Human Activity Recognition (HAR) plays a crucial role in numerous applications, ranging from healthcare to sports analytics. This study presents a novel approach to HAR that combines Artificial Neural Networks (ANNs) and Random Forests to enhance the accuracy of HAR in diverse real-world conditions, especially when dealing with noisy and imperfect data. ANNs are known for extracting intricate features from raw data, making them well-suited for classification. Random Forest excels at learning from multiple decision trees and utilizing collective knowledge to make predictions, making it suitable for handling data in real-world applications. Harnessing the power of ANNs in feature extraction, coupled with the collective decision-making capability of Random Forest, the combined model demonstrates improved accuracy in classifying human activities. This study showcases the potential of combining ANN and Random Forest in classifying multi-dimensional Inertial Measurement Unit (IMU) data, a widely-used data source in HAR. By leveraging the strengths of both ANN and Random Forest, the combined model addresses the challenges associated with real, imperfect data, leading to a more robust and accurate classification model. The results highlight the effectiveness of feature extraction by ANNs and underscore the importance of incorporating Random Forest in HAR systems to obtain 98.84% accuracy. The findings of this study offer valuable insights into the synergistic effects of combining ANN and Random Forest for HAR. The outcomes can contribute to developing more reliable and effective HAR systems, with potential applications in healthcare monitoring, activity recognition in smart environments, and other domains requiring accurate human activity classification.
This study proposed a control method, a data-driven predictive control (DDPC), for the hand exoskeleton used for active, passive, and resistive rehabilitation. DDPC is a model-free approach based on past system data. One of the strengths of DDPC is that constraints of states can be added to the controller while performing the controller design. These features of the control algorithm eliminate an essential problem for rehabilitation robots in terms of easy customization and safe repetitive rehabilitation tasks that can be planned within certain constraints. Experiments were carried out with a designed hand rehabilitation system under repetitive and various therapy tasks. Real-time experiment results demonstrate the feasibility and efficiency of the proposed control approach to rehabilitation systems.
The Internet of Things (IoT) is a collection of inexpensive, semi-autonomous, Internet-connected devices that sense and interact within the physical world. IoT security is of paramount concern, because most IoT devices use weak or no encryption at all. This concern is exacerbated by the fact that the number of IoT deployments continues grow, IoT devices are being integrated into key infrastructures, and their weak or lack of encryption is being exploited. Specific Emitter Identification (SEI) is being investigated as an effective, cost-saving IoT security approach, because it is a passive technique that exploits inherent, distinct features that are unintentionally imparted to the waveform during its formation and transmission by the IoT device’s Radio Frequency (RF) front-end. Despite the amount of research conducted, SEI still faces roadblocks that hinder its integration within operational networks. Our work focuses on the lack of feature permanence across time and environments, which is designated herein as the “multi-day” problem. We present results and analysis for six distinct experiments focused on improving multi-day SEI performance through the use of multiple waveform representations, deeper Convolutional Neural Networks (CNNs), increasing numbers of waveforms, channel model impacts, and two channel mitigation techniques. Our work shows improved multi-day SEI performance using the waveform’s frequency-domain representation and a CNN comprised of four convolutional layers. However, the traditional channel model and both channel mitigation techniques fail to sufficiently mitigate or remove real-world channel impacts, which suggests that the channel may not be the dominant effect hindering multi-day SEI performance.
22 November 2022 in Journal of Testing and Evaluation
Soft pneumatic grippers (SPGs) contribute significantly to the development of soft robotic devices and help enable different and complex hardware designs for soft robots. Apart from the many advantages, the lack of position feedback of soft grippers presents a problem for position control. This study aims to examine this problem with the SPG, which is fabricated using additive manufacturing technology. The soft gripper is fabricated in two parts using the molding technique. The structure of the gripper consists of molded silicone (room-temperature-vulcanizing [room-temperature-vulcanizing]-3015) and polyethylene terephthalate (PET) film consisting of chambers. These sections are an active chamber consisting of 10 airtight sections that can be operated with compressed air and a passive chamber covered with PET film. During the bending of the SPG, the PET film inside the passive chamber begins to be compressed by the compressed air, resulting in a higher load-holding capacity of the gripper. In addition, the position of the feedback module made of flex sensor is integrated during fabrication. Various experiments were carried out to evaluate the performance of the SPG. Finite element analysis of the gripper was performed with COMSOL to compare the accuracy of real-time experiments. With the experimental studies’ results, the SPG’s motion characteristics, the relationship between the bending angle and the pressure, and the position-tracking performance were analyzed. Experimental results of the SPG were found to be in good agreement with the simulation results. The experimental results determined that the SPG showed successful trajectory tracking performance with the proportional-integral-derivative control algorithm. The root mean square value for the controller is calculated as 0.1245°.
In this study, the focus is on an implant used in the treatment of early-onset scoliosis called magnetically controlled growing rods (MCGR). The primary goal of the study is to address and propose solutions for the mechanical problems reported in the literature concerning MCGR. The problems of the MCGR are mainly due to excessive stress and mechanical bearing problems. Therefore, an MCGR removed from a patient is teardown and geometrically modeled. Then, eleven design parameters are determined on the MCGR for the mechanical problems experienced and these are evaluated by mechanical analysis over 14 control points. In this study, analysis processes are carried out with L12 orthogonal array for eleven design parameters and 2 levels using Taguchi’s experimental design method (DoE). With the obtained data by analyzing the experiments in L12, the fitness functions depending on the design parameters are created for 14 control points. Since the problem is multi-objective, a non-dominated sorting genetic algorithm (NSGA II) and multi-objective particle swarm optimization (MOPSO) are used to minimize stress and displacement in existing mechanical problems using fitness functions. The obtained design models from NSGA II and MOPSO are analyzed and evaluated in comparison with the existing mechanical model obtained through pre-optimization teardown study of MCGR.
This study has designed an easy-to-wear parallel continuum robot-based hand rehabilitation system that supports and enhances the finger’s flexion, extension, abduction, and adduction movements. The primary novelty of the proposed system lies in its ability to focus therapeutic exercises on a single joint, a feature not commonly found in existing rehabilitation robots. A kinematic model of the system was developed, and to perform both kinematic and dynamic analyses, a multibody model was constructed in the MATLAB Simulink environment. Joint angle control was implemented using a nominal controller, and to account for individual uncertainties in joint dynamics, a neuroadaptive controller was integrated with the nominal controller. This approach aims for the neural network architecture to learn these uncertainties during control iterations and incorporate them into the control, resulting in a robust controller. Thus, a model reference control approach was proposed for active and passive rehabilitation processes. The system model was tested in a simulation environment, and then all tests were repeated in the physical system. The simulation and real system results include the real system’s open-loop responses, nominal controller responses for each joint, responses, and the results for active, passive, and assistive control modes.