We have developed a virtual olfactory environment using an olfactory display and computational fluid dynamics (CFD) simulation to provide odor concentration change according to user's movement in real-time in a vi...
详细信息
ISBN:
(纸本)9798350374025;9798350374032
We have developed a virtual olfactory environment using an olfactory display and computational fluid dynamics (CFD) simulation to provide odor concentration change according to user's movement in real-time in a virtual space. Although CFD can calculate the odor distribution in the complicated geometry, its computational cost was expensive and did not work in real-time in the previous study. In this study, real-time CFD based on GPU calculation was introduced to generate olfactory VR environment. Using real-time CFD we investigated influence of the user's body with its location and orientation changing irregularly. In the sensory test to find the odor direction, the correct answer rate was over 70% when the body influence was considered, while it was just a chance hit when the influence was not considered. The experimental result indicates the usefulness of considering the effect of the user's body since we cannot avoid that influence.
We present an online algorithm for detecting changes in computer network activity. Anomalous activity in IT systems often appear as changes in network topology as edges evolve in a dynamic graph;identifying these beha...
详细信息
ISBN:
(纸本)9798350364941;9798350364958
We present an online algorithm for detecting changes in computer network activity. Anomalous activity in IT systems often appear as changes in network topology as edges evolve in a dynamic graph;identifying these behavioral changes can be a challenging task. We propose a method that uses principles of Hyperdimensional computing to encode graphs to a real valued space where anomalies are easily identifiable. With reasonable assumptions of the baseline edge generating process, our approach operates in realtime and can produce an anomaly score primitive for one sample independently of all others. This score lends itself easily to an online Bayesian confidence estimate in constant memory, which is essential for real-world applications where networks are extremely large and interpretable predictions are needed in realtime. We demonstrate the effectiveness of our approach on both synthetic and real-world datasets.
The proceedings contain 27 papers. The topics discussed include: high performance conductive composite hydrogel interface for epidermal electrophysiological monitoring;implementation of Kirigami design in flexible wea...
ISBN:
(纸本)9798350375213
The proceedings contain 27 papers. The topics discussed include: high performance conductive composite hydrogel interface for epidermal electrophysiological monitoring;implementation of Kirigami design in flexible wearable electronic by laser cutting for conformability;cuffless continuous BP monitoring from photoplethysmography signals using UNET and attention mechanism;composite graphene/Cu pressure sensor with one-step laser-fabricated microstructure: a one stone two birds strategy;an active interface pressure adapter for enhancing sensing capacity and stability of wearable electronics;3D porous stacked organic electrochemical transistor for wearable glucose sensing: enhancing comfort, breathability, and real-time monitoring;tracking of dual gas concentration using carbon nanotubes based reservoir computing;and a novel fluid classification unit based on moisture electricity generation mechanism.
AI-Powered Edge computing is accelerating the integration of the cyber world with the ever-growing list of new physical IoT devices and will fundamentally change and empower the way humans interact with the world. In ...
详细信息
ISBN:
(纸本)9798350376975;9798350376968
AI-Powered Edge computing is accelerating the integration of the cyber world with the ever-growing list of new physical IoT devices and will fundamentally change and empower the way humans interact with the world. In this paper, we prototyped and analyzed three edge computing architectures for running SmartFall, a real-time fall detection application that uses accelerometer data from the watch, to compare the trade-off in relationship to battery consumption, potential data loss, machine learning model's prediction accuracy, and latency in model inferencing. Our experiments show that running the machine learning prediction on the server using the TensorFlow native model format has achieved the best model accuracy without draining the battery power of the smartwatches. However, the optimal selection of the software architecture depends on the intended deployment environment, projected user numbers, users' privacy concerns, and network stability.
The emerging of Digital Twin (DT) technology facilitates the further development of industrial automation. However, real-time and accurate DTs modeling and updating require massive communication and computing resource...
详细信息
ISBN:
(纸本)9798350358513;9798350358520
The emerging of Digital Twin (DT) technology facilitates the further development of industrial automation. However, real-time and accurate DTs modeling and updating require massive communication and computing resources, which poses a challenge to limited resources. Edge computing as a distributed computing architecture offers the possibility of high-efficient resource scheduling in DTs. Motivated by this gap, this paper aim to solve the problem of real-time and high fidelity DTs modeling and updating. First, we represent the computing tasks of DTs in the form of Heterogeneous computing Task Graph (HCTG). Then, a Hierarchical Attention Mechanism (HAT) is proposed to obtain the latent representation vectors of the HCTG. Finally, we design Markov Decision Process (MDP), and propose Deep Reinforcement Learning (DRL)-based computing task scheduling approach (HAT-DRL) to satisfy the minimum total completion time requirement of different DTs. Experimental results demonstrate that the proposed algorithm has promising scheduling performance and outperforms other task scheduling algorithms.
Driver fatigue is a leading cause of road accidents. This research presents a real-time driver drowsiness detection system using deep learning, optimized for low-power embedded systems. By compressing a complex model ...
详细信息
The Internet of Vehicles (IoV) represents a critical component of modern mobile edge computing systems. IoV computing resources encompass backend cloud resources, roadside edge nodes, and vehicle-mounted units. Tradit...
详细信息
With the advancement of artificial intelligence (AI) technologies, novel and inventive approaches for addressing complex problems are coming to the forefront. Neuromorphic computing based on AI technologies stands as ...
详细信息
With the development of industrial Internet, numerous real-time applications emerge where a huge amount of sensing data should be captured from the environment and processed. In this paper, we consider a heterogeneous...
详细信息
ISBN:
(纸本)9798350378412
With the development of industrial Internet, numerous real-time applications emerge where a huge amount of sensing data should be captured from the environment and processed. In this paper, we consider a heterogeneous mobile edge computing (Het-MEC) network via cloud-edge collaboration, which integrates data sensing, computing, and communication. Considering the real-time demand of industrial production, the freshness and validity of data need to be guaranteed. Therefore, we formulate the average peak age of information (PAoI) minimization problem. A Proximal Policy Optimization (PPO) based deep reinforcement learning algorithm is proposed to solve it online by dynamically optimizing the task offloading strategy and the joint sensing-computing-communication resource allocation scheme. Simulation results show that compared with local computing and cloud computing schemes, the proposed algorithm can achieve a lower average PAoI of the system. Benefited from the learning and transfer characteristic of deep model, the time complexity can also be reduced in a long-term view compared to the heuristic algorithm.
暂无评论