Search

Locomotion in animals provides a model for adaptive behavior as it is able to deal with various kinds of perturbations. Work in insects suggests that this evolved flexibility results from a modular architecture, which can be characterized by a recurrent neural network allowing for various emerging attractor states. Whereas a lower control-level coordinates joint movements on a short timescale, a higher-level handles action selection on longer timescales. Implementation of such a control system on a walking hexapod robot was able to deal with various walking patterns including disturbances such as uneven terrain or loss of a leg. Here, we propose a cognitive expansion to the adaptive control system that allows dealing with novel challenging situations. This approach makes use of an internal simulation-based planner that is triggered when the model-free controller fails to recover from an unstable pose. Using a grounded internal body model, the planner then tries, in internal simulation, different solutions out of context, and thus, proposes a new plan to be executed on the real robot. We demonstrate the feasibility of this control approach for walking over terrain with uncertain footholds in three scenarios.

For incremental machine-learning applications it is often important to robustly estimate the system accuracy during training, especially if humans perform the supervised teaching. Cross-validation and interleaved test/train error are here the standard supervised approaches. We propose a novel semi-supervised accuracy estimation approach that clearly outperforms these two methods. We introduce the Configram Estimation (CGEM) approach to predict the accuracy of any classifier that delivers confidences. By calculating classification confidences for unseen samples, it is possible to train an offline regression model, capable of predicting the classifier&rsquo, s accuracy on novel data in a semi-supervised fashion. We evaluate our method with several diverse classifiers and on analytical and real-world benchmark data sets for both incremental and active learning. The results show that our novel method improves accuracy estimation over standard methods and requires less supervised training data after deployment of the model. We demonstrate the application of our approach to a challenging robot object recognition task, where the human teacher can use our method to judge sufficient training.

We examine the mechanisms required to handle everyday activities from the standpoint of cognitive robotics, distinguishing activities on the basis of complexity and transparency. Task complexity (simple or complex) reflects the intrinsic nature of a task, while task transparency (easy or difficult) reflects an agent's ability to identify a solution strategy in a given task. We show how the CRAM cognitive architecture allows a robot to carry out simple and complex activities such as laying a table for a meal and loading a dishwasher afterward. It achieves this by using generalized action plans that exploit reasoning with modular, composable knowledge chunks representing general knowledge to transform underdetermined everyday action requests into motion plans that successfully accomplish the required task. Noting that CRAM does not yet have the ability to deal with difficult activities, we leverage insights from the situation model perspective on the cognitive mechanisms underlying flexible context-sensitive behavior with a view to extending CRAM to overcome thisdeficit. © 2021 The Authors. Topics in Cognitive Science published by Wiley Periodicals LLC on behalf of Cognitive Science Society.

Deep Reinforcement Learning techniques demonstrate advances in the domain of robotics. One of the limiting factors is a large number of interaction samples usually required for training in simulated and real-world environments. In this work, we demonstrate for a set of simulated dexterous in-hand object manipulation tasks that tactile information can substantially increase sample efficiency for training (by up to more than threefold). We also observe an improvement in performance (up to 46%) after adding tactile information. To examine the role of tactile-sensor parameters in these improvements, we included experiments with varied sensor-measurement accuracy (ground truth continuous values, noisy continuous values, Boolean values), and varied spatial resolution of the tactile sensors (927 sensors, 92 sensors, and 16 pooled sensor areas in the hand). To facilitate further studies and comparisons, we make these touch-sensor extensions available as a part of the OpenAI Gym Shadow-Dexterous-Hand robotics environments.

We present an approach based on conditional generative adversarial networks (GANs) to generate grasps directly and in a feed-forward manner from a raw depth image input. Building on the recently introduced StyleGAN architecture we extend results from an earlier proof-of-concept paper [1] and demonstrate successful sim2real transfer of grasp outputs for a robot arm with a Shadow Dexterous Hand. We find that the GAN model, which was only trained on a limited set of primitive objects, was able to generalize to a range of everyday real-world objects that differed significantly from the primitive objects used in simulation training. In contrast to discriminative models, the approach learns a latent representation in the set of feasible grasps that can be used for navigation in grasp space and thus allows smooth integration with other motion planning tools.

In daily life, humans use their limbs to perform various movements to interact with an external environment. Thanks to limb's variable and adaptive stiffness, humans can adapt their movements to unstable dynamics of the external environments. The underlying adaptive mechanism has been investigated, employing a simple planar device perturbed by external 2D force patterns. In the present work, we will employ a more advanced, compliant robot arm to extend previous work to a more realistic 3D-setting. We study the adaptive mechanism and use machine learning to capture the human adaptation behavior. In order to model human's stiffness adaptive skill, we give human subjects the task to reach for a target by moving a handle assembled on the end-effector of a compliant robotic arm. The arm is force controlled and the human is required to navigate the handle inside a non-visible, virtual maze and explore it only through robot force feedback when contacting maze virtual walls. By sampling the hand's position and force data, a computational model based on a combination of model predictive control and nonlinear regression is used to predict participants' successful trials. Our study shows that participants selectively increased the stiffness within the axis direction of uncertainty in order to compensate for instability caused by a divergent external force field. The learned controller was able to successfully mimic this behavior. When it is deployed on the robot for the navigation task, the robot arm successfully adapt to the unstable dynamics in the virtual maze, in a similar manner as observed in the participants' adaptation skill.

Decentralization is a central characteristic of biological motor control that allows for fast responses relying on local sensory information. In contrast, the current trend of Deep Reinforcement Learning (DRL) based approaches to motor control follows a centralized paradigm using a single, holistic controller that has to untangle the whole input information space. This motivates to ask whether decentralization as seen in biological control architectures might also be beneficial for embodied sensori-motor control systems when using DRL. To answer this question, we provide an analysis and comparison of eight control architectures for adaptive locomotion that were derived for a four-legged agent, but with their degree of decentralization varying systematically between the extremes of fully centralized and fully decentralized. Our comparison shows that learning speed is significantly enhanced in distributed architectures—while still reaching the same high performance level of centralized architectures—due to smaller search spaces and local costs providing more focused information for learning. Second, we find an increased robustness of the learning process in the decentralized cases—it is less demanding to hyperparameter selection and less prone to becoming trapped in poor local minima. Finally, when examining generalization to uneven terrains—not used during training—we find best performance for an intermediate architecture that is decentralized, but integrates only local information from both neighboring legs. Together, these findings demonstrate beneficial effects of distributing control into decentralized units and relying on local information. This appears as a promising approach towards more robust DRL and better generalization towards adaptive behavior.

This work discusses a learning approach to mask rewarding objects in images using sparse reward signals from an imitation learning dataset. For that we train an Hourglass network using only feedback from a critic model. The Hourglass network learns to produce a mask to decrease the critic’s score of a high score image and increase the critic’s score of a low score image by swapping the masked areas between these two images. We trained the model on an imitation learning dataset from the NeurIPS 2020 MineRL Competition Track, where our model learned to mask rewarding objects in a complex interactive 3D environment with a sparse reward signal. This approach was part of the 1st place winning solution in this competition. Video demonstration and code: https://rebrand.ly/critic-guided-segmentation.

In the experiments of comparative visual search reported here, each half of a display contains simple geometrical objects of three different colors and forms. The two hemifields are identical except for one mismatch either in color or form. The subject’s task is to find this difference. Eye-movement recording yields insight into the interaction of mental processes involved in the completion of this demanding task. We present a hierarchical model of comparative visual search and its implementation as a computer simulation. The evaluation of simulation data shows that this Three-Level Model is able to explain about 98% of the empirical data collected in six different experiments.

We present our second generation tactile sensor for the Shadow Dexterous Hand's palm. We were able to significantly improve the tactile sensor characteristics by utilizing our latest barometer-based tactile sensing technology with linear (R² ≥ 0.9996) sensor output and no noticeable hysteresis. The sensitivity threshold of the tactile cells and the spatial density were both dramatically increased. We demonstrate the benefits of the new sensor by re-running an experiment to estimate the stiffness of different objects that we originally used to test our first generation palm sensor. The results underline a considerable performance boost in estimation accuracy, just due to the improved tactile skin. We also propose a revised neural network architecture that even further improves the average classification accuracy to 96% in a 5-fold cross-validation.

Because of the complex anatomy of the human hand, in the absence of external constraints, a large number of postures and force combinations can be used to attain a stable grasp. Motor synergies provide a viable strategy to solve this problem of motor redundancy. In this study, we exploited the technical advantages of an innovative sensorized object to study unconstrained hand grasping within the theoretical framework of motor synergies. Participants were required to grasp, lift, and hold the sensorized object. During the holding phase, we repetitively applied external disturbance forces and torques and recorded the spatiotemporal distribution of grip forces produced by each digit. We found that the time to reach the maximum grip force during each perturbation was roughly equal across fingers, consistent with a synchronous, synergistic stiffening across digits. We further evaluated this hypothesis by comparing the force distribution of human grasping vs. robotic grasping, where the control strategy was set by the experimenter. We controlled the global hand stiffness of the robotic hand and found that this control algorithm produced a force pattern qualitatively similar to human grasping performance. Our results suggest that the nervous system uses a default whole hand synergistic control to maintain a stable grasp regardless of the number of digits involved in the task, their position on the objects, and the type and frequency of external perturbations. NEW & NOTEWORTHY We studied hand grasping using a sensorized object allowing unconstrained finger placement. During object perturbation, the time to reach the peak force was roughly equal across fingers, consistently with a synergistic stiffening across fingers. Force distribution of a robotic grasping hand, where the control algorithm is based on global hand stiffness, was qualitatively similar to human grasping. This suggests that the central nervous system uses a default whole hand synergistic control to maintain a stable grasp.

Haptic exploration is a key skill for both robots and humans to discriminate and handle unknown objects or to recognize familiar objects. Its active nature is evident in humans who from early on reliably acquire sophisticated sensory-motor capabilities for active exploratory touch and directed manual exploration that associates surfaces and object properties with their spatial locations. This is in stark contrast to robotics. In this field, the relative lack of good real-world interaction models-along with very restricted sensors and a scarcity of suitable training data to leverage machine learning methods-has so far rendered haptic exploration a largely underdeveloped skill. In robot vision however, deep learning approaches and an abundance of available training data have triggered huge advances. In the present work, we connect recent advances in recurrent models of visual attention with previous insights about the organisation of human haptic search behavior, exploratory procedures and haptic glances for a novel architecture that learns a generative model of haptic exploration in a simulated three-dimensional environment. This environment contains a set of rigid static objects representing a selection of one-dimensional local shape features embedded in a 3D space: an edge, a flat and a convex surface. The proposed algorithm simultaneously optimizes main perception-action loop components: feature extraction, integration of features over time, and the control strategy, while continuously acquiring data online. Inspired by the Recurrent Attention Model, we formalize the target task of haptic object identification in a reinforcement learning framework and reward the learner in the case of success only. We perform a multi-module neural network training, including a feature extractor and a recurrent neural network module aiding pose control for storing and combining sequential sensory data. The resulting haptic meta-controller for the rigid 16 × 16 tactile sensor array moving in a physics-driven simulation environment, called the Haptic Attention Model, performs a sequence of haptic glances, and outputs corresponding force measurements. The resulting method has been successfully tested with four different objects. It achieved results close to 100% while performing object contour exploration that has been optimized for its own sensor morphology.

We present a distributed supervised learning architecture, which can generate trajectory data conditioned by control commands and learned from demonstrations. The architecture consists of an ensemble of neural networks (NNs) which learns the dynamic model and a separate addressing NN that decides from which NN to draw a prediction. We introduce an error-based method for automatic assignment of data subsets to the ensemble NNs for training using the loss profile of the ensemble. Our code is publicly available (Code: https://github.com/NicoBach/distributed-dynamics-model).

This work presents a novel object-level control framework for the dexterous in-hand manipulation of objects with torque-controlled robotic hands. The proposed impedance-based controller realizes the compliant 6-DOF positioning of a grasped object. Enabled by the in-hand localization, the slippage of contacts is avoided by actively maintaining the desired grasp configuration. The internal forces on the object are determined by solving a quadratic optimization problem, which explicitly considers friction constraints on the contacts and allows to gradually shift the load between fingers. The proposed framework is capable of dealing with dynamic changes in the grasp configuration, which makes it applicable for the control of the object during grasp acquisition or the reconfiguration of fingers (i.e. finger gaiting). A nullspace controller avoids joint limits and singularities. Experiments with the DLR robot David demonstrate the efficiency and performance of the controller in various grasping scenarios.

Deep Reinforcement Learning has recently seen progress for continuous control tasks, driven by yearly challenges such as the NeurIPS Competition Track. This work combines complementary characteristics of two current state of the art methods, Twin-Delayed Deep Deterministic Policy Gradient and Distributed Distributional Deep Deterministic Policy Gradient, and applied this in the state-of-the-art Learn to move—Walk Around locomotion control challenge which was part of the NeurIPS 2019 Competition Track. The combined approach showed improved results and achieved the 4th place in this competition. The article presents this combination and evaluates the performance.

Recent developments in machine-learning algorithms have led to impressive performance increases in many traditional application scenarios of artificial intelligence research. In the area of deep reinforcement learning, deep learning functional architectures are combined with incremental learning schemes for sequential tasks that include interaction-based, but often delayed feedback. Despite their impressive successes, modern machine-learning approaches, including deep reinforcement learning, still perform weakly when compared to flexibly adaptive biological systems in certain naturally occurring scenarios. Such scenarios include transfers to environments different than the ones in which the training took place or environments that dynamically change, both of which are often mastered by biological systems through a capability that we here term "fluid adaptivity" to contrast it from the much slower adaptivity ("crystallized adaptivity") of the prior learning from which the behavior emerged. In this article, we derive and discuss research strategies, based on analyzes of fluid adaptivity in biological systems and its neuronal modeling, that might aid in equipping future artificially intelligent systems with capabilities of fluid adaptivity more similar to those seen in some biologically intelligent systems. A key component of this research strategy is the dynamization of the problem space itself and the implementation of this dynamization by suitably designed flexibly interacting modules.

A lot of motion planning research in robotics focuses on efficient means to find trajectories between individual start and goal regions, but it remains challenging to specify and plan robotic manipulation actions which consist of multiple interdependent subtasks. The Task Constructor framework we present in this work provides a flexible and transparent way to define and plan such actions, enhancing the capabilities of the popular robotic manipulation framework MoveIt!.11The Task Constructor framework is publicly available at https://github.com/ros-planning/moveit_task_constructor Subproblems are solved in isolation in black-box planning stages and a common interface is used to pass solution hypotheses between stages. The framework enables the hierarchical organization of basic stages using containers, allowing for sequential as well as parallel compositions. The flexibility of the framework is illustrated in multiple scenarios performed on various robot platforms, including bimanual ones.

Brain-computer interfaces provide a means for controlling a device by brain activity alone. One major drawback of noninvasive BCIs is their low information transfer rate, obstructing a wider deployment outside the lab. BCIs based on codebook visually evoked potentials (cVEP) outperform all other state-of-the-art systems in that regard. Previous work investigated cVEPs for spelling applications. We present the first cVEP-based BCI for use in real-world settings to accomplish everyday tasks such as navigation or action selection. To this end, we developed and evaluated a cVEP-based on-line BCI that controls a virtual agent in a simulated, but realistic, 3-D kitchen scenario. We show that cVEPs can be reliably triggered with stimuli in less restricted presentation schemes, such as on dynamic, changing backgrounds. We introduce a novel, dynamic repetition algorithm that allows for optimizing the balance between accuracy and speed individually for each user. Using these novel mechanisms in a 12-command cVEP-BCI in the 3-D simulation results in ITRs of 50 bits/min on average and 68 bits/min maximum. Thus, this work supports the notion of cVEP-BCIs as a particular fast and robust approach suitable for real-world use.

We propose a reinforcement learning approach that combines an asynchronous actor-critic model with a recurrent model of visual attention. Instead of using the full visual information of the scene, the resulting model accumulates the foveal information of controlled glimpses and is thus able to reduce the complexity of the network. Using the designed model, an artificial agent is able to solve a challenging “mediated interaction” task. In these tasks, the desired effects cannot be created through direct interaction, but instead require the learner to discover how to exert suitable effects on the target object through involving a “tool”. To learn the given mediated interaction task, the agent is “actively” searching for salient points within the environment by taking a limited number of fovea-like glimpses. It then uses the accumulated information to decide which action to take next.

Single-trial classification of EEG data from Disorder of Consciousness patients (DoC) has proved particularly challenging. We present an approach that establishes a measure to relate the performance of single-trial classification of DoC patient EEG data with relational frequency bands and thus with their mental state. We evaluate our approach on 31 patient data sets from two studies, showing that our measure indicates for different data sets a particular likelihood for misclassifying either target or non-target class samples.

We investigate the effectiveness of transfer learn- ing for accelerating shallow and deep reinforcement learning of “mediated interaction tasks”. In these tasks, the desired effects cannot be created through direct interaction, but instead require the learner to discover how to exert suitable effects on the target object through involving a suitable “mediator object”. We focus on the case where transfer learning is applied to generalize experiences from source tasks that are solvable through direct, unmediated interaction, to target tasks that require mediated interaction for their solution. We find that transfer learning employed in this context leads to a significant acceleration of learning. A refinement of the basic transfer learning strategy that is motivated from the principle of scaffolding in psychology leads to further improvements, totalling to an overall speed-up factor of almost one order of magnitude for a reinforcement learner solving an “extension- of-reach” task in a 2D world with simulated physics.

Our fingernails help us to accomplish a variety of manual tasks, but surprisingly only a few robotic hands are equipped with nails. In this paper, we present a sensorized fingernail for mechatronic hands that can capture static and dynamic interaction forces with the nail. Over the course of several iterations, we have developed a very compact working prototype that fits together with our previously developed multi-cell tactile fingertip sensor into the cavity of the distal phalange of a human-sized robotic hand. We present the construction details, list the key performance characteristics and demonstrate an example application of finding the end of an adhesive tape roll using the signals captured by the sensors integrated in the nail. We conclude with a discussion about improvement ideas for future versions.

The ability to discriminate between target and distractors, using the information perceived by the hand over time, is essential to perform haptic search successfully, be it for a human hand or a suitably sensorized anthropomorphic robot hand. To address the latter, we train a binary classifier to perform this discrimination during unconstrained haptic search performed by sighted study participants who were blindfolded. In this work, we test different representational concepts and compare the results with the human classification performance. This approach both guides our understanding of human haptic interaction with the 3D environment and aids future modeling of artificial touch for anthropomorphic robot hands. Our contribution is three-fold. Firstly, we are able to acquire a synchronized multimodal time series of exceptionally high spatio-temporal resolution of both the 3D environment and the hand with our novel experimental setup. It includes our Modular Haptic Stimulus Board to represent a 3D environment and a novel tactile glove equipped with position tracking markers and joint angle sensors. Secondly, we introduce a machine learning approach inspired by a novel application of the feature guidance concept for vision (Wolfe et al., 2007 [1]) to modeling of haptic search in a 3D environment, focusing on the target-distractor discrimination. Finally, we compare results for two different types of artificial neural networks, a feedforward and a recurrent network. We show that using recurrent networks, and therefore integrating information over time, improves the classification results. The evaluation also shows that classification accuracy is the highest for the combination of both the tactile and the joint angle modalities.

Formulating problems rigorously in terms of optimization principles has become a dominating approach in the fields of machine learning and computer vision. However, the systems described in these fields are in some respects different to integrated, modular, and embodied systems, such as the ones we aim to build in robotics. While representing systems via optimality principles is a powerful approach, relying on it as the sole approach to robotics raises substantial challenges. In this article, we take this as a starting point to discuss which ways of representing problems should be best-suited for robotics. We argue that an adequate choice of system representation—e.g. via optimization principles—must allow us to reflect the structure of the problem domain. We discuss system design principles, such as modularity, redundancy, stability, and dynamic processes, and the degree to which they are compatible with the optimization stance or instead point to alternative paradigms in robotics research. This discussion, we hope, will bring attention to this important and often ignored system-level issue in the context of robotics research.

In the NIPS 2017 Learning to Run challenge, participants were tasked with building a controller for a musculoskeletal model to make it run as fast as possible through an obstacle course. Top participants were invited to describe their algorithms. In this work, we present eight solutions that used deep reinforcement learning approaches, based on algorithms such as Deep Deterministic Policy Gradient, Proximal Policy Optimization, and Trust Region Policy Optimization. Many solutions use similar relaxations and heuristics, such as reward shaping, frame skipping, discretization of the action space, symmetry, and policy blending. However, each of the eight teams implemented different modifications of the known algorithms., Comment: 27 pages, 17 figures

International audience; Tactile sensing technology has achieved much progress in last decades. Recently developed tactile sensing systems have been utilized for implementation of bodily-aware robots, enhancement of robotic hand's dexterous manipulation, development of wearable devices for facilitating the human robot interaction, and so on. While tactile sensing has resulted in many possibilities in robotic research, robot's manipulation is still far from that of its counterpart-human-in term of dexterity. To overcome this issue, we are considering the following three aspects that need to be tackled: - Available tactile sensing technology is struggling to meet the capabilities of humans’ sense of touch; - Tactile data processing algorithms and data-driven learning technology are still immature; - General and robust robot manipulation control frameworks are missing

Robotic manipulation of such highly deformable objects as clothes is a challenging problem. Robot-assisted dressing adds even more complexity as the garment motions must be aligned with a human body under conditions of strong and variable occlusion. As a step toward solutions for the general task, we consider the example of a dual-arm robot with attached anthropomorphic hands that learns to put a knit cap on a styrofoam head. Our approach avoids modeling the details of the garment and its deformations. Instead, we demonstrate that a head-centric policy parameterization, combined with a suitable objective function for determining the right amount of contact between the cap and the head, enables a direct policy search algorithm to find successful trajectories for this task. We also show how a toy problem that mirrors some of the task constraints can be used to efficiently structure hyperparameter search. Additionally, we suggest a point cloud based algorithm for modeling the head as an ellipsoid which is required for defining the policy space.

Recent advances in the development of optical head-mounted displays (HMDs), such as the Microsoft HoloLens, Google Glass, or Epson Moverio, which overlay visual information directly in the user's field of vision, have opened up new possibilities for augmented reality (AR) applications. We propose a system that uses such an optical HMD to assist the user during goal-oriented activities (e.g. manufacturing work) in an intuitive and unobtrusive way (Essig et al. 2016). To this end, our system observes and recognizes the user's actions and generates context-sensitive feedback. Figure 1 shows an overview of our approach, exemplified with the task of assembling a bird house.

In this paper we present a novel approach to model perceptual grouping based on phase and frequency synchronization in a network of coupled Kuramoto oscillators. Transferring the grouping concept from the Competitive Layer Model (CLM) to a network of Kuramoto oscillators, we preserve the excellent grouping capabilities of the CLM, while dramatically improving the convergence rate, robustness to noise, and computational performance, which is verified in a series of artificial grouping experiments and with real-world data.

Finding good principles to choose the actions of artificial agents like robots in the most beneficial way to optimize their control of the environment is very much in the focus of current research in the field of intelligent systems. Especially in reinforcement learning, where the agent learns through the direct interaction with the environment, a good choice of actions is essential. We propose a new approach that allows a predictive ranking of different action sets with regard to their influence on the learning performance of an artificial agent. Our approach is based on a measure of control that utilizes the concept of mutual information. To evaluate this approach, we investigate its prediction of the effectiveness of different sets of actions in “mediated interaction” scenarios. Our results indicate that the mutual information-based measure can yield useful predictions on the aptitude of action sets for the learning process.

This paper presents learning and generalization algorithms for Gaussian Mixture Model (GMM) in order to accurately encode 3-DoF orientations and Euclidean variables in a common model. We employ correct displacement, integration and weighted averaging arithmetics for unit quaternions to adapt the learning and generalization methods of standard GMMs. We validate the proposed method in three different applications, learning a 3-dimensional rotation matrix, learning reachable space of a robot, and learning the motion model from demonstrations. We show good experimental results compared to the state-of-the-art method. (C) 2016 Elsevier B.V. All rights reserved.

Detecting hand-object interactions is a challenging problem with many applications in the human-computer interaction domain. We present a real-time method that automatically detects hand-object interactions in RGBD sensor data and tracks the object's rigid pose over time. The detection is performed using a fully convolutional neural network, which is purposefully trained to discern the relationship between hands and objects and which predicts pixel-wise class probabilities. This output is used in a probabilistic pixel labeling strategy that explicitly accounts for the uncertainty of the prediction. Based on the labeling of object pixels, the object is tracked over time using model-based registration. We evaluate the accuracy and generalizability of our approach and make our annotated RGBD dataset as well as our trained models publicly available.

It has been suggested recently that it is a reasonable abstraction of evolutionary processes to use evolutionary algorithms that select individuals based on the novelty of their behavior instead of their fitness. Here we study the performance of fitness- and novelty-based search on several neuroevolution tasks. We also propose several new algorithms that select both for fit and for novel individuals, but without weighting these two criteria directly against each other. We find that behavioral speciation, behavioral near neutral speciation, and behavioral novelty speciation perform best on most tasks. Pure novelty search, as well as a number of hybrid methods without speciation mechanism, do not perform well on most tasks. Using behavioral criteria for speciation often yields better results than using genetic criteria.

Tangible Interaction is an established subfield of Human-Computer Interaction research. Actuated tangible user interfaces are still an unsufficiently investigated subject, which offers completely new possibilities for interaction design. We present mechanisms for saving and automatically restoring arrangements of tangible objects, two different communication concepts, an exploratory data analysis system for blind users, as well as multi-modal interaction approaches., Begreifbare Interaktion ist ein sich etablierendes Forschungsfeld der Mensch-Maschine-Interaktion. Ein längst noch nicht vollständig untersuchter Zweig dieses Feldes stellen sich selbst bewegende Objekte dar, die völlig neue, bisher nicht erschöpfend untersuchte Interaktionsmöglichkeiten bieten. Vorgestellt werden unter anderem Mechanismen zum Speichern und automatischen Laden von Anordnungen der begreifbaren Objekte, zwei verschiedene Konzepte zur Kommunikation, sowie ein Datenanalysesystem für Blinde und weitere Ansätze für den Einsatz multi-modaler Interaktionstechniken.

The coming decade will bring raw processing power and storage capacities of everyday computers to the same level as small brains. Considering this, current interaction technology seems archaic as it still forces humans to follow highly stereotyped, narrow and often error-prone procedures in order to make computers, robots, or other machinery obey them. The vision behind the Excellence Cluster Cognitive Interaction Technology (CITEC) is to develop ways of interacting with technical systems that are as natural and smooth as communication between humans.

(ABRIDGED) In previous work, two platforms have been developed for testing computer-vision algorithms for robotic planetary exploration (McGuire et al. 2004b,2005; Bartolo et al. 2007). The wearable-computer platform has been tested at geological and astrobiological field sites in Spain (Rivas Vaciamadrid and Riba de Santiuste), and the phone-camera has been tested at a geological field site in Malta. In this work, we (i) apply a Hopfield neural-network algorithm for novelty detection based upon color, (ii) integrate a field-capable digital microscope on the wearable computer platform, (iii) test this novelty detection with the digital microscope at Rivas Vaciamadrid, (iv) develop a Bluetooth communication mode for the phone-camera platform, in order to allow access to a mobile processing computer at the field sites, and (v) test the novelty detection on the Bluetooth-enabled phone-camera connected to a netbook computer at the Mars Desert Research Station in Utah. This systems engineering and field testing have together allowed us to develop a real-time computer-vision system that is capable, for example, of identifying lichens as novel within a series of images acquired in semi-arid desert environments. We acquired sequences of images of geologic outcrops in Utah and Spain consisting of various rock types and colors to test this algorithm. The algorithm robustly recognized previously-observed units by their color, while requiring only a single image or a few images to learn colors as familiar, demonstrating its fast learning capability., Comment: 28 pages, 12 figures, accepted for publication in the International Journal of Astrobiology

We present a biologically motivated architecture for object recognition that is capable of online learning of several objects based on interaction with a human teacher. The system combines biological principles such as appearance-based representation in topographical feature detection hierarchies and context-driven transfer between different levels of object memory. Training can be performed in an unconstrained environment by presenting objects in front of a stereo camera system and labeling them by speech input. The learning is fully online and thus avoids an artificial separation of the interaction into training and test phases. We demonstrate the performance on a challenging ensemble of 50 objects.

We believe that augmenting or complementing natural functions of the brain can be an effective way to develop the technology for more brain-adequate interfaces. We illustrate this line of approach with examples from our recent work and report briefly on a major research initiative launched recently at Bielefeld University, joining the efforts of computer science groups and linguists towards more brain-adequate human–computer interfaces by studying the topic of interactive alignment in communication.

This paper describes a system for visual object recognition based on mobile augmented reality gear. The user can train the system to the recognition of objects online using advanced methods of interaction with mobile systems: Hand gestures and speech input control “virtual menus,” which are displayed as overlays within the camera image. Here we focus on the underlying neural recognition system, which implements the key requirement of an online trainable system—fast adaptation to novel object data. The neural three-stage architecture can be adapted in two modes: In a fast training mode (FT), only the last stage is adapted, whereas complete training (CT) rebuilds the system from scratch. Using FT, online acquired views can be added at once to the classifier, the system being operational after a delay of less than a second, though still with reduced classification performance. In parallel, a new classifier is trained (CT) and loaded to the system when ready.

Optical motion tracking systems often require a lot of manual work to generate clean labeled trajectories. This can be a deterrent if the goal is the creation of large motion tracking datasets. Especially in the case of hand tracking, issues of occlusion (often self-occlusion by other fingers) make the post-processing task very difficult and time intensive. We introduce a fully automatic optical motion tracking method that utilizes a model based inverse kinematics approach. The Hungarian method is used to efficiently calculate associations between model markers and motion capture markers and we demonstrate an elegant solution to the problem of occlusions using a posture interpolation step.

We discuss the framework of Model-Based Sonification (MBS) and its contribution to a principled design of mediators between high-dimensional data spaces and perceptual spaces, particularly sound spaces. Data Crystallization Sonification, discussed in the reprinted paper, exemplifies the design of sonification models according to this framework. Finally, promising lines of development in this area are pointed out, concerning generalizations, applications, and open research directions.

We present an approach for the convenient labeling of image patches gathered from an unrestricted environment. The system is employed for a mobile Augmented Reality (AR) gear: while the user walks around with the head-mounted AR-gear, context-free modules for focus-of-attention permanently sample the most 'interesting' image patches. After this acquisition phase, a Self-Organizing Map (SOM) is trained on the complete set of patches, using combinations of MPEG-7 features as a data representation. The SOM allows visualization of the sampled patches and an easy manual sorting into categories. With very little effort, the user can compose a training set for a classifier, thus, unknown objects can be made known to the system. We evaluate the system for COIL-imagery and demonstrate that a user can reach satisfying categorization within few steps, even for image data sampled from walking in an office environment. (An abbreviated version of some portions of this article appeared in [Bekel, H., Heidemann, G., & Ritter, H. (2005). SOM Based Image Data Structuring in an Augmented Reality Scenario. In Proceedings of the International Joint Conference on Neural Networks, Montreal, Canada.], as part of the IJCNN 2005 conference proceedings, published under the IEEE copyright).

We introduce a new type of neural network--the dynamic wave expansion neural network (DWENN)--for path generation in a dynamic environment for both mobile robots and robotic manipulators. Our model is parameter-free, computationally efficient, and its complexity does not explicitly depend on the dimensionality of the configuration space. We give a review of existing neural networks for trajectory generation in a time-varying domain, which are compared to the presented model. We demonstrate several representative simulative comparisons as well as the results of long-run comparisons in a number of randomly-generated scenes, which reveal that the proposed model yields dominantly shorter paths, especially in highly-dynamic environments.

In this paper, we present a fully automated image segmentation method based on an algorithm that provides adaptive plasticity in function approximation problems: the deformable (feature) map (DM) algorithm. The DM approach reduces a class of similar function approximation problems to the explicit supervised one-shot training of a single data set. This is followed by a subsequent, appropriate similarity transformation, which is based on a self-organized deformation of the underlying multidimensional probability distributions. We apply this algorithm to the real-world problem of fully automated voxel-based multispectral image segmentation, employing magnetic resonance data sets of the human brain. In contrast to previous segmentation approaches, the knowledge obtained within the segmentation procedure of a single prototypical reference data set can be re-utilized for the segmentation of new, 'similar' data employing a strategy of incremental adaptive learning based on the DM algorithm. Thus, we obtain a fully automatic segmentation method that does neither require manual contour tracing of training regions, visual classification of voxel clusters, nor any other kind of human intervention. Our application demonstrates that flexible learning by a strategy of self-organized incremental model adaptation can contribute to increase the efficiency and practicability of biomedical image processing systems.

A key prerequisite to make user instruction of work tasks by interactive demonstration effective and convenient is situated multi-modal interaction aiming at an enhancement of robot learning beyond simple low-level skill acquisition. We report the status of the Bielefeld GRAVIS-robot system that combines visual attention and gestural instruction with an intelligent interface for speech recognition and linguistic interpretation to allow multi-modal task-oriented instructions. With respect to this platform, we discuss the essential role of learning for robust functioning of the robot and sketch the concept of an integrated architecture for situated learning on the system level. It has the long-term goal to demonstrate speech-supported imitation learning of robot actions. We describe the current state of its realization to enable imitation of human hand postures for flexible grasping and give quantitative results for grasping a broad range of everyday objects.

This paper presents a review of self-organizing feature maps (SOFMs), in particular, those based on the Kohonen algorithm, applied to adaptive modeling and control of robotic manipulators. Through a number of references we show how SOFMs can learn nonlinear input–output mappings needed to control robotic manipulators, thereby coping with important robotic issues such as the excess degrees of freedom, computation of inverse kinematics and dynamics, hand–eye coordination, path-planning, obstacle avoidance, and compliant motion. We conclude the paper arguing that SOFMs can be a much simpler, feasible alternative to MLP and RBF networks for function approximation and for the design of neurocontrollers. Comparison with other supervised/unsupervised approaches and directions for further work on the field are also provided.

Striving for an autonomous self-exploration of robots to learn their own body schema, i.e. body shape and appearance, kinematic and dynamic parameters, association of tactile stimuli to specific body locations, etc., we developed a tactile-servoing feedback controller that allows a robot to continuously acquire self-touch information while sliding a fingertip across its own body. In this manner one can quickly acquire a large amount of training data representing the body shape. We compare three approaches to track the common contact point observed when one robot arm is touching the other in a bimanual setup: feedforward control, solely relying on a coarse CAD-based kinematics performs worst, a solely feedback-based controller typically lacks behind, and only the combination of both approaches yields satisfactory tracking results. As a first, preliminary application, we use this self-touch capability to calibrate the closed kinematic chain formed by both arms touching each other. The obtained homogeneous transform describing the relative mounting pose of both arms improves end-effector position estimations by a magnitude.

In this paper, we present an algorithm that provides adaptive plasticity in function approximation problems: the deformable (feature) map (DM) algorithm. The DM approach reduces a class of similar function approximation problems to the explicit supervised one-shot training of a single data set. This is followed by a subsequent, appropriate similarity transformation which is based on a self-organized deformation of the underlying multi-dimensional probability distributions. After discussing the theory of the DM algorithm, we use computer simulations to visualize its effects on low-dimensional toy examples. Finally, we present results of its application to the real-world problem of automatic nonlinear multispectral image registration, employing magnetic resonance data sets of the human brain.

A system for the automatic segmentation of fluorescence micrographs is presented. In the first step, positions of fluorescent cells are detected by a fast learning neural network, which acquires the visual knowledge from a set of training cell-image patches selected by the user. Guided by the detected cell positions the system extracts in the second step the contours of the cells. For contour extraction, a recurrent neural network model is used to approximate the cell shapes. Even though the micrographs are noisy and the fluorescent cells vary in shape and size, the system detects at minimum 95% of the cells. (C) 2002 Elsevier Science B.V. All rights reserved.