Ahmetcan Erdoğan

We present the design, control, and human–machine interface of a series elastic holonomic mobile platform, AssistOn-Mobile, aimed to administer therapeutic table-top exercises to patients who have suffered injuries that affect the function of their upper extremities. The proposed mobile platform is a low-cost, portable, easy-to-use rehabilitation device targeted for home use. In particular, AssistOn-Mobile consists of a holonomic mobile platform with four actuated Mecanum wheels and a compliant, low-cost, multi-degrees-of-freedom series elastic element acting as its force sensing unit. Thanks to its series elastic actuation, AssistOn-Mobile is highly backdriveable and can provide assistance/resistance to patients, while performing omni-directional movements on plane. AssistOn-Mobile also features Passive Velocity Field Control (PVFC) to deliver human-in-the-loop contour tracking rehabilitation exercises. PVFC allows patients to complete the contour-tracking tasks at their preferred pace, while providing the proper amount of assistance as determined by the therapists. PVFC not only minimizes the contour error but also does so by rendering the closed-loop system passive with respect to externally applied forces; hence, ensures the coupled stability of the human-robot system. We evaluate the feasibility and effectiveness of AssistOn-Mobile with PVFC for rehabilitation and present experimental data collected during human subject experiments under three case studies. In particular, we utilize AssistOn-Mobile with PVFC (a) to administer contour following tasks where the pace of the tasks is left to the control of the patients, so that the patients can assume a natural and comfortable speed for the tasks, (b) to limit compensatory movements of the patients by integrating a RGB-D sensor to the system to continually monitor the movements of the patients and to modulate the task speeds to provide online feedback to the patients, and (c) to integrate a Brain–Computer Interface such that the brain activity of the patients is mapped to the robot speed along the contour following tasks, rendering an assist-as-needed protocol for the patients with severe disabilities. The feasibility studies indicate that AssistOn-Mobile holds promise in improving the accuracy and effectiveness of repetitive movement therapies, while also providing quantitative measures of patient progress.

In this paper, an electroencephalogram (EEG) based Brain-Computer Interface (BCI) is integrated with a robotic system designed to target rehabilitation therapies of stroke patients such that patients can control the rehabilitation robot by imagining movements of their right arm. In particular, the power density of frequency bands are used as features from the EEG signals recorded during the experiments and they are classified by Linear Discriminant Analysis (LDA). As one of the novel contributions of this study, the posterior probabilities extracted from the classifier are directly used as the continuous-valued outputs, instead of the discrete classification output commonly used by BCI systems, to control the speed of the therapeutic movements performed by the robotic system. Adjusting the exercise speed of patients online, as proposed in this study, according to the instantaneous levels of motor imagery during the movement, has the potential to increase efficacy of robot assisted therapies by ensuring active involvement of patients. The proposed BCI-based robotic rehabilitation system has been successfully implemented on physical setups in our laboratory and sample experimental data are presented.

We present a systematic approach that enables
online modification/adaptation of robot assisted rehabilitation
exercises by continuously monitoring intention levels of patients
utilizing an electroencephalogram (EEG) based Brain-Computer
Interface (BCI). In particular, we use Linear Discriminant
Analysis (LDA) to classify event-related synchronization (ERS)
and desynchronization (ERD) patterns associated with motor
imagery; however, instead of providing a binary classification
output, we utilize posterior probabilities extracted from LDA
classifier as the continuous-valued outputs to control a rehabilitation
robot. Passive velocity field control (PVFC) is used as
the underlying robot controller to map instantaneous levels of
motor imagery during the movement to the speed of contour
following tasks. PVFC not only allows decoupling of the task and
the speed of the task from each other, but also does so by ensuring
coupled stability of the overall robot patient system. The proposed
framework is implemented on ASSISTON-MOBILE—a holonomic
mobile platform based series elastic robot, and feasibility studies
with healthy volunteers have been conducted test efficacy and
effectiveness of the proposed approach. Giving patients online
control over the speed of the task, the proposed approach ensures
active involvement of patients throughout exercise routines and
has the potential to increase the efficacy of robot assisted
therapies.

This paper proposes a cross-disciplinary methodology for a fundamental question in product development: How can the innovation patterns during the evolution of an engineering system (ES) be encapsulated, so that it can later be mined through data mining analysis methods? Reverse engineering answers the question of which components a developed engineering system consists of, and how the components interact to make the working product. TRIZ answers the question of which problem-solving principles can be, or have been employed in developing that system, in comparison to its earlier versions, or with respect to similar systems. While these two methodologies have been very popular, to the best of our knowledge, there does not yet exist a methodology that reverseengineers and encapsulates and represents the information regarding the complete product development process in abstract terms. This paper suggests such a methodology, that consists of mathematical formalism, graph visualization, and database representation. The proposed approach is demonstrated by analyzing the design and development process for a prototype wrist-rehabilitation robot.

Passive velocity field control is advantageous to deliver human-in-the-loop contour tracking rehabilitation exercises, since patients can be allowed to proceed with their preferred pace, while assistance can still be provided as determined by the therapist with ensured coupled stability. We introduce a framework based on passive velocity field control for robot assisted rehabilitation that includes prevention mechanisms against undesired slacking behavior of patients. This framework not only provides systematic approaches to prevent slacking, but also can do so while ensuring coupled stability of the overall robot patient system, a property that cannot be assured with any of the other ad-hoc slacking prevention methods. In particular, the proposed approach enables seamless on-line modification of the task difficulty, speed of contour following, and the level of assistance, while preserving passivity of the system with respect to external forces. The proposed slacking prevention schemes encourage active participation of the patients in rehabilitation protocols with even increased number of repetitions and thanks to flexibility introduced by the controller, render delivery of “repetitive tasks without repeating the same task” possible. Experiments with an haptic interface are included to demonstrate the passivity of the proposed control framework and preliminary human subject experiments with healthy volunteers are presented to validate feasibility and usability of the proposed approaches.

In this study, a forearm-wrist rehabilitation robot is designed and controlled. In particular, to comply with the rotational workspace of the human forearm-wrist, a 3RPS-R mechanism is proposed as the underlying kinematic structure of the device. The equations governing kinematics of the 3RPS-R mechanism are derived using Euler parameters (unit quaternions) to avoid representation singularities and the device is controlled using quaternion feedback to ensure the geometrical consistency of resulting interactions. To this end, first the reference trajectories for the human forearm-wrist rotations are calculated as geodesics using spherical linear interpolation. Next, an impedance controller is implemented with proper orientation error in SO(3) for assistance and dynamic interaction. In order to motivate the active participation of the patient, a contour tracking algorithm is used with properly defining a potential field in the same manifold of the end-effector space and a passive velocity field controller is synthesized. The controllers are implemented on the forearm-wrist rehabilitation robot and experimental results are presented.

This paper presents design, implementation and control of a 3RPS-R exoskeleton, specifically built to impose targeted therapeutic exercises to forearm and wrist. Design of the exoskeleton features enhanced ergonomy, enlarged workspace and optimized device performance when compared to previous versions of the device. Passive velocity field control (PVFC) is implemented at the task space of the manipulator to provide assistance to the patients, such that the exoskeleton follows a desired velocity field asymptotically while maintaining passivity with respect to external applied torque inputs. PVFC is augmented with virtual tunnels and resulting control architecture is integrated into a virtual flight simulator with force-feedback. Experimental results are presented indicating the applicability and effectiveness of using PVFC on 3RPS-R exoskeleton to deliver therapeutic movement exercises.

A new approach to online generation of velocity fields for parametric curves is presented for implementation in passive velocity field controllers that enable human-in-the-loop contour following tasks. A feedback stabilized closest point tracking algorithm is utilized for real-time determination of the contour error and online construction of the velocity field. The algorithm augments the system dynamics with a new state, and implements a uniformly asymptotically stable controller to update this new state to continual track the closest point to the robot end-effector. Applicability and effectiveness of the approach have been demonstrated through simulations and experiments.

We consider a housekeeping domain with multiple cleaning robots and represent it in the action language C+. With such a formalization of the domain, a plan can be computed using the causal reasoner CCalc for each robot to tidy some part of the house. However, to find a plan that characterizes a feasible trajectory that does not collide with obstacles, we need to consider geometric reasoning as well. For that, we embed motion planning in the domain description using external predicates. For safe execution of feasible plans, we introduce a planning and monitoring algorithm so that the robots can recover from plan execution failures due to heavy objects that cannot be lifted alone. The coordination of robots to help each other is considered for such a recovery. We illustrate the applicability of this algorithm with a simulation of a housekeeping domain.

We explore the effects of parameters constituting a second order dynamic system on the rate of human motor adaptation while performing a rhythmic dynamic task. In our experiments, participants excite virtual second-order systems at resonance via a haptic interface. After overtraining subjects with a nominal system, we unexpectedly change the system parameters and study the resulting motor adaptation in catch trials. Through four experiment seatings, we demonstrate the effects of dynamic system parameters on human motor adaptation. Results indicate that gain and damping parameters significantly affect the rate of adaptation. In particular, as the effort required to complete the task increases, the rate of adaptation decreases, indicating a trade-off between task performance and the effort required to perform the task.

We embed knowledge representation and automated reasoning in each level of the classical 3-layer robot control architecture, in such a way as to tightly integrate these layers. At the high-level, we represent not only actions and change but also commonsense knowledge in the action description language C+. Geometric reasoning is lifted to the high-level by embedding motion planning in the domain description, using external predicates. Then a discrete plan is computed for each robot, using the causal reasoner CCALC. At the mid-level, if a continuous trajectory is not computed by a motion planner because the discrete plan is not feasible at the continuous-level, then formal queries are asked to the causal reasoner to find a different plan subject to some (temporal) conditions represented as formulas. At the low-level, if the plan execution fails, then a new continuous trajectory is computed by a motion planner at the mid-level or a new discrete plan is computed using an automated reasoner at the high-level. We apply this tightly integrated robot control architecture in a housekeeping domain with multiple autonomous robots, and illustrate this application with a simulation.

We formalize actions and change in a housekeeping domain with multiple cleaning robots, and commonsense knowledge about this domain, in the action language C+. Geometric reasoning is lifted to high-level representation by embedding motion planning in the domain description using external predicates. With such a formalization of the domain, a plan can be computed using the causal reasoner CCalc for each robot to tidy some part of the house. We introduce a planning and monitoring algorithm for safe execution of these plans, so that it can recover from plan failures due to collision with movable objects whose presence and location are not known in advance or due to heavy objects that cannot be lifted alone. We illustrate the applicability of this algorithm with a simulation of a housekeeping domain.

This paper proposes a multi-lateral shared control concept for robot assisted rehabilitation. In particular,
a dual-user force-feedback teleoperation control architecture is implemented on a forearm-wrist rehabilitation
system consisting of two kinematically dissimilar robotic devices. The multi-lateral rehabilitation system
allows for patients to train with on-line virtual dynamic tasks in collaboration with a therapist. Different
control authority can be assigned to each agent so that therapists can guide or evaluate movements of
patients, or share the control with them. The collaboration is implemented using a dual-user force-feedback
teleoperation control architecture, in which a dominance factor determines the authority of each agent in
commanding the virtual task. The effectiveness of the controller and regulation of the dominance for each
agent is experimentally verified.

This paper deals with model based fault detection and isolation of a pickling process within the steel industry. The model is based on the grey box methodology and reflects the physical behaviour of the process. Possible faults are included in the model as parameters, which are estimated on line. The estimation is based on minimizing a loss function using past data from a defined moving time window. The procedure of finding the faults starts by estimate all defined fault parameters. One fault parameter is removed from the set of prospective list of faults by removing the parameter with the smallest saliency. The saliency is defined as the quote between the parameter estimate and the corresponding element of the inverse of the hessian matrix. The parameter with the smallest saliency gives a measure of the relevance of the estimated parameters relative all estimated parameters. The procedure is repeated until all fault parameters are eliminated from the list. To isolate the faults, the Akaike’s Information Criterion (AIC) is used to detect faults. This gives the threshold when a fault relevant parameter is removed from the list of prospective faults.

This paper presents the design and the optimal dimensional synthesis of a reconfigurable, parallel mechanism based, force feedback exoskeleton for the human ankle. The primary use for the device is aimed as a balance/proprioception trainer, while the exoskeleton can also be employed to accommodate range of motion (RoM)/strengthening exercises. Multiple design objectives for several physical therapy exercises are discussed and classified for the design of the rehabilitation device. After kinematic structures of mechanisms that are best suited for the physical therapy applications are selected, an optimization problem to study the trade-offs between multiple design criteria is formulated. Dimensional synthesis is performed to achieve maximum stiffness of the device, while simultaneously maximizing the actuator utilization, using a Pareto-front based framework. An “optimal”�design is selected studying the Pareto-front curve and taking the primary and secondary design criteria into account. Once the dimensions of the device are decided, reconfigurability is built into the design such that the device can be arranged as a 3UPS manipulator to administer RoM/strengthing exercises and as a 3RPS manipulator to support balance/proprioception training. Metatarsophalangeal joint exercises are also enabled through a reconfigurable design of the foot plate. Details of the final design of the ankle exoskeleton and CAD snapshots of the device are also presented.

This paper presents the design, analysis, and a clinical application of a reconfigurable, parallel mechanism based, force feedback exoskeleton for the human ankle. The device can either be employed as a balance/proprioception trainer or configured to accommodate range of motion (RoM)/strengthening exercises. The exoskeleton can be utilized as a clinical measurement tool to estimate dynamic parameters of the ankle and to assess ankle joint properties in physiological and pathological conditions. Kinematic analysis and control of the device are detailed and a protocol for utilization of the exoskeleton to determine ankle impedance is discussed. The prototype of the device is also presented.