Alcimar Barbosa Soares

The Development of a Virtual Myoelectric Prosthesis Controlled by an EMG Pattern Recognition System Based on Neural Networks

One of the major difficulties faced by those who are fitted with prosthetic devices is the great mental effort needed during the first stages of training. When working with myoelectric prosthesis, that effort increases dramatically. In this sense, the authors decided to devise a mechanism to help patients during the learning stages, without actually having to wear the prosthesis. The system is based on a real hardware and software for detecting and processing electromyografic (EMG) signals. The association of autoregressive (AR) models and a neural network is used for EMG pattern discrimination. The outputs of the neural network are then used to control the movements of a virtual prosthesis that mimics what the real limb should be doing. This strategy resulted in rates of success of 100% when discriminating EMG signals collected from the upper arm muscle groups. The results show a very easy-to-use system that can greatly reduce the duration of the training stages.

On the use of Augmented Reality techniques in learning and interpretation of cardiologic data

Augmented Reality is a technology which provides people with more intuitive ways of interaction and visualization, close to those in real world. The amount of applications using Augmented Reality is growing every day, and results can be already seen in several fields such as Education, Training, Entertainment and Medicine. The system proposed in this article intends to provide a friendly and intuitive interface based on Augmented Reality for heart beating evaluation and visualization. Cardiologic data is loaded from several distinct sources: simple standards of heart beating frequencies (for example situations like running or sleeping), files of heart beating signals, scanned electrocardiographs and real time data acquisition of patients heart beating. All this data is processed to produce visualization within Augmented Reality environments. The results obtained in this research have shown that the developed system is able to simplify the understanding of concepts about heart beating and its functioning. Furthermore, the system can help health professionals in the task of retrieving, processing and converting data from all the sources handled by the system, with the support of an edition and visualization mode.

Classification of EMG signals using artificial neural networks for virtual hand prosthesis control

Computer-based training systems have been widely studied in the field of human rehabilitation. In health applications, Virtual Reality presents itself as an appropriate tool to simulate training environments without exposing the patients to risks. In particular, virtual prosthetic devices have been used to reduce the great mental effort needed by patients fitted with myoelectric prosthesis, during the training stage. In this paper, the application of Virtual Reality in a hand prosthesis training system is presented. To achieve this, the possibility of exploring Neural Networks in a real-time classification system is discussed. The classification technique used in this work resulted in a 95% success rate when discriminating 4 different hand movements.

On the use of Virtual and Augmented Reality for upper limb prostheses training and simulation

Accidents happen and unfortunately people may loose part of their body members. Studies have shown that in this case, most individuals suffer physically and psychologically. For this reason, actions to restore the patients freedom and mobility are imperative. Traditional solutions require ways to adapt the individual to prosthetic devices. This idea is also applied to patients who have congenital limitations. However, one of the major difficulties faced by those who are fitted with these devices is the great mental effort needed during first stages of training. As a result, a meaningful number of patients give up the use of theses devices very soon. Thus, this article reports on a solution designed by the authors to help patients during the learning phases, without actually having to wear the prosthesis. This solution considers Virtual (VR) and Augmented Reality (AR) techniques to mimic the prosthesis natural counterparts. Thus, it is expected that problems such as weight, heat and pain should not contribute to an already hard task.

Neuromimetic Event-Based Detection for Closed-Loop Tactile Feedback Control of Upper Limb Prostheses

Upper limb amputees lack the valuable tactile sensing that helps provide context about the surrounding environment. Here, we utilize tactile information to provide active touch feedback to a prosthetic hand. First, we developed fingertip tactile sensors for producing biomimetic spiking responses for monitoring contact, release, and slip of an object grasped by a prosthetic hand. We convert the sensor output into pulses, mimicking the rapid and slowly adapting spiking responses of receptor afferents found in the human body. Second, we designed and implemented two neuromimetic event-based algorithms, Compliant Grasping and Slip Prevention, on a prosthesis to create a local closed-loop tactile feedback control system (i.e., tactile information is sent to the prosthesis). Grasping experiments were designed to assess the benefit of this biologically inspired neuromimetic tactile feedback to a prosthesis. Results from able-bodied and amputee subjects show the average number of objects that broke or slipped during grasping decreased by over 50 percent and the average time to complete a grasping task decreased by at least 10 percent for most trials when comparing neuromimetic tactile feedback with no feedback on a prosthesis. Our neuromimetic method of closed-loop tactile sensing is a novel approach to improving the function of upper limb prostheses.

Functional Languages in Signal Processing Applied to Prosthetic Limb Control

This article describes how one can use functional languages to develop a dedicated system for controlling a prosthetic arm. It shows the prototype artificial limb along with the development of the various algorithms and software used to process electromyographic (EMG) signals, to be used as inputs for the control mechanism. Great emphasis is also laid on the parametric modelling used to extract the necessary features from the EMG signals. An iterative least mean square (LMS) algorithm has been used to improve the efficiency of traditional LMS algorithms, which greatly enhanced the performance of the system. The use of the functional paradigm lead to a fast developing stage and a very compact set of programs that will run as fast as (sometimes even faster than) traditional C/C++ programs.

Using Augmented Reality Techniques to Simulate Myoelectric Upper Limb Prostheses

This article proposes the use of Augmented Reality (AR) techniques for control and simulation of myoelectric prostheses. The system has been designed so that it is able to reproduce the operation of a real prosthesis in an immersive AR environment, using a virtual device that operates in similar fashion to the real one, resulting in a training environment for users and therapists. Motion and posture of the virtual prosthesis is controlled by EMG signals collected via surface electrodes and classified into four classes of movements. The results of tests with non-amputee volunteers show that the system is capable of generating the correct prosthesis motion and posture in the AR environment, in real time.

EMG Decomposition and Artefact Removal

Traditionally, in clinical electromyography (EMG), neurophysiologists assess the state of the muscle by studying basic units of an EMG signal, which are referred to as motor unit action potentials (MUAPs). Information regarding themorphology and rate of occurrence ofMUAPs is often used for diagnosis of neuromuscular disorders. In addition, recent studies have shown that the analysis of the energy content of MUAPs is a possible way for discriminating among normal, neurogenic, and myopathic MUAPs [38], illustrating, thus, the clinical value of the interpretation of MUAP information.

Mobile Augmented Reality enhances indoor navigation for wheelchair users

Introduction: Individuals with mobility impairments associated with lower limb disabilities often face enormous challenges to participate in routine activities and to move around various environments. For many, the use of wheelchairs is paramount to provide mobility and social inclusion. Nevertheless, they still face a number of challenges to properly function in our society. Among the many difficulties, one in particular stands out: navigating in complex internal environments (indoors). The main objective of this work is to propose an architecture based on Mobile Augmented Reality to support the development of indoor navigation systems dedicated to wheelchair users, that is also capable of recording CAD drawings of the buildings and dealing with accessibility issues for that population. Methods Overall, five main functional requirements are proposed: the ability to allow for indoor navigation by means of Mobile Augmented Reality techniques; the capacity to register and configure building CAD drawings and the position of fiducial markers, points of interest and obstacles to be avoided by the wheelchair user; the capacity to find the best route for wheelchair indoor navigation, taking stairs and other obstacles into account; allow for the visualization of virtual directional arrows in the smartphone displays; and incorporate touch or voice commands to interact with the application. The architecture is proposed as a combination of four layers: User interface; Control; Service; and Infrastructure. A proof-of-concept application was developed and tests were performed with disable volunteers operating manual and electric wheelchairs. Results The application was implemented in Java for the Android operational system. A local database was used to store the test building CAD drawings and the position of fiducial markers and points of interest. The Android Augmented Reality library was used to implement Augmented Reality and the Blender open source library handled the basis for implementing directional navigation arrows. OpenGL ES provided support for various graphics and mathematical transformations for embedded systems, such as smartphones. Experiments were performed in an academic building with various labs, classrooms and male and female bathrooms. Two disable volunteers using wheelchairs showed no difficulties to interact with the application, either by entering touch or voice commands, and to navigate within the testing environment with the help of the navigational arrows implemented by the augmented reality modules. Conclusion The novel features implemented in the proposed architecture, with special emphasis on the use of Mobile Augmented Reality and the ability to identify the best routes free of potential hazards for wheelchair users, were capable of providing significant benefits for wheelchair indoor navigation when compared to current techniques described in the literature.

Towards better understanding and reducing the effect of limb position on myoelectric upper-limb prostheses

Myoelectric control of prosthetic devices tend to rely on classification schemes of extracted features of EMG data. Those features however, may be sensitive to arm position resulting in decreased performance in real-world applications. The effect of varying limb position in a pattern recognition system have been illustrated by documenting the change in classification accuracy as the user achieves particular limb configurations. We continue to investigate this limb position effect by observing its impact on classification accuracy as well as through an analysis of how each extracted feature of the raw EMG varies in each position. Finally, LDA classification schemes are applied both to demonstrate the effect varying limb position has on classification accuracy and to increase classification accuracy without the use of additional hardware or sensors such as accelerometers as has been done in the past. It is shown that high classification accuracy can be achieved by (1) training an LDA classifier with data from many positions, as well as (2) by utilizing an extra position LDA classifier which can weigh the grasp classifiers appropriately. The classification accuracies achieved by these methods approached that of a model relying on a perfect knowledge of arm position.