Showing papers in "Frontiers in Neurorobotics in 2020"

Current Status, Challenges, and Possible Solutions of EEG-Based Brain-Computer Interface: A Comprehensive Review

Abstract: Brain-Computer Interface (BCI), in essence, aims at controlling different assistive devices through the utilization of brain waves. It is worth noting that the application of BCI is not limited to medical applications, and hence, the research in this field has gained due attention. Moreover, the significant number of related publications over the past two decades further indicates the consistent improvements and breakthroughs that have been made in this particular field. Nonetheless, it is also worth mentioning that with these improvements, new challenges are constantly discovered. This article provides a comprehensive review of the state-of-the-art of a complete BCI system. First, a brief overview of electroencephalogram (EEG)-based BCI systems is given. Secondly, a considerable number of popular BCI applications are reviewed in terms of electrophysiological control signals, feature extraction, classification algorithms, and performance evaluation metrics. Finally, the challenges to the recent BCI systems are discussed, and possible solutions to mitigate the issues are recommended.

A Review of Robot-Assisted Lower-Limb Stroke Therapy: Unexplored Paths and Future Directions in Gait Rehabilitation.

Abstract: Stroke affects one out of every six people on Earth. Approximately 90% of stroke survivors have some functional disability with mobility being a major impairment, which not only affects important daily activities but also increases the likelihood of falling. Originally intended to supplement traditional post-stroke gait rehabilitation, robotic systems have gained remarkable attention in recent years as a tool to decrease the strain on physical therapists while increasing the precision and repeatability of the therapy. While some of the current methods for robot-assisted rehabilitation have had many positive and promising outcomes, there is moderate evidence of improvement in walking and motor recovery using robotic devices compared to traditional practice. In order to better understand how and where robot-assisted rehabilitation has been effective, it is imperative to identify the main schools of thought that have prevailed. This review intends to observe those perspectives through three different lenses: the goal and type of interaction, the physical implementation, and the sensorimotor pathways targeted by robotic devices. The ways that researchers approach the problem of restoring gait function are grouped together in an intuitive way. Seeing robot-assisted rehabilitation in this unique light can naturally provoke the development of new directions to potentially fill the current research gaps and eventually discover more effective ways to provide therapy. In particular, the idea of utilizing the human inter-limb coordination mechanisms is brought up as an especially promising area for rehabilitation and is extensively discussed.

The Path Planning of Mobile Robot by Neural Networks and Hierarchical Reinforcement Learning.

Abstract: Existing mobile robots cannot complete some functions. To solve these problems, which include autonomous learning in path planning, the slow convergence of path planning, and planned paths that are not smooth, it is possible to utilize neural networks to enable to the robot to perceive the environment and perform feature extraction, which enables them to have a fitness of environment to state action function. By mapping the current state of these actions through Hierarchical Reinforcement Learning (HRL), the needs of mobile robots are met. It is possible to construct a path planning model for mobile robots based on neural networks and HRL. In this article, the proposed algorithm is compared with different algorithms in path planning. It underwent a performance evaluation to obtain an optimal learning algorithm system. The optimal algorithm system was tested in different environments and scenarios to obtain optimal learning conditions, thereby verifying the effectiveness of the proposed algorithm. Deep Deterministic Policy Gradient (DDPG), a path planning algorithm for mobile robots based on neural networks and hierarchical reinforcement learning, performed better in all aspects than other algorithms. Specifically, when compared with Double Deep Q-Learning (DDQN), DDPG has a shorter path planning time and a reduced number of path steps. When introducing an influence value, this algorithm shortens the convergence time by 91% compared with the Q-learning algorithm and improves the smoothness of the planned path by 79%. The algorithm has a good generalization effect in different scenarios. These results have significance for research on guiding, the precise positioning, and path planning of mobile robots.

PD Control Compensation Based on a Cascade Neural Network Applied to a Robot Manipulator.

Abstract: A Proportional Integral Derivative (PID) controller is commonly used to carry out tasks like position tracking in the industrial robot manipulator controller; however, over time, the PID integral gain generates degradation within the controller, which then produces reduced stability and bandwidth. A proportional derivative (PD) controller has been proposed to deal with the increase in integral gain but is limited if gravity is not compensated for. In practice, the dynamic system non-linearities frequently are unknown or hard to obtain. Adaptive controllers are online schemes that are used to deal with systems that present non-linear and uncertainties dynamics. Adaptive controller use measured data of system trajectory in order to learn and compensate the uncertainties and external disturbances. However, these techniques can adopt more efficient learning methods in order to improve their performance. In this work, a nominal control law is used to achieve a sub-optimal performance, and a scheme based on a cascade neural network is implemented to act as a non-linear compensation whose task is to improve upon the performance of the nominal controller. The main contributions of this work are neural compensation based on a cascade neural networks and the function to update the weights of neural network used. The algorithm is implemented using radial basis function neural networks and a recompense function that leads longer traces for an identification problem. A two-degree-of-freedom robot manipulator is proposed to validate the proposed scheme and compare it with conventional PD control compensation.

A Piezoresistive Array Armband With Reduced Number of Sensors for Hand Gesture Recognition.

Abstract: Human machine interfaces (HMIs) are employed in a broad range of applications, spanning from assistive devices for disability to remote manipulation and gaming controllers. In this study, a new piezoresistive sensors array armband is proposed for hand gesture recognition. The armband encloses only three sensors targeting specific forearm muscles, with the aim to discriminate eight hand movements. Each sensor is made by a force-sensitive resistor (FSR) with a dedicated mechanical coupler and is designed to sense muscle swelling during contraction. The armband is designed to be easily wearable and adjustable for any user and was tested on 10 volunteers. Hand gestures are classified by means of different machine learning algorithms, and classification performances are assessed applying both, the 10-fold and leave-one-out cross-validations. A linear support vector machine provided 96% mean accuracy across all participants. Ultimately, this classifier was implemented on an Arduino platform and allowed successful control for videogames in real-time. The low power consumption together with the high level of accuracy suggests the potential of this device for exergames commonly employed for neuromotor rehabilitation. The reduced number of sensors makes this HMI also suitable for hand-prosthesis control.

Nengo and Low-Power AI Hardware for Robust, Embedded Neurorobotics.

Abstract: In this paper we demonstrate how the Nengo neural modeling and simulation libraries enable users to quickly develop robotic perception and action neural networks for simulation on neuromorphic hardware using tools they are already familiar with, such as Keras and Python. We identify four primary challenges in building robust, embedded neurorobotic systems, including: (1) developing infrastructure for interfacing with the environment and sensors; (2) processing task specific sensory signals; (3) generating robust, explainable control signals; and (4) compiling neural networks to run on target hardware. Nengo helps to address these challenges by: (1) providing the NengoInterfaces library, which defines a simple but powerful API for users to interact with simulations and hardware; (2) providing the NengoDL library, which lets users use the Keras and TensorFlow API to develop Nengo models; (3) implementing the Neural Engineering Framework, which provides white-box methods for implementing known functions and circuits; and (4) providing multiple backend libraries, such as NengoLoihi, that enable users to compile the same model to different hardware. We present two examples using Nengo to develop neural networks that run on CPUs and GPUs as well as Intel's neuromorphic chip, Loihi, to demonstrate two variations on this workflow. The first example is an implementation of an end-to-end spiking neural network in Nengo that controls a rover simulated in Mujoco. The network integrates a deep convolutional network that processes visual input from cameras mounted on the rover to track a target, and a control system implementing steering and drive functions in connection weights to guide the rover to the target. The second example uses Nengo as a smaller component in a system that has addressed some but not all of those challenges. Specifically it is used to augment a force-based operational space controller with neural adaptive control to improve performance during a reaching task using a real-world Kinova Jaco2 robotic arm. The code and implementation details are provided, with the intent of enabling other researchers to build and run their own neurorobotic systems.

A Path Toward Explainable AI and Autonomous Adaptive Intelligence: Deep Learning, Adaptive Resonance, and Models of Perception, Emotion, and Action.

Abstract: Biological neural network models whereby brains make minds help to understand autonomous adaptive intelligence This article summarizes why the dynamics and emergent properties of such models for perception, cognition, emotion, and action are explainable, and thus amenable to being confidently implemented in large-scale applications Key to their explainability is how these models combine fast activations, or short-term memory (STM) traces, and learned weights, or long-term memory (LTM) traces Visual and auditory perceptual models have explainable conscious STM representations of visual surfaces and auditory streams in surface-shroud resonances and stream-shroud resonances, respectively Deep Learning is often used to classify data However, Deep Learning can experience catastrophic forgetting: At any stage of learning, an unpredictable part of its memory can collapse Even if it makes some accurate classifications, they are not explainable and thus cannot be used with confidence Deep Learning shares these problems with the back propagation algorithm, whose computational problems due to non-local weight transport during mismatch learning were described in the 1980s Deep Learning became popular after very fast computers and huge online databases became available that enabled new applications despite these problems Adaptive Resonance Theory, or ART, algorithms overcome the computational problems of back propagation and Deep Learning ART is a self-organizing production system that incrementally learns, using arbitrary combinations of unsupervised and supervised learning and only locally computable quantities, to rapidly classify large non-stationary databases without experiencing catastrophic forgetting ART classifications and predictions are explainable using the attended critical feature patterns in STM on which they build The LTM adaptive weights of the fuzzy ARTMAP algorithm induce fuzzy IF-THEN rules that explain what feature combinations predict successful outcomes ART has been successfully used in multiple large-scale real world applications, including remote sensing, medical database prediction, and social media data clustering Also explainable are the MOTIVATOR model of reinforcement learning and cognitive-emotional interactions, and the VITE, DIRECT, DIVA, and SOVEREIGN models for reaching, speech production, spatial navigation, and autonomous adaptive intelligence These biological models exemplify complementary computing, and use local laws for match learning and mismatch learning that avoid the problems of Deep Learning

Neuromusculoskeletal Arm Prostheses: Personal and Social Implications of Living With an Intimately Integrated Bionic Arm

Abstract: People with limb loss are for the first time living chronically and uninterruptedly with intimately integrated neuromusculoskeletal prostheses. This new generation of artificial limbs are fixated to the skeleton and operated by bidirectionally transferred neural information. This unprecedented level of human-machine integration is bound to have profound psychosocial effects on the individuals living with these prostheses. Here, we examined the psychosociological impact on people as they integrate neuromusculoskeletal prostheses into their bodies and lives. Three people with transhumeral amputations participated in this study, all of whom had been living with neuromusculoskeletal prostheses in their daily lives between 2 and 6 years at the time of the interview. Direct neural sensory feedback had been enabled for 6 months to 2 years. Participants were interviewed about their experiences living with the neuromusculoskeletal prostheses in their home and professional daily lives. We analyzed these interviews to elucidate themes using an interpretive phenomenological approach that regards participants' own experiences as forms of expertise and knowledge-making. Our participant-generated results indicate that people adapted and integrated the technology into functional and social arenas of daily living, with positive psychosocial effects on self-esteem, self-image, and social relations intimately linked to improved trust of the prostheses. Participants expressed enhanced prosthetic function, increased and more diverse prosthesis use in tasks of daily living, and improved relationships between their prosthesis and phantom limb. Our interviews with patients also generated critiques of the language commonly used to describe human-prosthetic relations, including terms such as "embodiment," and the need for specificity surrounding the term "natural" with regard to control versus sensory feedback. Experiences living with neuromusculoskeletal prostheses were complex and subject-dependent, and therefore future research should consider human-machine interaction as a relationship that is constantly enacted, negotiated, and deeply contextualized.

Path Planning of Mobile Robot With Improved Ant Colony Algorithm and MDP to Produce Smooth Trajectory in Grid-Based Environment

Abstract: This approach has been derived mainly to improve quality and efficiency of global path planning for a mobile robot with unknown static obstacle avoidance features in grid-based environment. The quality of the global path in terms of smoothness, path consistency and safety can affect the autonomous behavior of a robot. In this paper, the efficiency of Ant Colony Optimization (ACO) algorithm has improved with additional assistance of A* Multi-Directional algorithm. In the first part, A* Multi-directional algorithm starts to search in map and stores the best nodes area between start and destination with optimal heuristic value and that area of nodes has been chosen for path search by ACO to avoid blind search at initial iterations. The path obtained in grid-based environment consist of points in Cartesian coordinates connected through line segments with sharp bends. Therefore, Markov Decision Process (MDP) trajectory evaluation model is introduced with a novel reward policy to filter and reduce the sharpness in global path generated in grid environment. With arc-length parameterization, a curvilinear smooth route has been generated among filtered waypoints and produces consistency and smoothness in the global path. To achieve a comfort drive and safety for robot, lateral and longitudinal control has been utilized to form a set of optimal trajectories along the reference route, as well as, minimizing total cost. The total cost includes curvature, lateral and longitudinal coordinates constraints. Additionally, for collision detection, at every step the set of optimal local trajectories have been checked for any unexpected obstacle. The results have been verified through simulations in MATLAB compared with previous global path planning algorithms to differentiate the efficiency and quality of derived approach in different constraint environments.

Pneumatic Quasi-Passive Actuation for Soft Assistive Lower Limbs Exoskeleton

Abstract: There is a growing international interest in developing soft wearable robotic devices to improve mobility and daily life autonomy as well as for rehabilitation purposes. Usability, comfort and acceptance of such devices will affect their uptakes in mainstream daily life. The XoSoft EU project developed a modular soft lower-limb exoskeleton to assist people with low mobility impairments. This paper presents the bio-inspired design of a soft, modular exoskeleton for lower limb assistance based on pneumatic quasi-passive actuation. The design of a modular reconfigurable prototype and its performance are presented. This actuation centers on an active mechanical element to modulate the assistance generated by a traditional passive component, in this case an elastic belt. This study assesses the feasibility of this type of assistive device by evaluating the energetic outcomes on a healthy subject during a walking task. Human-exoskeleton interaction in relation to task-based biological power assistance and kinematics variations of the gait are evaluated. The resultant assistance, in terms of overall power ratio (Λ) between the exoskeleton and the assisted joint, was 26.6% for hip actuation, 9.3% for the knee and 12.6% for the ankle. The released maximum power supplied on each articulation, was 113.6% for the hip, 93.2% for the knee, and 150.8% for the ankle.

A Hybrid BCI Based on SSVEP and EOG for Robotic Arm Control.

Abstract: Brain-computer interface (BCI) for robotic arm control has been studied to improve the life quality of people with severe motor disabilities. There are still challenges for robotic arm control in accomplishing a complex task with a series of actions. An efficient switch and a timely cancel command are helpful in the application of robotic arm. Based on the above, we proposed an asynchronous hybrid BCI in this study. The basic control of a robotic arm with six degrees of freedom was a steady-state visual evoked potential (SSVEP) based BCI with fifteen target classes. We designed an EOG-based switch which used a triple blink to either activate or deactivate the flash of SSVEP-based BCI. Stopping flash in the idle state can help to reduce visual fatigue and false activation rate (FAR). Additionally, users were allowed to cancel the current command simply by a wink in the feedback phase to avoid executing the incorrect command. Fifteen subjects participated and completed the experiments. The cue-based experiment obtained an average accuracy of 92.09%, and the information transfer rates (ITR) resulted in 35.98 bits/min. The mean FAR of the switch was 0.01/min. Furthermore, all subjects succeeded in asynchronously operating the robotic arm to grasp, lift, and move a target object from the initial position to a specific location. The results indicated the feasibility of the combination of EOG and SSVEP signals and the flexibility of EOG signal in BCI to complete a complicated task of robotic arm control.

CNN Based Detectors on Planetary Environments: A Performance Evaluation.

Abstract: An essential characteristic that an exploration robot must possess is to be autonomous. This is necessary because it will usually do its task in remote or hard-to-reach places. One of the primary elements of a navigation system is the information that can be acquired by the sensors of the environment in which it will operate. For this reason, an algorithm based on convolutional neural networks is proposed for the detection of rocks in environments similar to Mars. The methodology proposed here is based on the use of a Single-Shot-Detector (SSD) network architecture, which has been modified to evaluate the performance. The main contribution of this study is to provide an alternative methodology to detect rocks in planetary images because most of the previous works only focus on classification problems and used handmade feature vectors.

Design and Performance Evaluation of a Novel Wearable Parallel Mechanism for Ankle Rehabilitation.

Abstract: Repetitive and intensive physiotherapy is indispensable to patients with ankle disabilities. Increasingly robot-assisted technology has been employed in the treatment to reduce the burden of the therapists and the related costs of the patients. This paper proposes a configuration of a wearable parallel mechanism to supplement the equipment selection for ankle rehabilitation. The kinematic analysis, i.e., the inverse position solution and Jacobian matrices, is elaborated. Several performance indices, including the reachable workspace index, motion isotropy index, force transfer index, and maximum torque index, are developed based on the derived kinematic solution. Moreover, according to the proposed kinematic configuration and wearable design concept, the mechanical structure that contains a basic machine-drive system and a multi-model position/force data collection system is designed in detail. Finally, the results of the performance evaluation indicate that the wearable parallel robot possesses sufficient motion isotropy, high force transfer performance, and large maximum torque performance within a large workspace that can cover all possible range of motion of human ankle complex, and is suitable for ankle rehabilitation.

Hybrid Human-Machine Interface for Gait Decoding Through Bayesian Fusion of EEG and EMG Classifiers.

Abstract: Despite the advances in the field of brain computer interfaces (BCI), the use of the sole electroencephalography (EEG) signal to control walking rehabilitation devices is currently not viable in clinical settings, due to its unreliability. Hybrid interfaces (hHMIs) represent a very recent solution to enhance the performance of single-signal approaches. These are classification approaches that combine multiple human-machine interfaces, normally including at least one BCI with other biosignals, such as the electromyography (EMG). However, their use for the decoding of gait activity is still limited. In this work, we propose and evaluate a hybrid human-machine interface (hHMI) to decode walking phases of both legs from the Bayesian fusion of EEG and EMG signals. The proposed hHMI significantly outperforms its single-signal counterparts, by providing high and stable performance even when the reliability of the muscular activity is compromised temporarily (e.g., fatigue) or permanently (e.g., weakness). Indeed, the hybrid approach shows a smooth degradation of classification performance after temporary EMG alteration, with more than 75% of accuracy at 30% of EMG amplitude, with respect to the EMG classifier whose performance decreases below 60% of accuracy. Moreover, the fusion of EEG and EMG information helps keeping a stable recognition rate of each gait phase of more than 80% independently on the permanent level of EMG degradation. From our study and findings from the literature, we suggest that the use of hybrid interfaces may be the key to enhance the usability of technologies restoring or assisting the locomotion on a wider population of patients in clinical applications and outside the laboratory environment.

Modeling and Control of a Cable-Driven Rotary Series Elastic Actuator for an Upper Limb Rehabilitation Robot

Abstract: This paper focuses on the design, modeling, and control of a novel remote actuation, including a compact rotary series elastic actuator (SEA) and Bowden cable. This kind of remote actuation is used for an upper limb rehabilitation robot (ULRR) with four powered degrees of freedom (DOFs). The SEA mainly consists of a DC motor with planetary gearheads, inner/outer sleeves, and eight linearly translational springs. The key innovations include (1) an encoder for direct spring displacement measurement, which can be used to calculate the output torque of SEA equivalently, (2) the embedded springs can absorb the negative impact of backlash on SEA control performance, (3) and the Bowden cable enables long-distance actuation and reduces the bulky structure on the robotic joint. In modeling of this actuation, the SEA's stiffness coefficient, the dynamics of the SEA, and the force transmission of the Bowden cable are considered for computing the inputs on each powered joint of the robot. Then, both torque and impedance controllers consisting of proportional-derivative (PD) feedback, disturbance observer (DOB), and feedforward compensation terms are developed. Simulation and experimental results verify the performance of these controllers. The preliminary results show that this new kind of actuation can not only implement stable and friendly actuation over a long distance but also be customized to meet the requirements of other robotic system design.

Processing Surface EMG Signals for Exoskeleton Motion Control.

Abstract: The surface electromyography (sEMG) signal has been used for volitional control of robotic assistive devices. There are still challenges in improving system performance accuracy and signal processing to remove systematic noise. This study presents procedures and a pilot validation of the EMG-driven speed-control of exoskeleton and integrated treadmill with a goal to provide better interaction between a user and the system. The gait cycle duration (GCD) was extracted from sEMG signals using the autocorrelation algorithm and Bayesian fusion algorithm. GCDs of various walking speeds were then programmed to control the motion speed of exoskeleton robotic system. The performance and efficiency of this sEMG-controlled robotic assistive ambulation system was tested and validated among 6 healthy volunteers. The results demonstrated that the autocorrelation algorithm extracted the GCD from individual muscle contraction. The GCDs of individual muscles had variability between different walking steps under a designated walking speed. Bayesian fusion algorithms processed the GCDs of multiple muscles yielding a final GCD with the least variance. The fused GCD effectively controlled the motion speeds of exoskeleton and treadmill. The higher amplitude of EMG signals with shorter GCD was found during a faster walking speed. The algorithms using fused GCDs and gait stride length yielded trajectory joint motion tracks in a shape of sine curve waveform. The joint angles of the exoskeleton measured by a decoder mounted on the hip turned out to be in sine waveforms. The hip joint motion track of the exoskeleton matched the angles projected by trajectory curve generated by computer algorithms based on the fused GCDs with high agreement. The EMG-driven speed-control provided the human-machine inter-limb coordination mechanisms for an intuitive speed control of the exoskeleton-treadmill system at the user's intents. Potentially the whole system can be used for gait rehabilitation of incomplete spinal cord hemispheric stroke patients as goal-directed and task-oriented training tool.

The Human Takes It All: Humanlike Synthesized Voices Are Perceived as Less Eerie and More Likable. Evidence From a Subjective Ratings Study

Abstract: Background: The increasing involvement of social robots in human lives raises the question how humans perceive social robots. Little is known about human perception of synthesized voices. Aim: To investigate which synthesized voice parameters predict the speaker’s eeriness and likability; to determine if individual listener characteristics (e.g., personality, attitude towards robots, age) influence synthesized voice evaluations; and to explore which paralinguistic features subjectively distinguish humans from robots/artificial agents. Methods: 95 adults (62 females) listened to randomly presented three categories of audio-clips: synthesized (IBM Watson), humanoid (robot Sophia, Hanson Robotics) and human voices (5 clips/category). Voices were rated on intelligibility, prosody, trustworthiness, confidence, enthusiasm, pleasantness, human-likeness and naturalness. Speakers were rated on appeal, credibility, human-likeness, likability and eeriness. Participants’ personality traits, attitudes to robots and demographics were obtained. Results: The human voice and human speaker characteristics received reliably higher scores on all dimensions except for eeriness. Synthesized voice ratings were positively related to participants’ agreeableness and neuroticism. Females rated synthesized voices more positively on most dimensions. Surprisingly, interest in social robots and attitudes to robots played almost no role in voice evaluation. Contrary to the expectations of an uncanny valley, the higher the ratings of the human-likeness for both the voice and the speaker characteristics were, the less eerie they seemed to the participants. Moreover, the more human-like the speaker voice was the more it was liked by the participants. The latter was only applicable to one of the synthesized voices. Finally, pleasantness and trustworthiness of the synthesized voice predicted the likability of the speaker. Qualitative content analysis identified intonation, sound, emotion and imageability/embodiment as diagnostic features. Discussion: Humans clearly prefer human voices, but manipulating diagnostic speech features might increase acceptance of synthesized voices and thereby support human-robot interaction. No evidence for an auditory uncanny valley was found.

Robotic Exoskeleton Gait Training During Acute Stroke Inpatient Rehabilitation.

Abstract: Stroke is the leading cause of severe disability in adults resulting in mobility, balance, and coordination deficits. Robotic exoskeletons (REs) for stroke rehabilitation can provide the user with consistent, high dose repetition of movement, as well as balance and stability. The goal of this intervention study is to evaluate the ability of a RE to provide high dose gait therapy and the resulting effect on functional recovery for individuals with acute stroke. The investigation included a total of 44 participants. Twenty-two participants received RE gait training during inpatient rehabilitation (RE+SOC Group), and a matched sample of 22 individuals admitted to the same inpatient rehabilitation facility-receiving conventional standard of care treatment (SOC group). The effect of RE training was quantified using total distance walked during inpatient rehabilitation and functional independence measure (FIM). The total distance walked during inpatient rehabilitation showed a significant difference between the SOC and RE+SOC groups. RE+SOC walked twice the distance as SOC during the same duration (time spent in inpatient rehabilitation) of training. In addition, the average change in motor FIM showed a significant difference between the SOC and RE+SOC groups, where the average difference in motor FIM was higher in RE+SOC compared to the SOC group. The results suggest that RE provided increased dosing of gait training without increasing the duration of training during acute stroke rehabilitation. The RE+SOC group increased their motor FIM score (change from admission to discharge) compared to SOC group, both groups were matched for admission motor FIM scores suggesting that increased dosing may have improved motor function.

Lizard Brain: Tackling Locally Low-Dimensional Yet Globally Complex Organization of Multi-Dimensional Datasets.

Abstract: Machine learning deals with datasets characterized by high dimensionality. However, in many cases, the intrinsic dimensionality of the datasets is surprisingly low. For example, the dimensionality of a robot's perception space can be large and multi-modal but its variables can have more or less complex non-linear interdependencies. Thus multidimensional data point clouds can be effectively located in the vicinity of principal varieties possessing locally small dimensionality, but having a globally complicated organization which is sometimes difficult to represent with regular mathematical objects (such as manifolds). We review modern machine learning approaches for extracting low-dimensional geometries from multi-dimensional data and their applications in various scientific fields.

Gait Neural Network for Human-Exoskeleton Interaction.

Abstract: Robotic exoskeletons are developed with the aim of enhancing convenience and physical possibilities in daily life. However, at present, these devices lack sufficient synchronization with human movements. To optimize human-exoskeleton interaction, this article proposes a gait recognition and prediction model, called the gait neural network (GNN), which is based on the temporal convolutional network. It consists of an intermediate network, a target network, and a recognition and prediction model. The novel structure of the algorithm can make full use of the historical information from sensors. The performance of the GNN is evaluated based on the publicly available HuGaDB dataset, as well as on data collected by an inertial-based wearable motion capture device. The results show that the proposed approach is highly effective and achieves superior performance compared with existing methods.

Flexible Spiking CPGs for Online Manipulation During Hexapod Walking

Abstract: Neural signals for locomotion are influenced both by the neural network architecture and sensory inputs coordinating and adapting the gait to the environment. Adaptation relies on the ability to change amplitude, frequency, and phase of the signals within the sensorimotor loop in response to external stimuli. However, in order to experiment with closed-loop control, we first need a better understanding of the dynamics of the system and how adaptation works. Based on insights from biology, we developed a spiking neural network capable of continuously changing amplitude, frequency, and phase online. The resulting network is deployed on a hexapod robot in order to observe the walking behavior. The morphology and parameters of the network results in a tripod gait, demonstrating that a design without afferent feedback is sufficient to maintain a stable gait. This is comparable to results from biology showing that deafferented samples exhibit a tripod-like gait \citep{mantziaris2017intra} and adds to the evidence for a meaningful role of network topology in locomotion. Further, this work enables research into the role of sensory feedback and high-level control signals in the adaptation of gait types. A better understanding of the neural control of locomotion relates back to biology where it can provide evidence for theories that are currently not testable on live insects. In engineering, the use of spiking neural networks for robot locomotion facilitates the use of low-power, neuromorphic hardware, enabling new applications on significantly smaller robots or direct interfacing with biological tissue.

Reduction of Onset Delay in Functional Near-Infrared Spectroscopy: Prediction of HbO/HbR Signals.

Abstract: An intrinsic problem when using hemodynamic responses for the brain-machine interface is the slow nature of the physiological process. In this paper, a novel method that estimates the oxyhemoglobin changes caused by neuronal activations is proposed and validated. In monitoring the time responses of blood-oxygen-level-dependent signals with functional near-infrared spectroscopy (fNIRS), the early trajectories of both oxy- and deoxy-hemoglobins in their phase space are scrutinized. Furthermore, to reduce the detection time, a prediction method based upon a kernel-based recursive least squares (KRLS) algorithm is implemented. In validating the proposed approach, the fNIRS signals of finger tapping tasks measured from the left motor cortex are examined. The results show that the KRLS algorithm using the Gaussian kernel yields the best fitting for both ΔHbO (i.e., 87.5%) and ΔHbR (i.e., 85.2%) at q = 15 steps ahead (i.e., 1.63 s ahead at a sampling frequency of 9.19 Hz). This concludes that a neuronal activation can be concluded in about 0.1 s with fNIRS using prediction, which enables an almost real-time practice if combined with EEG.

Statistical Inter-stimulus Interval Window Estimation for Transient Neuromodulation via Paired Mechanical and Brain Stimulation.

Abstract: 3 For achieving motor recovery in individuals with sensorimotor deficits, augmented activation of 4 the appropriate sensorimotor system and facilitated induction of neural plasticity are essential. 5 An emerging procedure that combines peripheral nerve stimulation and its associative stimulation 6 with central brain stimulation is known to enhance the excitability of the motor cortex. In order to 7 effectively apply this paired stimulation technique, timing between central and peripheral stimuli 8 must be individually adjusted. There is a small range of effective timings between two stimuli, or 9 the inter-stimulus interval window (ISI-W). Properties of ISI-W from neuromodulation in response 10 to mechanical stimulation (Mstim) of muscles have been understudied because of the absence of 11 a versatile and reliable mechanical stimulator. This paper adopted a combination of transcranial 12 magnetic stimulation (TMS) and Mstim by using a high-precision robotic mechanical stimulator. 13 A pneumatically operated robotic tendon tapping device was applied. A low-friction linear cylinder 14 achieved high stimulation precision in time and low electromagnetic artifacts in physiological 15 measurements. This paper describes a procedure to effectively estimate an individual ISI-W from 16 the transiently enhanced motor evoked potential (MEP) with a reduced number of paired Mstim 17 and sub-threshold TMS trials by applying statistical sampling and regression technique. This paper 18 applied a total of four parametric and nonparametric statistical regression methods for ISI-W 19 estimation. The developed procedure helps to reduce time for individually adjusting effective ISI, 20 reducing physical burden on the subject.

Prerequisites for an Artificial Self.

Abstract: Traditionally investigated in philosophy, body ownership and agency-two main components of the minimal self-have recently gained attention from other disciplines, such as brain, cognitive and behavioral sciences, and even robotics and artificial intelligence. In robotics, intuitive human interaction in natural and dynamic environments becomes more and more important, and requires skills such as self-other distinction and an understanding of agency effects. In a previous review article, we investigated studies on mechanisms for the development of motor and cognitive skills in robots (Schillaci et al., 2016). In this review article, we argue that these mechanisms also build the foundation for an understanding of an artificial self. In particular, we look at developmental processes of the minimal self in biological systems, transfer principles of those to the development of an artificial self, and suggest metrics for agency and body ownership in an artificial self.

Crossmodal Language Grounding in an Embodied Neurocognitive Model.

Abstract: Human infants are able to acquire natural language seemingly easily at an early age. Their language learning seems to occur simultaneously with learning other cognitive functions as well as with playful interactions with the environment and caregivers. From a neuroscientific perspective, natural language is embodied, grounded in most, if not all, sensory and sensorimotor modalities, and acquired by means of crossmodal integration. However, characterising the underlying mechanisms in the brain is difficult and explaining the grounding of language in crossmodal perception and action remains challenging. In this paper, we present a neurocognitive model for language grounding which reflects bio-inspired mechanisms such as an implicit adaptation of timescales as well as end-to-end multimodal abstraction. It addresses developmental robotic interaction and extends its learning capabilities using larger-scale knowledge-based data. In our scenario, we utilise the humanoid robot NICO in obtaining the EMIL data collection, in which the cognitive robot interacts with objects in a children's playground environment while receiving linguistic labels from a caregiver. The model analysis shows that crossmodally integrated representations are sufficient for acquiring language merely from sensory input through interaction with objects in an environment. The representations self-organise hierarchically and embed temporal and spatial information through composition and decomposition. This model can also provide the basis for further crossmodal integration of perceptually grounded cognitive representations.

An Enhanced Robot Massage System in Smart Homes Using Force Sensing and a Dynamic Movement Primitive.

Abstract: With the requirements for improving life quality, smart home and healthcare have gradually become a future lifestyle In particular, the service robots with human behavioural sensing for private or personal use in the home have attracted a lot of research attention thanks to their advantages in releasing the high labor costs and the fatigue of human assistance In this paper, a novel force sensing and robotic learning algorithm based teaching interface for robot massaging has been proposed For the teaching purposes, a human operator physically holds the end-effector of the robot to perform the demonstration At this stage, the end position data are outputted and sent to be segmented via the Finite Difference (FD) method Dynamic Movement Primitive (DMP) is utilized to model and generalize the human-like movements In order to learn from multiple demonstrations, Dynamic Time Warping (DTW) is used for the preprocessing of the data recorded on the robot platform, and Gaussian Mixture Model (GMM) is employed for the evaluation of DMP to generate multiple patterns after the completion of the teaching process After that, Gaussian Mixture Regression (GMR) algorithm is applied to generate a synthesized trajectory to minimize position errors Then a hybrid position/force controller is integrated to track the desired trajectory in the task space considering the safety of human-robot interaction The validation of our proposed method has been performed and proved by conducting massage tasks on a KUKA LBR iiwa robot platform

Surgeon-Centered Analysis of Robot-Assisted Needle Driving Under Different Force Feedback Conditions.

Abstract: Robotic assisted minimally invasive surgery (RAMIS) systems present many advantages to the surgeon and patient over open and standard laparoscopic surgery. However, haptic feedback, which is crucial for the success of many surgical procedures, is still an open challenge in RAMIS. Understanding the way that haptic feedback affects performance and learning can be useful in the development of haptic feedback algorithms and teleoperation control systems. In this study, we examined the performance and learning of inexperienced participants under different haptic feedback conditions in a task of surgical needle driving via a soft homogeneous deformable object-an artificial tissue. We designed an experimental setup to characterize their movement trajectories and the forces that they applied on the artificial tissue. Participants first performed the task in an open condition, with a standard surgical needle holder, followed by teleoperation in one of three feedback conditions: (1) no haptic feedback, (2) haptic feedback based on position exchange, and (3) haptic feedback based on direct recording from a force sensor, and then again with the open needle holder. To quantify the effect of different force feedback conditions on the quality of needle driving, we developed novel metrics that assess the kinematics of needle driving and the tissue interaction forces, and we combined our novel metrics with classical metrics. We analyzed the final teleoperated performance in each condition, the improvement during teleoperation, and the aftereffect of teleoperation on the performance when using the open needle driver. We found that there is no significant difference in the final performance and in the aftereffect between the 3 conditions. Only the two conditions with force feedback presented statistically significant improvement during teleoperation in several of the metrics, but when we compared directly between the improvements in the three different feedback conditions none of the effects reached statistical significance. We discuss possible explanations for the relative similarity in performance. We conclude that we developed several new metrics for the quality of surgical needle driving, but even with these detailed metrics, the advantage of state of the art force feedback methods to tasks that require interaction with homogeneous soft tissue is questionable.

Parking Slot Detection on Around-View Images Using DCNN.

Abstract: Due to the complex visual environment and incomplete display of parking slots on around-view images, vision-based parking slot detection is a major challenge. Previous studies in this field mostly use the existing models to solve the problem, the steps of which are cumbersome. In this paper, we propose a parking slot detection method that uses directional entrance line regression and classification based on a deep convolutional neural network (DCNN) to make it robust and simple. For parking slots with different shapes and observed from different angles, we represent the parking slot as a directional entrance line. Subsequently, we design a DCNN detector to simultaneously obtain the type, position, length, and direction of the entrance line. After that, the complete parking slot can be easily inferred using the detection results and prior geometric information. To verify our method, we conduct experiments on the public ps2.0 dataset and self-annotated parking slot dataset with 2,135 images. The results show that our method not only outperforms state-of-the-art competitors with a precision rate of 99.68% and a recall rate of 99.41% on the ps2.0 dataset but also performs a satisfying generalization on the self-annotated dataset. Moreover, it achieves a real-time detection speed of 13 ms per frame on Titan Xp. By converting the parking slot into a directional entrance line, the specially designed DCNN detector can quickly and effectively detect various types of parking slots.

Neurorobots as a Means Toward Neuroethology and Explainable AI.

Abstract: Understanding why deep neural networks and machine learning algorithms act as they do is a difficult endeavor. Neuroscientists are faced with similar problems. One way biologists address this issue is by closely observing behavior while recording neurons or manipulating brain circuits. This has been called neuroethology. In a similar way, neurorobotics can be used to explain how neural network activity leads to behavior. In real world settings, neurorobots have been shown to perform behaviors analogous to animals. Moreover, a neuroroboticist has total control over the network, and by analyzing different neural groups or studying the effect of network perturbations (e.g., simulated lesions), they may be able to explain how the robot's behavior arises from artificial brain activity. In this paper, we review neurorobot experiments by focusing on how the robot's behavior leads to a qualitative and quantitative explanation of neural activity, and vice versa, that is, how neural activity leads to behavior. We suggest that using neurorobots as a form of computational neuroethology can be a powerful methodology for understanding neuroscience, as well as for artificial intelligence and machine learning.