Gräser, Axel

Preferred name

Gräser, Axel

Official Name

Gräser, Axel

Gerade angezeigt 1 - 10 von 16

Some of the metrics are blocked by your
consent settings
Machine Learning for Gait Classification
(2017-12-19)
Wang, Xingchen
;
Frese, Udo
;
Gräser, Axel
;
Gräser, Axel
Machine learning is a powerful tool for making predictions and has been widely used for solving various classification problems in last decades. As one of important applications of machine learning, gait classification focuses on distinguishing different gait patterns by investigating the quality of gait of individuals and categorizing them as belonging to particular classes. The most studied gait pattern classes are the normal gait patterns of healthy people, i.e., gait of people who do not have any gait disability caused by an illness or an injury, and the pathological gait of patients suffering from illnesses which cause gait disorders such as neurodegenerative diseases (NDDs). There has been significant research work trying to solve the gait classification problems using advanced machine learning techniques, as the results may be beneficial for the early detection of underlined NDDs and for the monitoring of the gait rehabilitation progress. Despite the huge development in the field of gait analysis and classification, there are still a number of challenges open to further research. One challenge is the optimization of applied machine learning strategies to achieve better classification results. Another challenge is to solve gait classification problems even in the case when only limited amount of data are available. Further, a challenge is the development of machine learning-based methods that could provide more precise results to evaluate the level of gait quality or gait disorder, in contrast of just classifying gait pattern as belonging to healthy or pathological gait. The focus of this thesis is on the development, implementation and evaluation of a novel and reliable solution for the complex gait classification problems by addressing the current challenges. This solution is presented as a classification framework that can be applied to different types of gait signals, such as lower-limbs joint angle signals, trunk acceleration signals, and stride interval signals. Developed framework incorporates a hybrid solution which combines two models to enhance the classification performance. In order to provide a large number of samples for training the models, a sample generation method is developed which could segments the gait signals into smaller fragments. Classification is firstly performed on the data sample level, and then the results are utilized to generate the subject-level results using a majority voting scheme. Besides the class labels, a confidence score is computed to interpret the level of gait quality. In order to significantly improve the gait classification performances, in this thesis a novel feature extraction methods are also proposed using statistical methods, as well as machine learning approaches. Gaussian mixture model (GMM), least square regression, and k-nearest neighbors (kNN) are employed to provide additional significant features. Promising classification results are achieved using the proposed framework and the extracted features. The framework is ultimately applied to the management of patients and their rehabilitation, and is proved to be feasible in many clinical scenarios, such as the evaluation of medication effect on Parkinsona s disease (PD) patientsa gait, the long-term gait monitoring of the hereditary spastic paraplegia (HSP) patient under physical therapy.
Dissertation
705 741
Some of the metrics are blocked by your
consent settings
Practical Brain Computer Interfacing
(2011-06-14)
Valbuena Varon, Diana Alexandra
;
Peleska, Jan
;
Gräser, Axel
;
Gräser, Axel
A brain-computer interface (BCI) is a communication system that enables users to voluntary send messages or commands without movement. The classical goal of BCI research is to support communication and control for users with impaired communication due to illness or injury. Typical BCI applications are the operation of computer cursors, spelling programs or external devices, such as wheelchairs, robots and neural prostheses. The user sends modulated information to the BCI by engaging in mental tasks that produce distinct brain patterns. The BCI acquires signals from the user's brain and translates them into suitable communication. This dissertation aims to develop faster and more reliable non-invasive BCI communication based on the study of users learning process and their interaction with the BCI transducer. To date, BCI research has focused on the development of advanced pattern recognition and classification algorithms to improve accuracy and reliability of the classified patterns. However, even with optimal detection methods, successful BCI operation depends on the degree to which the users can voluntary modulate their brain signals. Therefore, learning to operate a BCI requires repeated practice with feedback that engages learning mechanisms in the brain. In this work, several aspects including signal processing techniques, feedback methods, experimental and training protocols, demographics, and applications were explored and investigated. Research was focused on two BCI paradigms, steady-state visual evoked potentials (SSVEP) and event-related (de-)synchronization (ERD/ERS). Signal processing algorithms for the detection of both brain patterns were applied and evaluated. A general application interface for BCI feedback tasks was developed to evaluate the practicability, reliability and acceptance of new feedback methods. The role of feedback and training was fully investigated on studies conducted with healthy subjects. The influence of demographics on BCIs was explored in two field studies with a large number of subjects. Results were supported through advanced statistical analysis. Furthermore, the BCI control was evaluated in a spelling application and a service robotic application. This dissertation demonstrates that BCIs can provide effective communication for most subjects. Presented results showed that improvements in the BCI transducer, training protocols, and feedback methods constituted the basis to achieve faster and more reliable BCI communication. Nevertheless, expert assistance is necessary for both initial configuration and daily operation, which reduces the practicability of BCIs for people who really need them.
Dissertation
444 147
Some of the metrics are blocked by your
consent settings
Methoden zur adaptiven Benutzerinteraktion bei der semi-autonomen Aufgabenbearbeitung in Rehabilitationsszenarien
(2012-09-05)
Cyriacks, Marco
;
Frese, Udo
;
Gräser, Axel
;
Gräser, Axel
The ever increasing performance of modern computing systems enables the realization of more challenging functionalities in software and mechatronic systems. This tendency results in an increase in system complexity and also makes the operation by users more difficult. Therefore, recent developments are focusing more strongly on the usability of technical systems, especially in case of systems that do not only communicate with users via a user interface but also interact with them physically. Systems that support social reintegration of persons with disabilities, so-called rehabilitation or support robots, fall into this area. This thesis focuses on the development of methods for adaptive user interactions within a software architecture for rehabilitation robots. The objective is the development of a software framework that acts as a basis for the adaptability of the graphical user interface. The methods presented to realize adaptivity are based on a user interface modularization by encapsulating all functionalities into modules. These modules can be activated or deactivated during run-time depending on the availability of resources. Furthermore, a bi-directional communication channel between the user interface and active modules as well as among modules is established. Thus it becomes possible to source common functionality out into modules and to have it reused by other modules. The communication is based on a specification language that has been developed to enable validation and to reach robust run-time behavior. An extensive review of the software architecture used for the target system identified open problems that previously prevented the realization of adaptivity within the user interface. By using another specification language, finding solutions for those open problems becomes possible as well as achieving the set objective. The development is based on an abstraction layer between the user interface and the remaining layers of the software architecture. This realizes full decoupling of the user interface from system specificfunctionality. To proof the concept for adaptivity within the user interface, the implementation of a module integrating an algorithm for pattern recognition is exemplarily shown with the aim to predict future actions of the user by evaluating previous actions.
Dissertation
392 238
Some of the metrics are blocked by your
consent settings
Selective Darkening Filter and Welding Arc Observation for the Manual Welding Process
(2011-10-06)
Hillers, Bernd
;
Michels, Kai
;
Gräser, Axel
;
Gräser, Axel
An optical see-through LCD (GLCD) with a resolution of n x m pixels gives the ability to selectively control the darkening in the welders view. The setup of such a Selective Auto Darkening Filter is developed and its applicability tested. The setup is done by integrating a camera into the welding operation for extracting the welding arc position properly. A prototype of a GLCD taylored for welding is mounted in the welder's view. The extraction of the welding arc position requires an enhanced video acquisition during welding. The observation of scenes with high dynamic contrast is an outstanding problem which occurs if very high differences between the darkest and the brightest spot in a scene occur. The application to welding with its harsh conditions needs the development of supporting hardware. The synchronization of the camera with the flickering light conditions of pulsed welding processes in Gas Metal Arc Welding (GMAW) stabilizes the acquisition process and allows the scene to be flashed precisely if required by compact high power LEDs. The image acquisition is enhanced by merging two different exposed images for the resulting image. These source images cover a wider histogram range than it is possible by using only a single shot image with optimal camera parameters. After testing different standard contrast enhancement algorithm a novel content based algorithm is developed. It segments the image into areas with similar content and enhances these independently.
Dissertation
521 413
Some of the metrics are blocked by your
consent settings
Verfahren zur Optimierung des MSG-Schweißens mit Impulslichtbogen auf der Basis visueller und elektrischer Informationen
(2005-01-27)
Nordbruch, Stefan
;
Thiele, Georg
;
Gräser, Axel
;
Gräser, Axel
The visual observation of the metal transfer and the recording of electrical parameters of the pulsed gas metal arc welding are very important. These methods are used in all phases of research, development and application. Normally, high speed cameras in combination with an additionally lighting unit were used for the visual observation of the metal transfer. However, these systems have some disadvantages for many applications. Furthermore, no methods exist for the analysis of the recorded visual and electrical data.In this work the development and conversion of a concept for the visual online observation of all states of the metal transfer in combination with the simultanious and synchronized measurement of relevant electrical welding process signals (current, voltage, etc.) is described. Additionally, the concept contains the automatic analysis of the recorded data refering to relevant visual parameters of the metal transfer and relevant parameters of the electrical welding process signals. Aim of the concept is the inspection, analysis, optimisation and documentation of the welding process.In detail, this work focuses on the following tasks:* The development of a concept for the visual observation of all states of the metal transfer incombination with the simultanious and synchronized measurement of relevant electrical weldingprocess signals.* The analysis of the high dynamic CMOS technology based on the applicability of differrentCMOS cameras for visual observation of the metal transfer. Important points of the analysiswere the brightness dynamic of the arc, the velocity and the acceleration of the metal transferdrops. For a better evaluation of the characteristic and results of the CMOS technology, it hasben compared with the CCD technology (on the basis of a CCD camera).* The development of methods for the automatic calculation of characteristic parameters out ofthe measured electrical welding signals.* The development of methods for the automatic extraction of visual parameters out of the recordedmetal transfer images.* The development of methods for the automatic preparation and analysis of the caluculated visualand electrical parameters.* The analysis of the developed methods and the entire sensor systems on their ability for a controlof the process using the visual and electrical parameters as basis for future works.
Dissertation
295 898
Some of the metrics are blocked by your
consent settings
Robust Machine Vision for Service Robotics
(2010-01-20)
Grigorescu, Sorin Mihai
;
Anheier, Walter
;
Gräser, Axel
;
Gräser, Axel
In this thesis the vision architecture ROVIS (RObust machine VIsion for Service robotics ) is suggested. The purpose of the architecture is to improve the robustness and accuracy of visual perceptual capabilities of service robotic systems. In comparison to traditional industrial robot vision where the working environment is predefined, service robots have to cope with variable illumination conditions and cluttered scenes. The key concept for robustness in this thesis is the inclusion of feedback structures within the image processing operations and between the components of ROVIS. Using this approach a consistent processing of visual data is achieved.Specific for the suggested vision system are the novel methods used in two important areas of ROVIS: definition of an image ROI, on which further image processing algorithms are to be applied, and robust object recognition for reliable 3D object reconstruction. The ROI definition process, build around the well known "bottom-up top-down" framework, uses either pixel level information to construct a ROI bounding the object to be manipulated, or contextual knowledge from the working scene for bounding certain areas in the imaged environment. The object recognition and 3D reconstruction chain is developed for two cases: region and boundary based segmented objects. Since vision in ROVIS relies on image segmentation on each processing stage, that is image ROI definition and object recognition, robust segmentation methods had to be developed. As said before, the robustness of the proposed algorithms, and consequently of ROVIS, is represented by the inclusion of feedback mechanisms at image processing levels. The validation of the ROVIS system is performed through its integration in the overall control architecture of the service robot FRIEND. The performance of the proposed closed-loop vision methods is evaluated against their open-loop counterparts.
Dissertation
309 134
Some of the metrics are blocked by your
consent settings
Machine Learning Techniques for Autonomous Multi-Sensor Long-Range Environmental Perception System
(2021-01-28)
Haseeb, Muhammad Abdul
;
Frese, Udo
;
Gräser, Axel
;
Gräser, Axel
An environment perception system is one of the most critical components of an automated vehicle, which is defined as a vehicle where the driver does not require to monitor the vehicle’s behavior and its surroundings during driving. This thesis addresses some of the main challenges in the development of vision-based environment perception methods for automated driving, focusing on railway vehicles. The thesis aims at developing methods for detecting obstacles on the rail tracks in front of a moving train to reduce the number of collisions between trains and various obstacles, thus increasing the safety of rail transport. In the field of autonomous obstacle detection for automated driving, besides recognising the objects on the way, the crucial information for collision avoidance is estimated distances between the vehicle and the recognised objects (e.g. cars, pedestrians, cyclists). With the limited capabilities of current state-of-the-art sensor-based environment perception approaches, it is unrealistic to detect distant objects and estimates the distance to them. Mid-to-long-range obstacle detection system is one of the fundamental requirements for heavy vehicles such as railway vehicles or trucks, due to required long braking distance. However, this problem is unaddressed in the computer vision community. The emphasis of this thesis is on the development of robust and reliable algorithms for real-time vision-based mid-to-long-range obstacle detection. In this thesis, the algorithms for obstacle detection from single cameras were developed and evaluated on images captured from RGB, Thermal and Night-Vision Cameras. The developed algorithms are based on advanced machine/deep learning techniques. The development of machine-learning-based algorithms was supported by a novel mid-to-long-range obstacle detection dataset for railways that is proposed in the thesis, which compiles annotated images with the object class, bounding box, and ground truth distance to the object. The developed novel methods for autonomous long-range obstacle detection, tracking, and distance estimation for railways were evaluated on real-world images, which were recorded in different illumination and weather conditions by the obstacle detection system mounted on a static test-bed set-up on the straight rail track and as well on a moving train. Although the focus is on railways, the developed algorithms are also capable to use for road vehicles, hence evaluated on the images of road-scene captured by a camera mounted on moving cars.
Dissertation
522 554
Some of the metrics are blocked by your
consent settings
Gaze-Based Control of Robot Arm in Three-Dimensional Space
(2018-03-01)
Alsharif, Shiva
;
Anheier, Walter
;
Gräser, Axel
;
Gräser, Axel
Eye tracking technology has opened up a new communication channel for people with very restricted body movements. These devices had already been successfully applied as a human computer interface, e.g. for writing a text, or to control different devices like a wheelchair. This thesis proposes a Human Robot Interface (HRI) that enables the user to control a robot arm in 3-Dimensional space using only 2-Dimensional gaze direction and the states of the eyes. The introduced interface provides all required commands to translate, rotate, open or close the gripper with the definition of different control modes. In each mode, different commands are provided and direct gaze direction of the user is applied to generate continuous robot commands. To distinguish between natural inspection eye movements and the eye movements that intent to control the robot arm, dynamic command areas are proposed. The dynamic command areas are defined around the robot gripper and are updated with its movements. To provide a direct interaction of the user, gaze gestures and states of the eyes are used to switch between different control modes. For the purpose of this thesis, two versions of the above-introduced HRI were developed. In the first version of the HRI, only two simple gaze gestures and two states of the eye (closed eyes and eye winking) are used for switching. In the second version, instead of the two simple gestures, four complex gaze gestures were applied and the positions of the dynamic command areas were optimized. The complex gaze gestures enable the user to switch directly from initial mode to the desired control mode. These gestures are flexible and can be generated directly in the robot environments. For the recognition of complex gaze gestures, a novel algorithm based on Dynamic Time Warping (DTW) is proposed. The results of the studies conducted with both HRIs confirmed their feasibility and showed the high potential of the proposed interfaces as hands-free interfaces. Furthermore, the results of subjective and objective measurements showed that the usability of the interface with simple gaze gestures had been improved with the integration of complex gaze gestures and the new positions of the dynamic command areas.
Dissertation
1006 616
Some of the metrics are blocked by your
consent settings
Joint Trajectory Generation and High-level Control for Patient-tailored Robotic Gait Rehabilitation
(2017-05-23)
Focke Martinez, Santiago
;
Frese, Udo
;
Gräser, Axel
;
Gräser, Axel
This dissertation presents a group of novel methods for robot-based gait rehabilitation which were developed aiming to offer more individualized therapies based on the specific condition of each patient, as well as to improve the overall rehabilitation experience for both patient and therapist. A novel methodology for gait pattern generation is proposed, which offers estimated hip and knee joint trajectories corresponding to healthy walking, and allows the therapist to graphically adapt the reference trajectories in order to fit better the patient's needs and disabilities. Additionally, the motion controllers for the hip and knee joints, mobile platform, and pelvic mechanism of an over-ground gait rehabilitation robotic system are also presented, as well as some proposed methods for "assist as needed" therapy. Two robot-patient synchronization approaches are also included in this work, together with a novel algorithm for online hip trajectory adaptation developed to reduce obstructive forces applied to the patient during therapy with compliant robotic systems. Finally, a prototype graphical user interface for the therapist is also presented.
Dissertation
381 270
Some of the metrics are blocked by your
consent settings
Programming by demonstration in the application of rehabilitation robots
(2005-07-11)
She, Haiying
;
Zhang, Jianwei
;
Gräser, Axel
;
Gräser, Axel
The rising number of elderly and disabled persons leads to the increasing demand for reha-bilitation robots for the support in daily life actions. The major factor limiting the application of robots in this domain is the programming of the robot since a detailed knowledge of the daily life activities as well as an effective mechanism to deal with the diverse environment is not available. The most promising way to alleviate this problem is to take advantage of the human experience and skills in performing daily activities and to transfer them to the robot. For this, the concept of Programming by Demonstration (PbD) is believed to be an intuitive and preferable solution. However, classic PbD methods will fail in the situations where the environment in which the task is to be performed differs from that in the demonstration, since important parts of the human skill are lost in that programming process. This work proposes a comprehensive solution for this problem, which is implemented and tested for the exemplary daily life action 'pouring a beverage from a bottle to a glass' (pour-in in short). This solution consists of three advanced methods: The parameter-based trajectory generation, the inclusion of the closed control loop into the PbD process and the implementa-tion of an anti-dripping movement into a pour-in. In the parameter-based trajectory genera-tion, human demonstration is used to obtain the characteristics of a pour-in trajectory repre-senting the human skills. This information is used to autonomously generate a general pour-in trajectory that is optimized to the current task conditions (e.g. the size of bottle and glass). This general pour-in trajectory is then processed by a closed control loop that manages other task restrictions like the target filling liquid. Based on the concept of cascade control loops two control systems reflecting different control strategies of human beings in a pour-in are designed in this work: The first is the three-point based cascade weight/flow rate control (TCC), which applies the human strategy to control the flow rate within minimum and maxi-mum limits. These limits depend on the filling ratio of the liquid. The second is the monitored cascade weight/flow rate control (MCC) based on the imitation of a particularity of the human demonstration, which is the unidirectivity of the movement. Experiments with the rehabilita-tion robot system FRIEND reveal that the lager delay time of the system influences the con-trol accuracy considerable due to the corresponding delays in the system response. Therefore, a number of strategies are developed and implemented into the control loops to dampen the impact of the time delay. Finally, the pour-in process is accomplished and optimized by the inclusion of an anti-dripping movement, whose implementation is based on human demon-strations. By the application of particular trajectory transformations it is possible to apply only one anti-dripping trajectory to any pour-in process. Experiments prove that a robust task execution under various task conditions is achieved to-gether with a high accuracy in reaching the target filling weight. This shows that the applica-tion of these advanced PbD methods for the implementation of pour-in processes into a reha-bilitation robot is successful and that this comprehensive solution is able to confer the abilities of human autonomous environment adaptation. The promising results qualify the developed methods to serve as a starting point for future implementations of additional daily life tasks.
Dissertation
238 126

Gräser, Axel

Filter

Einstellungen

Sortieren nach

Ergebnisse pro Seite