Melanie Kolditz

Actuator Design for Stabilizing Single Tendon Platforms

In this paper we illustrate the feasibility of using control moment gyroscopes (CMGs) for the stabilization of free swinging robots hanging from single tendons. Such systems may provide robotic workspaces of unprecedented size, especially in the vertical. Taking typical base reaction forces of industrial robots we show that control moment gyroscopes may provide means for compensation. From the basic principles of CMGs we derive design criteria for a free swinging robot platform. These criteria are illustrated in the design of a scissored pair CMG for a single DoF demonstrator.

Lidar-based measurement of surface roughness features of single tree crowns

In remote sensing data, trees have a low interspecies variability and show a high variability within the tree species. Therefore, specific features that distinguish between unique properties of two tree species are required for a single tree based genera classification. To improve classification results, the suitability of seven surface roughness features, calculated on single tree crown regions, is studied. The algorithms developed to provide roughness parameters can be validated and prototyped in a Virtual Forest testbed. The features are extracted from a normalized digital surface model with a resolution of 0.4m per pixel. Within the test area of 340km2 more than 4000 single trees of eleven different species and additionally 200 buildings are available as reference data. Technical standards define several parameters to describe surface properties. These roughness features are evaluated in the context of single tree crowns. All of these features are based on the deviation of the height values of the tree crown to its mean height. As an additional feature the relationship between the crowns surface area and its occupied ground area is used. The evaluation results of these features regarding the discrimination of tree species on different levels - eleven single tree species, seven tree classes, deciduous and coniferous - and also towards discrimination of trees from buildings will be presented.

Isokinematic leg extension training with an industrial robot

Resistance training of the leg extensor muscles is an important intervention in rehabilitation and prevention of musculoskeletal disorders such as hip or knee arthrosis and osteoporosis. With current training equipment, neither the exercise trajectory can be optimized nor the loadings on structures of the musculoskeletal system can be controlled. To overcome these limitations an experimental research platform for the development of new training scenarios is developed using an industrial robot for maximum flexibility together with kinetic and kinematic data and musculoskeletal models for estimating loadings on target structures. The focus of this paper lies on the implementation of isokinematic exercise, i.e. leg extension and flexion with constant velocity. A force triggered trajectory with smooth transitions between two points needs to be planned for the robot. An algorithm which uses continuous polynomials is proposed. It consists of three parts. First, the trajectory is planned in Cartesian space by intuitive definitions of e.g. start and end point or desired velocity and minimum resistive force. The trajectory can be visualized and optimized using OpenSim together with a model of the research platform, which makes the system usable for non experts in the field of robotics. Second, a smooth trajectory in joint space is generated from the planning points, using a third order polynomial for joint velocities between two adjacent points. Third, the trajectory is adapted to the measured force at the end effector, as the robot should only move along the trajectory, if the applied force by the user is high enough. The proposed algorithm is furthermore easily expandable to arbitrary force triggered motions with definable position and velocity profiles.

Evaluation of foot position and orientation as manipulated variables to control external knee adduction moments in leg extension training

BACKGROUND AND OBJECTIVE Effective leg extension training at a leg press requires high forces, which need to be controlled to avoid training-induced damage. In order to avoid high external knee adduction moments, which are one reason for unphysiological loadings on knee joint structures, both training movements and the whole reaction force vector need to be observed. In this study, the applicability of lateral and medial changes in foot orientation and position as possible manipulated variables to control external knee adduction moments is investigated. As secondary parameters both the medio-lateral position of the center of pressure and the frontal-plane orientation of the reaction force vector are analyzed. METHODS Knee adduction moments are estimated using a dynamic model of the musculoskeletal system together with the measured reaction force vector and the motion of the subject by solving the inverse kinematic and dynamic problem. Six different foot conditions with varying positions and orientations of the foot in a static leg press are evaluated and compared to a neutral foot position. RESULTS Both lateral and medial wedges under the foot and medial and lateral shifts of the foot can influence external knee adduction moments in the presented study with six healthy subjects. Different effects are observed with the varying conditions: the pose of the leg is changed and the direction and center of pressure of the reaction force vector is influenced. Each effect results in a different direction or center of pressure of the reaction force vector. CONCLUSIONS The results allow the conclusion that foot position and orientation can be used as manipulated variables in a control loop to actively control knee adduction moments in leg extension training.

Robotergestütztes System für ein verbessertes neuromuskuläres Aufbautraining der Beinstrecker

Zusammenfassung Neuromuskuläres Aufbautraining der Beinstrecker ist ein wichtiger Bestandteil in der Rehabilitation und Prävention von Muskel-Skelett-Erkrankungen. Effektives Training erfordert hohe Muskelkräfte, die gleichzeitig hohe Belastungen von bereits geschädigten Strukturen bedeuten. Um trainingsinduzierte Schädigungen zu vermeiden, müssen diese Kräfte kontrolliert werden. Mit heutigen Trainingsgeräten können diese Ziele allerdings nicht erreicht werden. Für ein sicheres und effektives Training sollen durch den Einsatz der Robotik, Sensorik, eines Regelkreises sowie Muskel-Skelett-Modellen Belastungen am Zielgewebe direkt berechnet und kontrolliert werden. Auf Basis zweier Vorstudien zu möglichen Stellgrößen wird der Aufbau eines robotischen Systems vorgestellt, das sowohl für Forschungszwecke als auch zur Entwicklung neuartiger Trainingsgeräte verwendet werden kann.