DRL has emerged as a promising approach for mobile robot navigation in unknown environments without a prior map. However, the performance of DRL methods for this task varies greatly, depending on the choice of algorit...
DRL has emerged as a promising approach for mobile robot navigation in unknown environments without a prior map. However, the performance of DRL methods for this task varies greatly, depending on the choice of algorithm, state representation, and training procedure. In this paper we explore various cutting-edge DRL algorithms, such as policy-, value-, and actor-critic-based approaches. Our results demonstrate the effectiveness of the ranging sensor approach, which achieves robust navigation policies capable of generalizing to unseen virtual environments with a high success rate. We combine Behavior Cloning with Imitation Learning to expedite the training process, leveraging expert demonstrations and reinforcement learning. Our methodology enables faster training while enhancing the learning efficiency and performance of the robot, obtaining better results in terms of crash and success rate, and being able to reach a cumulative reward of approximately 12000.
The recent proliferation of hyper-realistic deepfake videos has drawn attention to the threat of audio and visual forgeries. Most previous studies on detecting artificial intelligence-generated fake videos only utiliz...
详细信息
Path planning is a crucial component of autonomous navigation and frequently demands different priorities such as path length, safety, or energy consumption, with the latter being particularly important in the context...
Path planning is a crucial component of autonomous navigation and frequently demands different priorities such as path length, safety, or energy consumption, with the latter being particularly important in the context of unmanned autonomous vehicles. In many applications, the agent may have to react to environment shifting. Algorithms such as geometric and dynamic programming as well as techniques such as artificial potential fields have been employed to tackle this local planning problem. In recent years, machine learning has gained more evidence in many research fields due to its flexibility and generalization capabilities. In this study, we propose a Q-learning-based approach to local planning, which weighs three crucial factors- path length, safety, and energy consumption- that can be freely adjusted by the user to suit its application’s needs. The performance of the proposed method was tested in simulated static and dynamic scenarios as well as benchmarked with a baseline approach. The results show that it can perform well in both kinds of environments without struggling with the commom pitfalls of other local planning algorithms.
Neurodegenerative disease is a growing global problem. Many of these diseases such as Parkinson's disease can cause grip strength weakness. In this work, we focused on developing an e-textile based EMG acquisition...
详细信息
ISBN:
(数字)9798350309652
ISBN:
(纸本)9798350309669
Neurodegenerative disease is a growing global problem. Many of these diseases such as Parkinson's disease can cause grip strength weakness. In this work, we focused on developing an e-textile based EMG acquisition system interfaced with an adaptive VR game engine for detecting hand squeeze actions. We prototyped and performed preliminary tests on a grip detection system consisting of: 1) an e-textile forearm band (E-band), 2) a portable data acquisition module (DAM), and 3) an adaptive 3D VR game engine. Digital health solutions such as our VRGrip system widen the scope of in-home rehabilitation and assessment, and we aimed to make the first step in creating a system for grip strength rehabilitation. To validate our concept, we performed preliminary tests. These tests revealed that the Bluetooth communication protocol between DAM and the game engine was reliable, that the adaptive game engine successfully tailors the game's difficulty level to the user's performance, and that the system was reliable at detecting hand squeezes (F=0.96).
An increasing number of people are being diagnosed with neurodegenerative diseases every year. One of the aims of digital health is to provide innovative solutions for rehabilitation or pre-diagnosis of these diseases...
An increasing number of people are being diagnosed with neurodegenerative diseases every year. One of the aims of digital health is to provide innovative solutions for rehabilitation or pre-diagnosis of these diseases. In this vein, e-textiles can allow for the construction of flexible and comfortable devices to record biosignals and develop novel rehabilitation platforms. In this study, an e-textile forearm band was constructed to acquire an electromyogram (EMG: a muscle response to a nerve’s stimulation of the muscle) from the flexor muscles of the hand to detect hand grips/squeezes. A light and portable data acquisition module, as well as a computer-based virtual reality game (controlled by the acquired EMG signal) were designed. Our pilot study showed that the system is reliable in detecting hand squeezes in both 6 healthy participants (F-score: 0.88) and 6 participants with Parkinson’s disease (F-score: 0.88). Participants also rated the system to be enjoyable, comfortable, and easy to use.
Contactless sensors embedded in the ambient environment have broad applications in unobtrusive, long-term health monitoring for preventative and personalized healthcare. Microwave radar sensors are an attractive candi...
Contactless sensors embedded in the ambient environment have broad applications in unobtrusive, long-term health monitoring for preventative and personalized healthcare. Microwave radar sensors are an attractive candidate for ambient sensing due to their high sensitivity to physiological motions, ability to penetrate through obstacles and privacy-preserving properties, but practical applications in complex real-world environments have been limited because of challenges associated with background clutter and interference. In this work, we propose a thin and soft textile sensor based on microwave metamaterials that can be easily integrated into ordinary furniture for contactless ambient monitoring of multiple cardiovascular signals in a localized manner. Evaluations of our sensor’s performance in human subjects show high accuracy of heartbeat and arterial pulse detection, with ≥ 96.5% sensitivity and < 5% mean absolute relative error (MARE) across all subjects. We demonstrate our sensor’s utility for cuffless blood pressure monitoring on a human subject over a continuous 10-minute period. Our results highlight the potential of metamaterial textile sensors in ambient health and wellness monitoring *** relevance—The contactless metamaterial textile sensors demonstrated in this paper provide unobtrusive, convenient and long-term monitoring of multiple cardiovascular health metrics, including heart rate, pulse rate and cuffless blood pressure, which can facilitate preventative and personalized healthcare.
Wound monitoring is crucial for effective healing, as nonhealing wounds can lead to tissue ulceration and necrosis. Evaluating wound recovery involves observing changes in angiogenesis. Laser speckle contrast imaging ...
详细信息
The demand for electricity has increased rapidly and, for this reason, there is a need to efficiently use it. In this way, the identification of residential appliances enables such use for consumers and is crucial for...
The demand for electricity has increased rapidly and, for this reason, there is a need to efficiently use it. In this way, the identification of residential appliances enables such use for consumers and is crucial for demand response programs. Due to the variety of appliances in homes and their dynamic behavior, the search for patterns that explain and allow the correct labeling of temporal windows becomes a challenging task, since a window may contain more than one appliance. In this sense, the present paper proposes the transformation of time-series into images, using Gramian angular field and recurrence plots. The dataset composed of images was submitted to the labeling process, considering the use of convolutional neural networks. A comparative analysis was performed using the UK-DALE dataset. The results demonstrated the effectiveness of the proposed feature engineering stage, since the labeling task reached F1-scores until 94 %.
This paper introduces a novel forecastings technique based on randomized fuzzy cognitive maps (FCM), called LRHFCM (or large reservoir of randomized high-order FCM) for predicting univariate time series. LR-HFCM is a ...
详细信息
ISBN:
(数字)9798350366235
ISBN:
(纸本)9798350366242
This paper introduces a novel forecastings technique based on randomized fuzzy cognitive maps (FCM), called LRHFCM (or large reservoir of randomized high-order FCM) for predicting univariate time series. LR-HFCM is a hybrid method combining fuzzy time series (FTS), FCMs, and reservoir computing. It is a type of echo state network (ESN) consisting of the input layer, intermediate (or large reservoir) layer, and output layer, where LASSO regression is applied to train the output layer. The novelty of this approach is that the internal layer includes a very large reservoir, considering different combinations from the sets of concepts and order using a certain number of sub-reservoirs to capture different dynamics of input time series. It is important to highlight that the weights within each sub-reservoir are chosen randomly and remain constant throughout the training process. The validity of the LR-HFCM approach is evaluated across 15 different time series datasets. The results highlight the outperformance of the LR-HFCM technique in comparison to various baseline models.
Left Ventricular Assist Devices have been successfully used for the treatment of Congestive Heart failure in patients who are not eligible for heart transplantation. This paper describes the implementation and compari...
Left Ventricular Assist Devices have been successfully used for the treatment of Congestive Heart failure in patients who are not eligible for heart transplantation. This paper describes the implementation and comparison of the performance of a pressure sensor-based feedback controller. The strategies were tested on a mock loop of the systemic circulation. The results show that the use of pressure sensors generated a more accurate response of the controller compared to the use of *** Relevance— The study describes the integration of a LVAD physiological controller to a dynamic wireless monitoring system.
暂无评论