Actual Deployment Research Articles

Optimal deployment for carbon capture enables more than half of China's coal-fired power plant to achieve low-carbon transformation.

Carbon capture, utilization and storage (CCUS) technology is critical to global net-zero emission goal, whereas actual deployment is well below expectations. This study constructs a comprehensive framework, integrating nonlinear dynamic optimization, real option and technology learning curve, to explore optimal CCUS deployment for China's coal-fired power plant toward carbon neutrality. The commercialization application will occur in 2030-2035, with the optimal potential ranging between 248.54 GW and 564.90 GW. East China has the greatest potential, reaching 196.85 GW, followed by North China with the potential of 116.29 GW. The cost of second-generation capture technology will decrease from 219 CNY/ton CO2 to 165 CNY/ton CO2 during 2030-2031.The annual corporate expenditure (R&D investment and capture cost) and government expenditure (subsidy) will peak at 23.92 billion CNY in 2035 and 63.71 billion CNY in 2044, respectively. The financial burden can be lessened by carbon trading market and third-party intervention in the later period.

Advances in stimuli-responsive systems for pesticides delivery: Recent efforts and future outlook.

Effective pest management for enhanced crop output is one of the primary goals of establishing sustainable agricultural practices in the world. Pesticides are critical in preventing biological disasters, ensuring crop productivity, and fostering sustainable agricultural production growth. Studies showed that crops are unable to properly utilize pesticides because of several limiting factors, such as leaching and bioconversion, thereby damaging ecosystems and human health. In recent years, stimuli-responsive systems for pesticides delivery (SRSP) by nanotechnology demonstrated excellent promise in enhancing the effectiveness and safety of pesticides. SRSP are being developed with the goal of delivering precise amounts of active substances in response to biological needs and environmental factors. An in-depth analysis of carrier materials, design fundamentals, and classification of SRSP were provided. The adhesion of SRSP to crop tissue, absorption, translocation in and within plants, mobility in the soil, and toxicity were also discussed. The problems and shortcomings that need be resolved to accelerate the actual deployment of SRSP were highlighted in this review.

Stochastic Geometry Analysis of Sojourn Time in RF/VLC Hybrid Networks

The spectrum scarcity in the radio frequency (RF) communication in indoor environments motivates the integration of an alternative technology like visible light communication (VLC) with the existing RF architecture that results in a hybrid RF/VLC network. While VLC helps offloading the congested RF spectrum by offering capacity-per-area improvements, the resulting heterogeneity and narrow coverage areas of optical base stations (BSs) impose several challenges for user mobility such as unnecessary handovers. To help addressing these challenges, in this paper, we derive the mean and the distribution of sojourn time in RF/VLC hybrid networks. The mathematical analysis conducted in this paper makes use of the tools from stochastic geometry and abstracting the BSs’ locations via two independent homogeneous Poisson point processes (PPPs). Since PPP modeling is yet to be well established for RF/VLC hybrid networks, we compare the PPP based analytical results to those obtained for an actual deployment, a Matérn hard-core point process (MHCPP) based deployment, and a deterministic square lattice deployment of VLC luminaries. Furthermore, we utilize the sojourn time distribution to calculate the unnecessary handover probability. Our numerical results show the interplay between the sojourn time and the receiver field of view as a function of BS density and they highlight the cost of BS densification in terms of unnecessary handovers.

Simulation study to assess the effectiveness of gamma radiation for inactivation of viruses on food packaging material

The recent COVID-19 pandemic spread across the globe has raised the concern about the possible transmission of viruses through food packaging material during domestic and international trade. Therefore, mitigation strategies are needed to address these safety issues. Preliminary in-silico study showed that interactions between food packaging material and viral surface proteins were possibly hydrophobic in nature with most favourable interaction having a binding free energy of −5.24 kcal/mol. Since these interactions can cause viruses to adsorb on the food packets and get transmitted during supply chain, it is necessary to inactivate the viruses. In this context, efficacy of gamma irradiation in inactivating the viruses on the food packaging material was assessed. For this simulation study P1 (virulent) bacteriophage of E. coli was used as a model system. Gamma irradiation of food packets at an absorbed dose >8 kGy was found to completely inactivate the infectivity of P1(virulent) bacteriophage when co-cultured with E. coli host and assayed for viral plaque formation. Reduction in infectivity of P1(vir) phage was more prominent at ambient temperature (25 ± 2 °C) as compared to cold temperature (6 ± 2 °C) when assayed after storage (one week). Gamma irradiation (2 kGy) completely inactivated the virus particles on food packets when stored for 1 week at both the above temperatures. It is thus proposed that gamma irradiation (2 kGy) can possibly be integrated as a final treatment of the packaged food products to rule out the possibility of viral transmission. However, the efficacy of radiation processing against different pathogenic viruses needs to be determined prior to actual commercial deployment.

An improved algae-YOLO model based on deep learning for object detection of ocean microalgae considering aquacultural lightweight deployment

Algae are widely distributed and have a considerable impact on water quality. Harmful algae can degrade water quality and be detrimental to aquaculture, while beneficial algae are widely used. The accuracy and speed of existing intelligent algae detection methods are available, but the size of parameters of models is large, the equipment requirements are high, the deployment costs are high, and there is still little research on lightweight detection methods in the area of algae detection. In this paper, we propose an improved Algae-YOLO object detection approach, which is based on ShuffleNetV2 as the YOLO backbone network to reduce the parameter space, adapting the ECA attention mechanism to improve detection accuracy, and redesigning the neck structure replacing the neck structure with ghost convolution module for reducing the size of parameters, finally the method achieved the comparable accuracy. Experiments showed that the Algal-YOLO approach in this paper reduces the size of parameters by 82.3%, and the computation (FLOPs) is decreased from 16G to 2.9G with less loss of accuracy, and mAP by only 0.007 when compared to the original YOLOv5s. With high accuracy, the smaller model size are achieved, which reduces the equipment cost during actual deployment and helps to promote the practical application of algae detection.

Optimizing Face Recognition Inference with a Collaborative Edge–Cloud Network

The rapid development of deep-learning-based edge artificial intelligence applications and their data-driven nature has led to several research issues. One key issue is the collaboration of the edge and cloud to optimize such applications by increasing inference speed and reducing latency. Some researchers have focused on simulations that verify that a collaborative edge-cloud network would be optimal, but the real-world implementation is not considered. Most researchers focus on the accuracy of the detection and recognition algorithm but not the inference speed in actual deployment. Others have implemented such networks with minimal pressure on the cloud node, thus defeating the purpose of an edge-cloud collaboration. In this study, we propose a method to increase inference speed and reduce latency by implementing a real-time face recognition system in which all face detection tasks are handled on the edge device and by forwarding cropped face images that are significantly smaller than the whole video frame, while face recognition tasks are processed at the cloud. In this system, both devices communicate using the TCP/IP protocol of wireless communication. Our experiment is executed using a Jetson Nano GPU board and a PC as the cloud. This framework is studied in terms of the frame-per-second (FPS) rate. We further compare our framework using two scenarios in which face detection and recognition tasks are deployed on the (1) edge and (2) cloud. The experimental results show that combining the edge and cloud is an effective way to accelerate the inferencing process because the maximum FPS achieved by the edge-cloud deployment was 1.91× more than the cloud deployment and 8.5× more than the edge deployment.

AN OPTIMAL DEPLOYMENT METHOD OF UWB POSITIONING BASE-STATION

Abstract. Aiming at the prominent problem of high deployment cost of UWB (Ultra Wideband) positioning system and the waste of resources caused by repeated coverage of UWB base-station signals, the optimal deployment of the location and number of UWB positioning base stations is studied. Using Genetic Optimization Algorithm to optimize the objective function with Position Dilution of Precision (PDOP) and base-station Sharing-Value. The simulation and experimental results show that for the optimization of the location of UWB base stations, the three-dimensional average positioning error of the optimized layout method can be reduced by at least 0.3m compared with other typical layout methods; for the optimization of the number of base stations, the number by the optimized layout method is significantly lower than the typical layout method and decreases more as the actual base station deployment area increases. Therefore, the performance of the proposed method for optimal deployment of UWB positioning base stations is fully verified, which provides a reference for the optimal deployment of UWB positioning base stations in practical applications.

A Collision Reduction Adaptive Data Rate Algorithm Based on the FSVM for a Low-Cost LoRa Gateway

LoRa (Long Range), a wireless communication technology for low power wide area networks (LPWANs), enables a wide range of IoT applications and inter-device communication, due to its openness and flexible network deployment. In the actual deployment and operation of LoRa networks, the static link transmission scheme does not make full use of the channel resources in the time-varying channel environment, resulting in a poor network performance. In this paper, we propose a more effective adaptive data rate (ADR) algorithm for low-cost gateways, we firstly analyze the impact of the different hardware parameters (RSSI, SNR) on the link quality and classify the link quality using the fuzzy support vector machine (FSVM). Secondly, we establish an end device (ED) throughput model and energy consumption model and design different adaptive rate algorithms, according to the different link quality considering both the link-level performance and the MAC layer performance. The proposed algorithm uses machine learning to classify the link quality, which can accurately classify the link quality using a small amount of data, compared to other adaptive rate algorithms, and the link parameter adaptation algorithm can maximize the throughput while ensuring the link stability, by considering the link-level performance and the MAC layer performance, compared to other algorithms. The results show that it outperforms the standard LoRaWAN ADR algorithm in both the single ED and the multi ED scenarios, in terms of the packets reception rate (PRR) and the network throughput. Compared to the LoRaWAN ADR in 32 multi-ED scenarios, the proposed algorithm improves the throughput by 34.12% and packets the reception rate by 26%, significantly improving the network throughput and the packets reception rate.

The Definitive Guide to Actual Sensor Network Deployments in Research Studies from 2013–2017: A Systematic Review

The research community has been working on sensor networks for more that seven decades and altogether more than a million research articles on sensor networks have been published, with this number growing every day. In this article, we try to provide a thorough and complete systematic review on the sensor network field resulting from representative subset of more than 3000 articles which include actual sensor network deployments. After a thorough analysis of these data, we provide a definitive overview of the trends in sensor network deployment described in published research articles.

Knowing and Feeling the “China Dream”: Logic and Rhetoric in the Political Language of Xi’s China

The Xi era can be characterised as an ambitious and nationalistic one. Aims such as rejuvenating the Chinese nation and fulfilling the country's historic mission to be a great, rich country are often referred to in elite leadership discourse and state produced propaganda. “China Dream” is amongst the most important slogans used in this language. In terms of the context in which this phrase occurs and the actual deployment, it is one that carries broad connotations and implications about an era where, in ways similar to politics in the West, issues around identity are key, and the important thing is to feel, not just to believe. “China Dream” is a form of language that speaks into the public's deeper instincts and sentiments.

Multi-AGV's Temporal Memory-Based RRT Exploration in Unknown Environment

With the increasing need for multi-robot for exploring the unknown region in a challenging environment, efficient collaborative exploration strategies are needed for achieving such feat. A frontier-based Rapidly-Exploring Random Tree (RRT) exploration can be deployed to explore an unknown environment. However, its' greedy behavior causes multiple robots to explore the region with the highest revenue, which leads to massive overlapping in exploration process. To address this issue, we present a temporal memory-based RRT (TM-RRT) exploration strategy for multi-robot to perform robust exploration in an unknown environment. It computes adaptive duration for each frontier assigned and calculates the frontier's revenue based on the relative position of each robot. In addition, each robot is equipped with a memory consisting of frontier assigned and share among fleets to prevent repeating assignment of same frontier. Through both simulation and actual deployment, we have shown the robustness of TM-RRT exploration strategy by completing the exploration in a <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$25.0\,\rm{m}\times \text{54.0}\,\rm{m}$</tex-math></inline-formula> ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\text{1350.0}\,\rm{m}^{2}$</tex-math></inline-formula> ) area, while the conventional RRT exploration strategy falls short.

Latency Analysis of Blockchain-Based SSI Applications

Several revolutionary applications have been built on the distributed ledgers of blockchain (BC) technology. Besides cryptocurrencies, many other application fields can be found in smart systems exploiting smart contracts and Self Sovereign Identity (SSI) management. The Hyperledger Indy platform is a suitable open-source solution for realizing permissioned BC systems for SSI projects. SSI applications usually require short response times from the underlying BC network, which may vary highly depending on the application type, the used BC software, and the actual BC deployment parameters. To support the developers and users of SSI applications, we present a detailed latency analysis of a permissioned BC system built with Indy and Aries. To streamline our experiments, we developed a Python application using containerized Indy and Aries components from official Hyperledger repositories. We deployed our experimental application on multiple virtual machines in the public Google Cloud Platform and on our local, private cloud using a Docker platform with Kubernetes. We evaluated and compared their performance benchmarked by Read and Write latencies. We found that the local Indy ledger reads and writes 30–50%, and 65–85% faster than the Indy ledger running on the Google Cloud Platform, respectively.

From compute to care: Lessons learned from deploying an early warning system into clinical practice.

BackgroundDeploying safe and effective machine learning models is essential to realize the promise of artificial intelligence for improved healthcare. Yet, there remains a large gap between the number of high-performing ML models trained on healthcare data and the actual deployment of these models. Here, we describe the deployment of CHARTwatch, an artificial intelligence-based early warning system designed to predict patient risk of clinical deterioration.MethodsWe describe the end-to-end infrastructure that was developed to deploy CHARTwatch and outline the process from data extraction to communicating patient risk scores in real-time to physicians and nurses. We then describe the various challenges that were faced in deployment, including technical issues (e.g., unstable database connections), process-related challenges (e.g., changes in how a critical lab is measured), and challenges related to deploying a clinical system in the middle of a pandemic. We report various measures to quantify the success of the deployment: model performance, adherence to workflows, and infrastructure uptime/downtime. Ultimately, success is driven by end-user adoption and impact on relevant clinical outcomes. We assess our deployment process by evaluating how closely we followed existing guidance for good machine learning practice (GMLP) and identify gaps that are not addressed in this guidance.ResultsThe model demonstrated strong and consistent performance in real-time in the first 19 months after deployment (AUC 0.76) as in the silent deployment heldout test data (AUC 0.79). The infrastructure remained online for >99% of time in the first year of deployment. Our deployment adhered to all 10 aspects of GMLP guiding principles. Several steps were crucial for deployment but are not mentioned or are missing details in the GMLP principles, including the need for a silent testing period, the creation of robust downtime protocols, and the importance of end-user engagement. Evaluation for impacts on clinical outcomes and adherence to clinical protocols is underway.ConclusionWe deployed an artificial intelligence-based early warning system to predict clinical deterioration in hospital. Careful attention to data infrastructure, identifying problems in a silent testing period, close monitoring during deployment, and strong engagement with end-users were critical for successful deployment.

Establishing and Operating Experiential Benchmark Setups of dc Microgrids: A Guideline

Increased attention is given to microgrids for their numerous advantages over conventional power networks. However, there are still a limited number of real microgrid networks that are currently in service, and this raises shortages in expertise on microgrids testing and evaluation before actual deployment. Low-cost, rapid, realistic, and flexible testing are the key challenges that face the microgrids’ validation. This article focuses on the importance of experiential validations to reduce the risks and ensure the desired microgrids performances. It highlights the major challenges for establishing and operating realistic and accurate experimental microgrid testbeds. It also provides guidelines and recommendations for the validation of microgrid studies. An experimental dc microgrid setup showcase is demonstrated, and helpful tips and remarks are suggested. Different case studies on the developed dc microgrid’s setup are presented, and the authors’ experiences and analyses are shared. The experimental results show the accuracy of microgrid testbeds and their expansion capabilities as well as their effectiveness in testing various microgrid conditions.

An Experimental Analysis of Cyclic and Reference Signals of 4G LTE for TOA Estimation and Positioning in Mobile Fading Environments

By design, the 4G LTE signals contain a few patterns that repeat in time and frequency (e.g., various reference signals and the cyclic prefix) to facilitate timing synchronization, frequency error estimation, and channel equalization. For communications, these signal components are considered “overhead” serving the purpose of extracting information-carrying symbols for data decoding. When used as signals of opportunity for navigation, they provide times of arrival (TOA), and changes in TOA estimates can be related to the receiver displacement. Indeed, successful use of LTE signals has been reported for ground vehicles, airborne platforms, and indoor positioning in the literature. Even though LTE signals are well specified in the 3GPP standards, some aspects are left to operators as implementation options that depend on the actual deployment environment and local traffic conditions. To better understand the practical aspects of in-the-air LTE signals, experiments were carried out and an initial study of the cyclic and reference patterns in LTE signals for TOA estimation is presented in this article and in particular the effect of antenna ports on TOA estimation. This article aims at providing essential and insightful information directly useable for developing TOA estimation algorithms and ultimately for navigation with LTE signals.

Reproducible and low‐power multistate bio‐memristor from interpenetrating network electrolyte design

AbstractBio‐memristor can address the inherent limitations of conventional memory components in artificial perceptual systems due to their biocompatibility with biological tissue. The actual deployment of bio‐memristor is restricted by the lack of reproducibility, high power consumption, and insufficient storage capacity. Here, a reproducible and low‐power multistate bio‐memristor is developed by designing the chitosan (CS)‐reduced graphene oxide (rGO) interpenetrating network electrolyte. The interpenetrating network structure of the CS‐rGO electrolyte reinforces structural stability and improves ionic conductivity. The bio‐memristor equipped with CS‐rGO active layer shows stable bipolar resistive switching up to 100 consecutive cycles, reproducible multistate storage with six different memory states, and low programming power of 9.4 μW. The fabricated biocompatible CS‐rGO device also exhibits deformation stability of memory operation over 103 bending cycles, high biocompatibility with HEK293 cells, and skin adhesion. This work provides an enlightening design strategy to develop high‐performance bio‐memristors for applications in artificial perceptual systems.image

Inspecting Decorative Ceramic Defects by Fusing Convolutional Neural Network and Image Recognition.

The intelligent inspection of ceramic decorative defects is one of the hot research at present. This work aims to improve the defect inspection automation of finished decorative ceramic workpieces. First, it introduces the multi-target detection algorithm and compares the performance of different network models on the public data set. Second, the initial images are collected on the spot. The initial pictures are easy to produce noise in actual deployment, affecting the image quality. Therefore, image preprocessing is performed for the initial images, and a median filtering method is used to calculate the denoising. Finally, the original You Only Look Once version 3 network model is realized. Based on this, the decorative ceramic-oriented Automated Surface Defect Inspection model is proposed. Then, decorative ceramic defect images are inputted for model training. The experimental conclusions are deeply studied and analyzed. The results show that the proposed decorative ceramic-oriented Automated Surface Defect Inspection model based on Deep Learning technology has good feature extraction and inspection ability. The detection accuracy is 94.90% on the test set, and the detection speed reaches 25 frames per second. Compared with the traditional manual inspection method, the proposed model greatly improves the inspection effect and can meet the on-site inspection requirements of surface defects of decorative ceramics under complex backgrounds. It is of great significance to improve the quality inspection efficiency and economic benefits of China's decorative ceramics industry.

A practical deep reinforcement learning framework for multivariate occupant-centric control in buildings

Reinforcement learning (RL) has been shown to have the potential for optimal control of heating, ventilation, and air conditioning (HVAC) systems. Although research on RL-based building control has received extensive attention in recent years, there is limited real-world implementation to evaluate its performance while keeping occupants in the loop. Additionally, many HVAC systems consist of multiple subsystems, but conventional RL algorithms face significant challenges when dealing with high-dimensional action spaces. This study proposes a practical deep reinforcement learning (DRL) based multivariate occupant-centric control framework that considers personalized thermal comfort and occupant presence. Specifically, Branching Dueling Q-network (BDQ) is leveraged as the learning agent to efficiently solve the multi-dimensional control task, and a tabular-based personal comfort modeling method is applied that is naturally integrated into human-in-the-loop operations. The BDQ agent is pre-trained in a virtual environment, followed by online deployment in a real office space for 5-dimensional action control. Based on the actual deployment and real-time comfort votes, our results showed a 14% reduction in cooling energy and an 11% improvement in total thermal acceptability.

Spatial attention-based convolutional transformer for bearing remaining useful life prediction

The remaining useful life (RUL) prediction is of significance to the health management of bearings. Recently, deep learning has been widely investigated for bearing RUL prediction due to its great success in sequence learning. However, the improvement of the prediction accuracy of existing deep learning algorithms heavily relies on feature engineering such as handcrafted feature generation and time–frequency transformation, which increase the complexity and difficulty of the actual deployment. In this paper, a novel spatial attention-based convolutional transformer (SAConvFormer) is proposed to establish an accurate bearing RUL prediction model based on raw vibration data without prior knowledge or feature engineering. In this algorithm, firstly, a convolutional neural network enhanced by a spatial attention mechanism is proposed to squeeze the feature maps and extract the local and global features from raw bearing vibration data effectively. Then, the extracted senior features are fed into a transformer network to further explore the sequential patterns relevant to the bearing RUL. An experimental study using the XJTU-SY rolling bearings dataset revealed the merits of the proposed deep learning algorithm in terms of root-mean-square-error (RMSE) and mean-absolute-error (MAE) in comparison with other state-of-the-art algorithms.

PV-EncoNet: Fast Object Detection Based on Colored Point Cloud

Object detection is the most critical and foundational sensing module for the autonomous movement platform. However, most of the existing deep learning solutions are based on GPU servers, which limits their actual deployment. We present an efficient multi-sensor fusion based object detection model that can be deployed on the off-the-shelf edge computing device for the vehicle platform. To achieve real-time target detection, the model eliminates a large number of invalid point clouds through ground filtering algorithm, and then adds texture information (fused from camera image) through point cloud coloring to enhance features. The proposed PV-EncoNet efficiently encodes both the spatial and texture features of each colored point through point-wise and voxel-wise encoding, and then predicts the position, heading and class of the objects. The final model can achieve about 17.92 and 24.25 Frame per Second (FPS) on two different edge computing platforms, and the detection accuracy is comparable with the state-of-the-art models on the KITTI public dataset (i.e., 88.54% for cars, 71.94% for pedestrians and 73.04% for cyclists). The robustness and generalization ability of the PV-EncoNet for the 3D colored point cloud detection task is also verified by deploying it on the local vehicle platform and testing it on real road conditions.