Bio-Signals such as electroencephalography (EEG) and electromyography (EMG) are widely used for the rehabilitation of physically disabled people characterization cognitive impairments. Successful decoding these bio-signals is however non-trivial because time-varying non-stationary characteristics. Furthermore, existence short- long-range dependencies in time-series signal makes even more challenging. State-of-the-art studies proposed Convolutional Neural Networks (CNNs) based architectures...

In the presented work we compare machine learning techniques in context of lane change behavior performed by humans a semi-naturalistic simulated environment. We evaluate different approaches using differing feature combinations order to identify appropriate feature, best combination, and most technique for described task. Based on data acquired from human drivers traffic simulator NISYS TRS <sup xmlns:mml="http://www.w3.org/1998/Math/MathML"...

Abstract Objective . In electrophysiology, microelectrodes are the primary source for recording neural data (single unit activity). These can be implanted individually or in form of arrays containing dozens to hundreds channels. Recordings some channels contain activity, which often contaminated with noise. Another fraction does not record any data, but only By noise, we mean physiological activities unrelated spiking, including technical artifacts and neurons that too far away from...

Abstract Brain-computer interfaces (BCIs) enable communication between humans and machines by translating brain activity into control commands. Electroencephalography (EEG) signals are one of the most used in non-invasive BCI applications but often contaminated with noise. Therefore, it is possible that meaningful patterns for classifying EEG deeply hidden. State-of-the-art deep-learning algorithms successful learning hidden, patterns. However, quality quantity presented inputs pivotal....

Based on the concepts of dynamic field theory (DFT), we present an architecture that autonomously generates scene representations by controlling gaze and attention, creating visual objects in foreground, tracking objects, reading them into working memory, taking account their visibility. At core this are three-dimensional neural fields (DNFs) link feature to spatial information. These couple lower dimensional fields, which provide links sensory surface motor systems. We discuss how DNFs can...

Objective. Advancements in electrode design have resulted micro-electrode arrays with hundreds of channels for single cell recordings. In the resulting electrophysiological recordings, each implanted can record spike activity (SA) one or more neurons along background (BA). The aim this study is to isolate SA neural source. This process called sorting classification. Advanced algorithms are time consuming because human intervention at various stages pipeline. Current approaches lack...

Abstract Objective . Accurate decoding of surface electromyography (sEMG) is pivotal for muscle-to-machine-interfaces and their application e.g. rehabilitation therapy. sEMG signals have high inter-subject variability, due to various factors, including skin thickness, body fat percentage, electrode placement. Deep learning algorithms require long training time tend overfit if only few samples are available. In this study, we aim investigate methods calibrate deep models a new user when...

We extend the attractor dynamics approach to generate goal-directed movement of a redundant, anthropomorphic arm while avoiding dynamic obstacles and respecting joint limits. To make robot's movements human-like, we approximately straight-line trajectories by using two heading direction angles tool-point quite analogously how is represented in primate central nervous system. Two additional control tool's spatial orientation so that it follows tool-point's collision-free path. A fifth...

To enable a robotic assistant to autonomously reach for and transport objects while avoiding obstacles we have generalized the attractor dynamics approach established vehicles trajectory formation in robot arms. This is able deal with time-varying environments that occur when human operator moves shared workspace. Stable fixed points (attractors) heading direction of end-effector shift during movement are being tracked by system. enables avoid spurious states hamper potential field methods....

The presented work formulates an framework in which early prediction of drivers lane change behavior is realized. We aim to build a representation order recognize and predict driver's intentions as first step towards realistic driver model. In the test bed Institute Neuroinformatik, based on traffic simulator NISYS TRS <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> , 10 individuals have driven experiments they performed more then 150...

CORA is a robotic assistant whose task to collaborate with human operator on simple manipulation or handling tasks. Its sensory channels comprising vision, audition, haptics, and force sensing are used extract perceptual information about speech, gestures gaze of the operator, object recognition. The anthropomorphic robot arm makes goal-directed movements pick up hand over objects. may mechanically interact by pushing it away (haptics) taking an out robot's gripper (force sensing). design...

For autonomous robots to manipulate objects in unknown environments, they must be able move their arms without colliding with nearby objects, other agents or humans. The simultaneous avoidance of multiple obstacles real time by all link segments a manipulator is still hard task both practice and theory. We present systematic scheme for the generation collision free movements redundant manipulators scenes arbitrarily many obstacles. Based on dynamical systems approach robotics, constraints...

Neurological conditions are a major source of movement disorders. Motion modelling and variability analysis have the potential to identify pathology but require profound data. We introduce systematic dataset 3D center-out task-space trajectories human hand transport movements in natural setting. The tasks this study consist grasping cylindric object from unified start position transporting it one nine target locations unconstrained operational space. measurement procedure is automatized...

For an autonomous robotic system, the ability to share same workspace and interact with humans is basis for cooperative behavior. In this work, we investigate human spatial language as communicative channel between robot human, facilitating their joint work on a tabletop. We specifically combine theory of Dynamic Neural Fields that represent perceptual cognitive states motor control linguistic input in demonstration. show such neural dynamic framework can integrate across symbolic,...

The movement of autonomous agents in natural environments is restricted by potentially large numbers constraints. To generate behavior that fulfills all given constraints simultaneously, the attractor dynamics approach to generation represents each constraint a dynamical system with attractors or repellors at desired undesired values relevant variable. These systems are transformed into vector fields over control variables robotic agent force state whole directions beneficial satisfaction...

We describe the general concept, system architecture, hardware, and behavioral abilities of CORA (Cooperative Robot Assistant), an autonomous nonmobile robot assistant. Outgoing from our basic assumption that behavior to perform determines internal external structure behaving system, we have designed anthropomorphic allow for humanlike strategies in solving complex tasks. Although was built as a prototype service assist human partner industrial assembly tasks, will show CORA's are also...

Advanced Driver Assistant Systems act, by definition in natural, often poorly structured, environments and are supposed to closely interact with human operators. Both, natural as well behaviour have no inherent metric can not be modelled/measured the classical way physically plausibly behaving systems described. This makes development of a new methodology especially an experimental environment necessary, reflecting all constraints task observed, incorporating influence each single component...

The underlying motivation of this work is to demonstrate that artificial muscle activity known and unknown motion can be generated based on parameters, such as angular position, acceleration, velocity each joint (or the end-effector instead), which are similarly represented in our brains. This model motivated by planning process central nervous system. That incorporates current body state from sensory systems previous experiences, might pre-learned inverse dynamics generate associated...

The ability to generate discrete movement with distinct and stable time courses is important for interaction scenarios both between different robots human partners, catching interception tasks, timed action sequences. In dynamic environments, where trajectories are evolving online, this not a trivial task. dynamical systems approach robotics provides framework robust incorporation of fluctuating sensor information, but control usually restricted rhythmic motion realized through limit cycles....

Movement generation in robotics is an old problem with many excellent solutions. Most of them, however, look for optimality according to some metrics, but have no biological inspiration or cannot be used imitate motion. For a human these techniques behave non-naturalistic way. This poses instance human-robot interaction and, general, good acceptance robots society. The present work presents new analysis the attractor dynamics approach movement anthropomorphic robot arm. Our points...

In this paper we describe a straight forward technique for tracking human hand based on images acquired by an active stereo camera system. We demonstrate the implementation of method anthropomorphic assistance robot as part multi-modal man-machine interaction system: detecting hand-position, can interprete pointing gesture specification target object to grasp.

We present an architecture based on the Dynamic Field Theory for problem of scene representation. At core this are three-dimensional neural fields linking feature to spatial information. These coupled lower-dimensional that provide both a close link sensory surface and motor behavior. highlight updating mechanism architecture, when single object is selected followed by robot's head in smooth pursuit multi-item tracking several items move simultaneously.