Search

This paper presents a cognitive model on electroconvulsive treatment to reduce the stress level in body. The stress reduction is triggered by a cognitive electroconvulsive treatment that uses persistent manipulation of this treatment. The goal of this treatment is to decrease the strength between certain parts of the brain which are in charge of the stress. The proposed adaptive cognitive model aims to illustrate the effect of the therapy on different components of the brain. The model begins with a state of tough, powerful, and consistent stress within a post-traumatic disorder patient, and after following electroconvulsive treatment, the stress level starts to decrease from time to time according to each treatment session. The results show that, in the end, the disorder person will have a declined stress in contrast to not performing electroconvulsive treatment.

Traffic signs are one of the most important information that guide cars to travel, and the detection of traffic signs is an important component of autonomous driving and intelligent transportation systems. Constructing a traffic sign dataset with many samples and sufficient attribute categories will promote the development of traffic sign detection research. In this paper, we propose a new Chinese traffic sign detection benchmark, which adds more than 4,000 real traffic scene images and corresponding detailed annotations based on our CCTSDB 2017, and replaces many original easily-detected images with difficult samples to adapt to the complex and changing detection environment. Due to the increase of the number of difficult samples, the new benchmark can improve the robustness of the detection network to some extent compared to the old version. At the same time, we create new dedicated test sets and categorize them according to three aspects: category meanings, sign sizes, and weather conditions. Finally, we present a comprehensive evaluation of nine classic traffic sign detection algorithms on the new benchmark. Our proposed benchmark can help determine the future research direction of the algorithm and develop a more precise traffic sign detection algorithm with higher robustness and real-time performance.

Wearable on body sensors have been employed in many applications including ambulatory monitoring and pervasive computing systems. In this work, a wearable assistant has been created for people suffering from Parkinson’s disease (PD), specifically with the Freezing of Gait (FoG) symptom. Wearable accelerometers were placed on the person’s body and used for movement measure. When FoG is detected, a rhythmic audio signal was given from the wearable assistant to motivate the wearer to continue walking. Long term monitoring results in collecting huge amounts of complex raw data; therefore, data analysis becomes impractical or infeasible resulting in the need for data reduction. In the present study, Discrete Wavelet Transform (DWT) has been used to extract the main features inherent in the key movement indicators for FoG detection. The discrimination capacities of these features were assessed using, i) Support Vector Machine (SVM) using a linear kernel function, and ii) Artificial Neural Network (ANN) with a two-layer feed-forward with hidden layer of 20 neurons that trained with conjugate gradient back- propagation. Using these two different machine learning techniques, we were capable of detecting FoG with an accuracy of 87.50% and 93.8%, respectively. Additionally, the comparison between the extracted features from DWT coefficients with those using Fast Fourier Transform (FFT) established accuracies of 93.8% and 81.3%, respectively. Finally, the discriminative features extracted from DWT yield to a robust multidimensional classification model compared to models in the literature based on a single feature. The work presented paves the way for reliable, real-time wearable sensors to aid people with PD.

In Wireless Sensor Networks (WSNs), clustering has been shown to be an efficient technique to improve scalability and network lifetime. In clustered networks, clustering creates unequal load distribution among Cluster Heads (CHs) and Cluster Member (CM) nodes. As a result, the entire network is subject to premature death because of the deficient active nodes within the network. In this paper, we present clustering-based routing algorithms that can balance out the trade-off between load distribution and network lifetime “green cluster-based routing scheme”. This paper proposes a new energy aware green cluster-based routing algorithm to preventing premature death of large scale dense WSNs. To deal with the uncertainty present in network information, a fuzzy rule-based node classification model is proposed for clustering. Its primary benefits are flexibility in selecting effective CHs, reliability in distributing CHs overload among the other nodes, and reducing communication overhead and cluster formation time in highly dense areas. In addition, we propose a routing scheme that balances the load among sensors. The proposed scheme is evaluated through simulations to compare our scheme with the existing algorithms available in the literature. The numerical results show the relevance and the improved efficiency of our scheme.

Wireless sensor networks (WSNs) have drawn much research attention in recent years due to the superior performance in multiple applications, such as military and industrial monitoring, smart home, disaster restoration etc. In such applications, massive sensor nodes are randomly deployed and they remain static after the deployment, to fully cover the target sensing area. This will usually cause coverage redundancy or coverage hole problem. In order to effectively deploy sensors to cover whole area, we present a novel node deployment algorithm based on mobile sensors. First, sensor nodes are randomly deployed in target area, and they remain static or switch to the sleep mode after deployment. Second, we partition the network into grids and calculate the coverage rate of each grid. We select grids with lower coverage rate as candidate grids. Finally, we awake mobile sensors from sleep mode to fix coverage hole, particle swarm optimization (PSO) algorithm is used to calculate moving position of mobile sensors. Simulation results show that our algorithm can effectively improve the coverage rate of WSNs.

AAL has benefitted tremendously from the near-ubiquity of powerful smartphones and very high data rates available over broadband and mobile networks. However, this is beyond the reach of many users. IoT systems offer the potential to extend some of the benefits to disadvantaged users. Such solutions will need to secure personal health information and provide a sufficient quality of service even when operating constrained user devices and communications.

The SPHERE project has developed a multi-modal sensor platform for health and behavior monitoring in residential environments. So far, the SPHERE platform has been deployed for data collection in approximately 50 homes for duration up to one year. This technical document describes the format and the expected content of the SPHERE dataset(s) under preparation. It includes a list of some data quality problems (both known to exist in the dataset(s) and potential ones), their workarounds, and other information important to people working with the SPHERE data, software, and hardware. This document does not aim to be an exhaustive descriptor of the SPHERE dataset(s); it also does not aim to discuss or validate the potential scientific uses of the SPHERE data.

Recently, many researchers have concentrated on using neural networks to learn features for Distant Supervised Relation Extraction (DSRE). These approaches generally use a softmax classifier with cross-entropy loss, which inevitably brings the noise of artificial class NA into classification process. To address the shortcoming, the classifier with ranking loss is employed to DSRE. Uniformly randomly selecting a relation or heuristically selecting the highest score among all incorrect relations are two common methods for generating a negative class in the ranking loss function. However, the majority of the generated negative class can be easily discriminated from positive class and will contribute little towards the training. Inspired by Generative Adversarial Networks (GANs), we use a neural network as the negative class generator to assist the training of our desired model, which acts as the discriminator in GANs. Through the alternating optimization of generator and discriminator, the generator is learning to produce more and more discriminable negative classes and the discriminator has to become better as well. This framework is independent of the concrete form of generator and discriminator. In this paper, we use a two layers fully-connected neural network as the generator and the Piecewise Convolutional Neural Networks (PCNNs) as the discriminator. Experiment results show that our proposed GAN-based method is effective and performs better than state-of-the-art methods.

The recent boom in the Internet of Things (IoT) will turn Smart Cities and Smart Homes (SH) from hype to reality. SH is the major building block for Smart Cities and have long been a dream for decades, hobbyists in the late 1970s made Home Automation (HA) possible when personal computers started invading home spaces. While SH can share most of the IoT technologies, there are unique characteristics that make SH special. From the result of a recent research survey on SH and IoT technologies, this paper defines the major requirements for building SH. Seven unique requirement recommendations are defined and classified according to the specific quality of the SH building blocks.

Wearable sensors and healthcare devices use small lightweight batteries to power their typical operations of monitoring and tracking. It becomes absolutely vital to effectively utilise all the available battery charge for device longevity between charges. The electrochemical recovery effect enables the extraction of more power from the battery when implementing idle times in between use cycles, and has been used to develop various power management techniques. However, there is no evidence concerning the actual increase in available power that can be attained using the recovery effect. Also, this property cannot be generalised on all the battery chemistries since it is an innate phenomenon, relying on the anode/cathode material. Indeed recent developments suggest that recovery effect does not exist at all. This paper presents experimental results to verify the presence and level of the recovery effect in commonly used battery chemistries in wearable sensors and healthcare devices. The results have revealed that the recovery effect significantly does exist in certain batteries, and importantly we show that it is also comprised of two different time constants. This novel finding has important implications for the development of power management techniques that utilise the recovery effect with application in a large range of battery devices.

Technological advances in sensors and communications have enabled discrete integration into everyday objects, both in the home and about the person. Information gathered by monitoring physiological, behavioural, and social aspects of our lives, can be used to achieve a positive impact on quality of life, health, and well-being. Wearable sensors are at the cusp of becoming truly pervasive, and could be woven into the clothes and accessories that we wear such that they become ubiquitous and transparent. To interpret the complex multidimensional information provided by these sensors, data fusion techniques are employed to provide a meaningful representation of the sensor outputs. This paper is intended to provide a short overview of data fusion techniques and algorithms that can be used to interpret wearable sensor data in the context of health monitoring applications. The application of these techniques are then described in the context of healthcare including activity and ambulatory monitoring, gait analysis, fall detection, and biometric monitoring. A snap-shot of current commercially available sensors is also provided, focusing on their sensing capability, and a commentary on the gaps that need to be bridged to bring research to market.

Technology has been seen as a possible solution to the increasing costs of healthcare and the globally aging population. It is known that many elderly people prefer to stay in their homes for as long as possible and remote monitoring can be a solution, but often such systems lack useful information or are prohibitive due to cost, ease of use/deployment and wireless coverage.\ud \ud This work presents a novel gateway software architecture based on threads being managed by dictionary memory. The architecture has been deployed in a distributed interconnected set of low-cost consumer grade gateway devices using Bluetooth Low Energy (BLE) that are positioned around the home. The gateway devices can then be used to listen, monitor or connect to BLE based healthcare sensors to continually reveal information about the user with full residential coverage. A further novelty of this work is the ability to maintain handover connections between many sensors and many gateways as a user moves throughout their home, thus the gateways can route information to/from sensors across the consumer’s home network. The system has been tested in an experimental house and is now poised to be initially deployed to 100 homes for residential healthcare monitoring before any public mass consumer deployment.

Supercapacitors are likely to be adopted as power sources for wearable sensors; in particular where the sensor mechanism relies on energy harvesting. A specific advantage of supercapacitors over traditional batteries is their performance over large numbers of discharge cycles. Likewise, in the case of wearable devices, it is essential to efficiently manage the available power. Supercapacitors exhibit a small recovery effect, in part due to ion diffusion. Modelling this effect allows an increase in available energy to be realized following the sleep times during a discharge cycle thus increasing the time between charging for wearables. This paper presents the increase in useful lifetime that can be achieved via the recovery effect in a typical wearable device.

Health monitoring technologies such as Body Area Network (BAN) systems has gathered a lot of attention during the past few years. Largely encouraged by the rapid increase in the cost of healthcare services and driven by the latest technological advances in Micro-Electro-Mechanical Systems (MEMS) and wireless communications. BAN technology comprises of a network of body worn or implanted sensors that continuously capture and measure the vital parameters such as heart rate, blood pressure, glucose levels and movement. The collected data must be transferred to a local base station in order to be further processed. Thus, wireless connectivity plays a vital role in such systems. However, wireless connectivity comes at a cost of increased power usage, mainly due to the high energy consumption during data transmission. Unfortunately, battery-operated devices are unable to operate for ultra-long duration of time and are expected to be recharged or replaced once they run out of energy. This is not a simple task especially in the case of implanted devices such as pacemakers. Therefore, prolonging the network lifetime in BAN systems is one of the greatest challenges. In order to achieve this goal, BAN systems take advantage of low-power in-body and on-body/off-body wireless communication technologies. This paper compares some of the existing and emerging low-power communication protocols that can potentially be employed to support the rapid development and deployment of BAN systems.

The creation of OFDM based Wireless Personal\ud Area Networks (WPANs) has allowed high bit-rate wireless\ud communication devices suitable for streaming High Definition\ud video between consumer products as demonstrated in Wireless-\ud USB. However, these devices need high clock rates, particularly\ud for the OFDM sections resulting in high silicon cost and high\ud electrical power. Acknowledging that electrical power in\ud wireless consumer devices is more critical than the number of\ud implemented logic gates, this paper presents a Double Data Rate\ud (DDR) architecture to reduce the OFDM input and output clock\ud rate by a factor of 2. The architecture has been implemented\ud and tested for Wireless-USB (ECMA-368) resulting in a\ud maximum clock of 264MHz instead of 528MHz existing\ud anywhere on the die.

Dual Carrier Modulation (DCM) is currently used\ud as the higher data rate modulation scheme for Multiband\ud Orthogonal Frequency Division Multiplexing (MB-OFDM) in the\ud ECMA-368 defined Ultra-Wideband (UWB) radio platform.\ud ECMA-368 has been chosen as the physical radio platform for\ud many systems including Wireless USB (W-USB), Bluetooth 3.0\ud and Wireless HDMI; hence ECMA-368 is an important issue to\ud consumer electronics and the user’s experience of these products.\ud In this paper, Log Likelihood Ratio (LLR) demapping method\ud is used for the DCM demaper implemented in fixed point model.\ud Channel State Information (CSI) aided scheme coupled with the\ud band hopping information is used as the further technique to\ud improve the DCM demapping performance. The receiver\ud performance for the fixed point DCM is simulated in realistic\ud multi-path environments.

As consumers demand more functionality) from their electronic devices and manufacturers supply the demand then electrical power and clock requirements tend to increase, however reassessing system architecture can fortunately lead to suitable counter reductions. To maintain low clock rates and therefore reduce electrical power, this paper presents a parallel convolutional coder for the transmit side in many wireless consumer devices. The coder accepts a parallel data input and directly computes punctured convolutional codes without the need for a separate puncturing operation while the coded bits are available at the output of the coder in a parallel fashion. Also as the computation is in parallel then the coder can be clocked at 7 times slower than the conventional shift-register based convolutional coder (using DVB 7/8 rate). The presented coder is directly relevant to the design of modern low-power consumer devices

The Universal Serial Bus (USB) is an extremely popular interface standard for computer peripheral connections and is widely used in consumer Mass Storage Devices (MSDs). While current consumer USB MSDs provide relatively high transmission speed and are convenient to carry, the use of USB MSDs has been prohibited in many commercial and everyday environments primarily due to security concerns. Security protocols have been previously proposed and a recent approach for the USB MSDs is to utilize multi-factor authentication. This paper proposes significant enhancements to the three-factor control protocol that now makes it secure under many types of attacks including the password guessing attack, the denial-of-service attack, and the replay attack. The proposed solution is presented with a rigorous security analysis and practical computational cost analysis to demonstrate the usefulness of this new security protocol for consumer USB MSDs.

Nowadays, broadcasters deliver ultra-high resolution video to their consumers. This live video is sent to a set-top box for display on a television. However, if one or more users in the home want to view the same video on their personal mobile devices with a lower display resolution and limited processing power, decoding the original ultra-high resolution video would result in stuttering and quickly drain the battery life on these devices. To enable a satisfactory consumer experience, the resolution of the video stream should be adapted to the target mobile device at the set-top box. The aim of this paper is to investigate the performance of different machine learning strategies to arbitrary downsize video pre-encoded with the high efficiency video coding standard (HEVC). These machine learning techniques exploit correlation between input and output coding information to predict the splitting behavior of HEVC coding units. Several machine learning algorithms are optimized. Additionally, both online and offline training strategies are tested. Of the tested algorithms, online-trained random forests achieve the best compression-efficiency with a bit rate increase of 5.4% and an average complexity reduction of 70%1.