Axel Graeser

Convolutional Neural Network with embedded Fourier Transform for EEG classification

In BCI (brain - computer interface) systems, brain signals must be processed to identify distinct activities that convey different mental states. We propose a new technique for the classification of electroencephalographic (EEG) steady-state visual evoked potential (SSVEP) activity for non-invasive BCI. The proposed method is based on a convolutional neural network that includes a Fourier transform between hidden layers in order to switch from the time domain to the frequency domain analysis in the network. The first step allows the creation of different channels. The second step is dedicated to the transformation of the signal in the frequency domain. The last step is the classification. It uses a hybrid rejection strategy that uses a junk class for the mental transition states and thresholds for the confidence values. The presented results with offline processing are obtained with 6 electrodes on 2 subjects with a time segment of 1s. The system is reliable for both subjects over 95%, with rejection criterion.

Brain-controlled finite state machine for wheelchair navigation

This proposal is about a brain-controlled electrically powered wheelchair. The system comprises a brain-computer interface based on steady-state visual evoked potentials and a processing unit relying on a finite state machine (FSM). Results of first simulation experiments comparing two different FSMs are presented.

Robust dazzling detection in a novel visual based dazzling avoidance system

This article presents a novel method of detecting the dazzling effect in a new vision based driving assistance system (ShadeVision) which aims to avoid the dazzling effect caused by strong external light sources. Different from existing systems which only detect the dazzling effect at night, our proposed system can also identify the dazzling effect during the day, particularly in dawn and dusk. To that end, the eye region of the driver and other regions surrounding the eyes, the interior of the vehicle, as well as the background light outside the window are all captured by a high speed camera with 1280×500 resolution and a frame rate of 1000 Hz mounted in the vehicle. Afterwards all these data are synthesized and analyzed to determine the dazzling effect. Moreover, the details of dazzling detection method for nighttime and daytime driving are presented explicitly. The final tests verify the effectiveness of our proposed approaches.

Contrast enhancement in video sequences using variable block shape adaptive histogram equalization

This paper describes a new method for enhancing the local contrast of high dynamical range images on conventional low dynamical range displays. We use the mean shift clustering algorithm to segment the image and enhance segments using contrast limited adaptive histogram equalization (CLAHE) in combination with a new kernel based interpolation technique. Our main application is the enhancement of welding image sequences, but we tested our method on a larger image database. Experiments demonstrate improvements over the traditional CLAHE based image enhancement.

Joint Trajectory Generation and Control for Overground Robot-based Gait Rehabilitation System MOPASS

Robotic gait rehabilitation systems have appeared in the last fifteen years as an alternative to traditional physiotherapy, offering tireless and precise performance. This paper presents the design of MOPASS, a robotic system for overground gait rehabilitation. Additionally, it presents the gait trajectory generator for hip and knee motion in sagittal plane that was implemented in the system. This generator creates healthy-like gait patterns based on extrema (or inflexion) points characteristic of each joint curve. Finally, the motion controllers designed for the active joints of MOPASS are presented.

Robust Face Detection with Eyes Occluded by the Shadow from Dazzling Avoidance System

Some novel vision-based dazzling avoidance systems, such as the ShadeVision system, can intelligently cast a shadow on drivers eye region to protect the eyes from the dazzling effect that are caused by the strong external light sources. However, during this protective state (eyes covered by shadow), the system still needs to detect the drivers face to realize further monitoring, e.g., the intention of the driver as well as the dazzling effect from another light source. The protective shadow will make it difficult to apply the state-of-the-art algorithms to detect the drivers face, as the eye information is of great importance to vision-based face detection. This paper presents a series of robust algorithms for the face detection with protective shadow cast on drivers eyes, among which the Partially Masked Training, the Consecutive Sub-block Training and the Overlapped Sub-block Training are proposed for the first time for this kind of task. The on-road experimental results verify the effectiveness of the designed algorithms.

sBCI-Headset—Wearable and Modular Device for Hybrid Brain-Computer Interface

Severely disabled people, like completely paralyzed persons either with tetraplegia or similar disabilities who cannot use their arms and hands, are often considered as a user group of Brain Computer Interfaces (BCI). In order to achieve high acceptance of the BCI by this user group and their supporters, the BCI system has to be integrated into their support infrastructure. Critical disadvantages of a BCI are the time consuming preparation of the user for the electroencephalography (EEG) measurements and the low information transfer rate of EEG based BCI. These disadvantages become apparent if a BCI is used to control complex devices. In this paper, a hybrid BCI is described that enables research for a Human Machine Interface (HMI) that is optimally adapted to requirements of the user and the tasks to be carried out. The solution is based on the integration of a Steady-state visual evoked potential (SSVEP)-BCI, an Event-related (de)-synchronization (ERD/ERS)-BCI, an eye tracker, an environmental observation camera, and a new EEG head cap for wearing comfort and easy preparation. The design of the new fast multimodal BCI (called sBCI) system is described and first test results, obtained in experiments with six healthy subjects, are presented. The sBCI concept may also become useful for healthy people in cases where a “hands-free” handling of devices is necessary.

Robust profile face detection and rotation angle estimation of the driver's head in a novel dazzling avoidance system

This paper presents robust methods of detecting the profile face and estimating the rotation angle of the drivers head in the ShadeVision system [1], which is a novel vision-based driving assistance system aiming to avoid the dazzling effect caused by strong external light. As the detection rate of dazzling effect highly relies on the brightness of the profile face [2], new algorithms need to be developed to realize the robust profile face detection in driving scenario. When there are multiple strong external light sources, the system needs to know where the drivers sight is focusing in order to determine the light source that is dangerous to the driver [1]. To this end the precise rotation angle of the drivers head is of great importance. The effectiveness of the developed algorithms is verified by the laboratory and field tests.

Extended RRT algorithm with dynamic N-dimensional cuboid domains

The paper presents an approach to path planning problem for manipulators operating in difficult environments with high probability of collisions. The approach improves the Rapidly exploring Random Trees (RRT) algorithm and produces a new approach with better results. It is achieved by expanding the forward and backward tree and reducing dynamically the sampling area of the free-configuration space (C-Space) by introducing a N-Dimensional cuboid regions. The place of this area could be dynamically or statically defined. The results show that this modification provides a very reasonable solution to the path planning problem applicable also to any kind of manipulators.

Ambient intelligence and rehabilitation robots — A necessary symbiosis for robust operation in unstructured environments

In this paper we will consider robots that are designed to support disabled people with quadriplegia. The patients are paralysed e.g. due to a spinal cord injury (SCI) at vertebrae c4 or above and depend all day from care persons. Robots like the care giving robot FRIEND will only be accepted by users if they provide them with independence from personal care persons for considerable time periods over the day in as many situations as possible and if the user can rely on the robot. Robot control must be very reliable and robust against many variations in environment. For robust behaviour, it is necessary to provide the robot control system with additional information from the environment. This information can be generated by a sensor at the robot or by smart environmental systems, which are equipped with additional sensors and actor elements that are used to support the robot system in difficult tasks. In this paper, ambient intelligence is especially considered to support action sequences of the robot and provide redundancy to the robot control system.