SensAble Tech., Inc.) and it was modified to enable manipulation with only one finger. 22 osteopathic medical students (16 first- and 6 second-year) participated in the evaluation of the system. The majority of the participating students (>;90.9%) thought that future practice with the system may help them develop their palpatory skills. The majority (>;77.3%) of the students also thought that the instructions on the module screens were clear. When the students were asked about the user interface, most of the students (>;86.4%) responded that it was clear and easy to interpret. Evaluation results also showed that when the students were asked whether they would like to use the modules in the future for training at least 90.9% of them answered “Yes” or “Maybe”. The achievement of purpose ratings for individual modules changed between 6.27 and 8.82 on a 10-point scale. This system can be used for unlimited student practice for improving skills from Osteopathic Manipulative Medicine laboratory and also as a repeatable and objective measure of palpatory skill to track student progress.

Dynamics and Control of the Robotic Lumbar Spine (RLS)

This paper presents the dynamics and nonlinear control of the Robotic Lumbar Spine (RLS). The RLS is a 15 degree-of-freedom, fully cable-actuated robotic lumbar spine which can mimic in vivo human lumbar spine movements to provide better hands-on training for medical students. The current design includes five active lumbar vertebrae and the sacrum, with dimensions of an average adult human spine. It is actuated by 20 cables connected to electric motors. Every vertebra is connected to the neighboring vertebrae by spherical joints. Medical schools can benefit from a tool, system, or method that will help instructors train students and assess their tactile proficiency throughout their education. The robotic lumbar spine has the potential to satisfy these needs in palpatory diagnosis. Additionally, a new approach to solve for positive and nonzero cable tensions that are also continuous in time is introduced.Copyright

Image display visualization in teleoperation

For successful teleoperation, an operator must be able to accurately inscribe the desired movement to the manipulandum. The goal of manipulating an image display for teleoperation is to perform visual transformations that minimize mental fatigue, performance time, and enhances the compatibility between the operator and the teleoperated system for successful task completion. A concept based on multiple physical and virtualized displays will identify the areas in which the increasing mental demand reduces the effectiveness of the operator. The effect of physical and virtual display configurations was investigated in an experimental study. A total of 18 participants teleoperated an industrial robot under three display configurations. The results of the study revealed that the virtualization of the display is a more effective way to present information to the operator of a teleoperated system when the goal-to-goal times and the total distance traveled during a task were considered.

Tracking a system of multiple cameras on a rotating spherical robot

For many applications in the operation of a spherical robot, it is necessary to use optical devices to observe, track and monitor the robots surroundings and environment. Estimating the rotation of a multiple camera system is crucial and can also be complex and computationally expensive. Demonstrating the multiple camera system with a shared central point can be simpler and require less computation. In this paper, we will demonstrate that a multiple camera system can be developed, an estimation of the movement of a wheel and sphere can be tracked, and the movement can be observed from a remote location.

Design of a Spherical Robot With Cable-Actuated Driving Mechanism

This paper presents the kinematics and dynamics of a spherical robot with a mechanical driving system that consists of four cable-actuated moving masses. Four cable-pulley systems control four tetrahedrally-located movable masses and the robot functions by shifting its center of mass to create rolling torque. The cable actuation decreases overall mass and, therefore, allow for less energy expenditure, as compared to other moving mass mechanisms that translate the masses by powered-screws. Additionally, the design allows the center of mass for the static (spherical shell, electronics, motors etc.) and dynamic mass (moving masses) to be at the geometric center at any given time, therefore has potential for tumbleweeding when needed. The derived equations of motion are verified by means of simulations.