Oncilla Robot: A Versatile Open-Source Quadruped Research Robot With Compliant Pantograph Legs.

We present Oncilla robot, a novel mobile, quadruped legged locomotion machine. This large-cat sized, 5.1 kg robot is one of a kind of a recent, bioinspired legged robot class designed with the capability of model-free locomotion control. Animal legged locomotion in rough terrain is clearly shaped by sensor feedback systems. Results with Oncilla robot show that agile and versatile locomotion is possible without sensory signals to some extend, and tracking becomes robust when feedback control is added (Ajallooeian, 2015). By incorporating mechanical and control blueprints inspired from animals, and by observing the resulting robot locomotion characteristics, we aim to understand the contribution of individual components. Legged robots have a wide mechanical and control design parameter space, and a unique potential as research tools to investigate principles of biomechanics and legged locomotion control. But the hardware and controller design can be a steep initial hurdle for academic research. To facilitate the easy start and development of legged robots, Oncilla-robot's blueprints are available through open-source. The robot's locomotion capabilities are shown in several scenarios. Specifically, its spring-loaded pantographic leg design compensates for overdetermined body and leg postures, i.e., during turning maneuvers, locomotion outdoors, or while going up and down slopes. The robot's active degree of freedom allow tight and swift direction changes, and turns on the spot. Presented hardware experiments are conducted in an open-loop manner, with little control and computational effort. For more versatile locomotion control, Oncilla-robot can sense leg joint rotations, and leg-trunk forces. Additional sensors can be included for feedback control with an open communication protocol interface. The robot's customized actuators are designed for robust actuation, and efficient locomotion. It trots with a cost of transport of 3.2 J/(Nm), at a speed of 0.63 m s-1 (Froude number 0.25). The robot trots inclined slopes up to 10°, at 0.25 m s-1. The multi-body Webots model of Oncilla robot, and Oncilla robot's extensive software architecture enables users to design and test scenarios in simulation. Controllers can directly be transferred to the real robot. Oncilla robot's blueprints are open-source published (hardware GLP v3, software LGPL v3).

On learning navigation behaviors for small mobile robots with reservoir computing architectures.

This paper proposes a general reservoir computing (RC) learning framework that can be used to learn navigation behaviors for mobile robots in simple and complex unknown partially observable environments. RC provides an efficient way to train recurrent neural networks by letting the recurrent part of the network (called reservoir) be fixed while only a linear readout output layer is trained. The proposed RC framework builds upon the notion of navigation attractor or behavior that can be embedded in the high-dimensional space of the reservoir after learning. The learning of multiple behaviors is possible because the dynamic robot behavior, consisting of a sensory-motor sequence, can be linearly discriminated in the high-dimensional nonlinear space of the dynamic reservoir. Three learning approaches for navigation behaviors are shown in this paper. The first approach learns multiple behaviors based on the examples of navigation behaviors generated by a supervisor, while the second approach learns goal-directed navigation behaviors based only on rewards. The third approach learns complex goal-directed behaviors, in a supervised way, using a hierarchical architecture whose internal predictions of contextual switches guide the sequence of basic navigation behaviors toward the goal.

Memristor models for machine learning.

In the quest for alternatives to traditional complementary metal-oxide-semiconductor, it is being suggested that digital computing efficiency and power can be improved by matching the precision to the application. Many applications do not need the high precision that is being used today. In particular, large gains in area and power efficiency could be achieved by dedicated analog realizations of approximate computing engines. In this work we explore the use of memristor networks for analog approximate computation, based on a machine learning framework called reservoir computing. Most experimental investigations on the dynamics of memristors focus on their nonvolatile behavior. Hence, the volatility that is present in the developed technologies is usually unwanted and is not included in simulation models. In contrast, in reservoir computing, volatility is not only desirable but necessary. Therefore, in this work, we propose two different ways to incorporate it into memristor simulation models. The first is an extension of Strukov's model, and the second is an equivalent Wiener model approximation. We analyze and compare the dynamical properties of these models and discuss their implications for the memory and the nonlinear processing capacity of memristor networks. Our results indicate that device variability, increasingly causing problems in traditional computer design, is an asset in the context of reservoir computing. We conclude that although both models could lead to useful memristor-based reservoir computing systems, their computational performance will differ. Therefore, experimental modeling research is required for the development of accurate volatile memristor models.

True zero-training brain-computer interfacing--an online study.

Despite several approaches to realize subject-to-subject transfer of pre-trained classifiers, the full performance of a Brain-Computer Interface (BCI) for a novel user can only be reached by presenting the BCI system with data from the novel user. In typical state-of-the-art BCI systems with a supervised classifier, the labeled data is collected during a calibration recording, in which the user is asked to perform a specific task. Based on the known labels of this recording, the BCI's classifier can learn to decode the individual's brain signals. Unfortunately, this calibration recording consumes valuable time. Furthermore, it is unproductive with respect to the final BCI application, e.g. text entry. Therefore, the calibration period must be reduced to a minimum, which is especially important for patients with a limited concentration ability. The main contribution of this manuscript is an online study on unsupervised learning in an auditory event-related potential (ERP) paradigm. Our results demonstrate that the calibration recording can be bypassed by utilizing an unsupervised trained classifier, that is initialized randomly and updated during usage. Initially, the unsupervised classifier tends to make decoding mistakes, as the classifier might not have seen enough data to build a reliable model. Using a constant re-analysis of the previously spelled symbols, these initially misspelled symbols can be rectified posthoc when the classifier has learned to decode the signals. We compare the spelling performance of our unsupervised approach and of the unsupervised posthoc approach to the standard supervised calibration-based dogma for n = 10 healthy users. To assess the learning behavior of our approach, it is unsupervised trained from scratch three times per user. Even with the relatively low SNR of an auditory ERP paradigm, the results show that after a limited number of trials (30 trials), the unsupervised approach performs comparably to a classic supervised model.

Design and control of compliant tensegrity robots through simulation and hardware validation.

To better understand the role of tensegrity structures in biological systems and their application to robotics, the Dynamic Tensegrity Robotics Lab at NASA Ames Research Center, Moffett Field, CA, USA, has developed and validated two software environments for the analysis, simulation and design of tensegrity robots. These tools, along with new control methodologies and the modular hardware components developed to validate them, are presented as a system for the design of actuated tensegrity structures. As evidenced from their appearance in many biological systems, tensegrity ('tensile-integrity') structures have unique physical properties that make them ideal for interaction with uncertain environments. Yet, these characteristics make design and control of bioinspired tensegrity robots extremely challenging. This work presents the progress our tools have made in tackling the design and control challenges of spherical tensegrity structures. We focus on this shape since it lends itself to rolling locomotion. The results of our analyses include multiple novel control approaches for mobility and terrain interaction of spherical tensegrity structures that have been tested in simulation. A hardware prototype of a spherical six-bar tensegrity, the Reservoir Compliant Tensegrity Robot, is used to empirically validate the accuracy of simulation.

Nanophotonic reservoir computing with photonic crystal cavities to generate periodic patterns.

Reservoir computing (RC) is a technique in machine learning inspired by neural systems. RC has been used successfully to solve complex problems such as signal classification and signal generation. These systems are mainly implemented in software, and thereby they are limited in speed and power efficiency. Several optical and optoelectronic implementations have been demonstrated, in which the system has signals with an amplitude and phase. It is proven that these enrich the dynamics of the system, which is beneficial for the performance. In this paper, we introduce a novel optical architecture based on nanophotonic crystal cavities. This allows us to integrate many neurons on one chip, which, compared with other photonic solutions, closest resembles a classical neural network. Furthermore, the components are passive, which simplifies the design and reduces the power consumption. To assess the performance of this network, we train a photonic network to generate periodic patterns, using an alternative online learning rule called first-order reduced and corrected error. For this, we first train a classical hyperbolic tangent reservoir, but then we vary some of the properties to incorporate typical aspects of a photonics reservoir, such as the use of continuous-time versus discrete-time signals and the use of complex-valued versus real-valued signals. Then, the nanophotonic reservoir is simulated and we explore the role of relevant parameters such as the topology, the phases between the resonators, the number of nodes that are biased and the delay between the resonators. It is important that these parameters are chosen such that no strong self-oscillations occur. Finally, our results show that for a signal generation task a complex-valued, continuous-time nanophotonic reservoir outperforms a classical (i.e., discrete-time, real-valued) leaky hyperbolic tangent reservoir (normalized root-mean-square errors=0.030 versus NRMSE=0.127).

Integrating dynamic stopping, transfer learning and language models in an adaptive zero-training ERP speller.

OBJECTIVE: Most BCIs have to undergo a calibration session in which data is recorded to train decoders with machine learning. Only recently zero-training methods have become a subject of study. This work proposes a probabilistic framework for BCI applications which exploit event-related potentials (ERPs). For the example of a visual P300 speller we show how the framework harvests the structure suitable to solve the decoding task by (a) transfer learning, (b) unsupervised adaptation, (c) language model and (d) dynamic stopping. APPROACH: A simulation study compares the proposed probabilistic zero framework (using transfer learning and task structure) to a state-of-the-art supervised model on n = 22 subjects. The individual influence of the involved components (a)-(d) are investigated. MAIN RESULTS: Without any need for a calibration session, the probabilistic zero-training framework with inter-subject transfer learning shows excellent performance--competitive to a state-of-the-art supervised method using calibration. Its decoding quality is carried mainly by the effect of transfer learning in combination with continuous unsupervised adaptation. SIGNIFICANCE: A high-performing zero-training BCI is within reach for one of the most popular BCI paradigms: ERP spelling. Recording calibration data for a supervised BCI would require valuable time which is lost for spelling. The time spent on calibration would allow a novel user to spell 29 symbols with our unsupervised approach. It could be of use for various clinical and non-clinical ERP-applications of BCI.

Toward unified hybrid simulation techniques for spiking neural networks.

In the field of neural network simulation techniques, the common conception is that spiking neural network simulators can be divided in two categories: time-step-based and event-driven methods. In this letter, we look at state-of-the art simulation techniques in both categories and show that a clear distinction between both methods is increasingly difficult to define. In an attempt to improve the weak points of each simulation method, ideas of the alternative method are, sometimes unknowingly, incorporated in the simulation engine. Clearly the ideal simulation method is a mix of both methods. We formulate the key properties of such an efficient and generally applicable hybrid approach.

Experimental demonstration of reservoir computing on a silicon photonics chip.

In today's age, companies employ machine learning to extract information from large quantities of data. One of those techniques, reservoir computing (RC), is a decade old and has achieved state-of-the-art performance for processing sequential data. Dedicated hardware realizations of RC could enable speed gains and power savings. Here we propose the first integrated passive silicon photonics reservoir. We demonstrate experimentally and through simulations that, thanks to the RC paradigm, this generic chip can be used to perform arbitrary Boolean logic operations with memory as well as 5-bit header recognition up to 12.5 Gbit s(-1), without power consumption in the reservoir. It can also perform isolated spoken digit recognition. Our realization exploits optical phase for computing. It is scalable to larger networks and much higher bitrates, up to speeds >100 Gbit s(-1). These results pave the way for the application of integrated photonic RC for a wide range of applications.

Frequency modulation of large oscillatory neural networks.

Dynamical systems which generate periodic signals are of interest as models of biological central pattern generators and in a number of robotic applications. A basic functionality that is required in both biological modelling and robotics is frequency modulation. This leads to the question of whether there are generic mechanisms to control the frequency of neural oscillators. Here we describe why this objective is of a different nature, and more difficult to achieve, than modulating other oscillation characteristics (like amplitude, offset, signal shape). We propose a generic way to solve this task which makes use of a simple linear controller. It rests on the insight that there is a bidirectional dependency between the frequency of an oscillation and geometric properties of the neural oscillator's phase portrait. By controlling the geometry of the neural state orbits, it is possible to control the frequency on the condition that the state space can be shaped such that it can be pushed easily to any frequency.

Automated design of complex dynamic systems.

Several fields of study are concerned with uniting the concept of computation with that of the design of physical systems. For example, a recent trend in robotics is to design robots in such a way that they require a minimal control effort. Another example is found in the domain of photonics, where recent efforts try to benefit directly from the complex nonlinear dynamics to achieve more efficient signal processing. The underlying goal of these and similar research efforts is to internalize a large part of the necessary computations within the physical system itself by exploiting its inherent non-linear dynamics. This, however, often requires the optimization of large numbers of system parameters, related to both the system's structure as well as its material properties. In addition, many of these parameters are subject to fabrication variability or to variations through time. In this paper we apply a machine learning algorithm to optimize physical dynamic systems. We show that such algorithms, which are normally applied on abstract computational entities, can be extended to the field of differential equations and used to optimize an associated set of parameters which determine their behavior. We show that machine learning training methodologies are highly useful in designing robust systems, and we provide a set of both simple and complex examples using models of physical dynamical systems. Interestingly, the derived optimization method is intimately related to direct collocation a method known in the field of optimal control. Our work suggests that the application domains of both machine learning and optimal control have a largely unexplored overlapping area which envelopes a novel design methodology of smart and highly complex physical systems.

MACOP modular architecture with control primitives.

Walking, catching a ball and reaching are all tasks in which humans and animals exhibit advanced motor skills. Findings in biological research concerning motor control suggest a modular control hierarchy which combines movement/motor primitives into complex and natural movements. Engineers inspire their research on these findings in the quest for adaptive and skillful control for robots. In this work we propose a modular architecture with control primitives (MACOP) which uses a set of controllers, where each controller becomes specialized in a subregion of its joint and task-space. Instead of having a single controller being used in this subregion [such as MOSAIC (modular selection and identification for control) on which MACOP is inspired], MACOP relates more to the idea of continuously mixing a limited set of primitive controllers. By enforcing a set of desired properties on the mixing mechanism, a mixture of primitives emerges unsupervised which successfully solves the control task. We evaluate MACOP on a numerical model of a robot arm by training it to generate desired trajectories. We investigate how the tracking performance is affected by the number of controllers in MACOP and examine how the individual controllers and their generated control primitives contribute to solving the task. Furthermore, we show how MACOP compensates for the dynamic effects caused by a fixed control rate and the inertia of the robot.

A unified probabilistic approach to improve spelling in an event-related potential-based brain-computer interface.

In recent years, in an attempt to maximize performance, machine learning approaches for event-related potential (ERP) spelling have become more and more complex. In this paper, we have taken a step back as we wanted to improve the performance without building an overly complex model, that cannot be used by the community. Our research resulted in a unified probabilistic model for ERP spelling, which is based on only three assumptions and incorporates language information. On top of that, the probabilistic nature of our classifier yields a natural dynamic stopping strategy. Furthermore, our method uses the same parameters across 25 subjects from three different datasets. We show that our classifier, when enhanced with language models and dynamic stopping, improves the spelling speed and accuracy drastically. Additionally, we would like to point out that as our model is entirely probabilistic, it can easily be used as the foundation for complex systems in future work. All our experiments are executed on publicly available datasets to allow for future comparison with similar techniques.

Event management for large scale event-driven digital hardware spiking neural networks.

The interest in brain-like computation has led to the design of a plethora of innovative neuromorphic systems. Individually, spiking neural networks (SNNs), event-driven simulation and digital hardware neuromorphic systems get a lot of attention. Despite the popularity of event-driven SNNs in software, very few digital hardware architectures are found. This is because existing hardware solutions for event management scale badly with the number of events. This paper introduces the structured heap queue, a pipelined digital hardware data structure, and demonstrates its suitability for event management. The structured heap queue scales gracefully with the number of events, allowing the efficient implementation of large scale digital hardware event-driven SNNs. The scaling is linear for memory, logarithmic for logic resources and constant for processing time. The use of the structured heap queue is demonstrated on a field-programmable gate array (FPGA) with an image segmentation experiment and a SNN of 65,536 neurons and 513,184 synapses. Events can be processed at the rate of 1 every 7 clock cycles and a 406×158 pixel image is segmented in 200 ms.

Real-time detection of epileptic seizures in animal models using reservoir computing.

In recent years, an increasing number of studies have investigated the effects of closed-loop anti-epileptic treatments. Most of the current research still is very labour intensive: real-time treatment is manually triggered and conclusions can only be drawn after multiple days of manual review and annotation of the electroencephalogram (EEG). In this paper we propose a technique based on reservoir computing (RC) to automatically and in real-time detect epileptic seizures in the intra-cranial EEG (iEEG) of epileptic rats in order to immediately trigger seizure treatment. The performance of the system is evaluated in two different seizure types: absence seizures from genetic absence epilepsy rats from Strasbourg (GAERS) and limbic seizures from post status epilepticus (PSE) rats. The dataset consists of 452 hours iEEG from 23 GAERS and 2083 hours iEEG from 22 PSE rats. In the default set-up the system detects 0.09 and 0.13 false positives per seizure and misses 0.07 and 0.005 events per seizure for GAERS and PSE rats respectively. It achieves an average detection delay below 1s in GAERS and less than 10s in the PSE data. This detection delay and the number of missed seizures can be further decreased when a higher false positive rate is allowed. Our method outperforms state-of-the-art detection techniques and only a few parameters require optimization on a limited training set. It is therefore suited for automatic seizure detection based on iEEG and may serve as a useful tool for epilepsy researchers. The technique avoids the time-consuming manual review and annotation of EEG and can be incorporated in a closed-loop treatment strategy.

Cascadable excitability in microrings.

To emulate a spiking neuron, a photonic component needs to be excitable. In this paper, we theoretically simulate and experimentally demonstrate cascadable excitability near a self-pulsation regime in high-Q-factor silicon-on-insulator microrings. For the theoretical study we use Coupled Mode Theory. While neglecting the fast energy and phase dynamics of the cavity light, we can still preserve the most important microring dynamics, by only keeping the temperature difference with the surroundings and the amount of free carriers as dynamical variables of the system. Therefore we can analyse the microring dynamics in a 2D phase portrait. For some wavelengths, when changing the input power, the microring undergoes a subcritical Andronov-Hopf bifurcation at the self-pulsation onset. As a consequence the system shows class II excitability. Experimental single ring excitability and self-pulsation behaviour follows the theoretic predictions. Moreover, simulations and experiments show that this excitation mechanism is cascadable.

Information processing capacity of dynamical systems.

Many dynamical systems, both natural and artificial, are stimulated by time dependent external signals, somehow processing the information contained therein. We demonstrate how to quantify the different modes in which information can be processed by such systems and combine them to define the computational capacity of a dynamical system. This is bounded by the number of linearly independent state variables of the dynamical system, equaling it if the system obeys the fading memory condition. It can be interpreted as the total number of linearly independent functions of its stimuli the system can compute. Our theory combines concepts from machine learning (reservoir computing), system modeling, stochastic processes, and functional analysis. We illustrate our theory by numerical simulations for the logistic map, a recurrent neural network, and a two-dimensional reaction diffusion system, uncovering universal trade-offs between the non-linearity of the computation and the system's short-term memory.

A bayesian model for exploiting application constraints to enable unsupervised training of a P300-based BCI.

This work introduces a novel classifier for a P300-based speller, which, contrary to common methods, can be trained entirely unsupervisedly using an Expectation Maximization approach, eliminating the need for costly dataset collection or tedious calibration sessions. We use publicly available datasets for validation of our method and show that our unsupervised classifier performs competitively with supervised state-of-the-art spellers. Finally, we demonstrate the added value of our method in different experimental settings which reflect realistic usage situations of increasing difficulty and which would be difficult or impossible to tackle with existing supervised or adaptive methods.

Learning slow features with reservoir computing for biologically-inspired robot localization.

This work proposes a hierarchical biologically-inspired architecture for learning sensor-based spatial representations of a robot environment in an unsupervised way. The first layer is comprised of a fixed randomly generated recurrent neural network, the reservoir, which projects the input into a high-dimensional, dynamic space. The second layer learns instantaneous slowly-varying signals from the reservoir states using Slow Feature Analysis (SFA), whereas the third layer learns a sparse coding on the SFA layer using Independent Component Analysis (ICA). While the SFA layer generates non-localized activations in space, the ICA layer presents high place selectivity, forming a localized spatial activation, characteristic of place cells found in the hippocampus area of the rodent's brain. We show that, using a limited number of noisy short-range distance sensors as input, the proposed system learns a spatial representation of the environment which can be used to predict the actual location of simulated and real robots, without the use of odometry. The results confirm that the reservoir layer is essential for learning spatial representations from low-dimensional input such as distance sensors. The main reason is that the reservoir state reflects the recent history of the input stream. Thus, this fading memory is essential for detecting locations, mainly when locations are ambiguous and characterized by similar sensor readings.