Non-nucleophilic and non-fluorinated compounds are the most important class of solvents to enable sustainable rechargeable magnesium (Mg) batteries; however, they suffer from poor stability due to the formation of an unstable solid–electrolyte-interphase (SEI). Here, we design a hybrid-solvent electrolyte by dispersing a non-nucleophilic Mg(TFSI)2–MgCl2-1,2-dimethoxyethane (DME) solution in a non-fluorinated weakly coordinating solvent (tetrahydrofuran, THF) to prevent the decomposition of DME and form a stable SEI. This strategy significantly improved the cycle life of a Mg anode from 24 to 4000 hours (in Mg–Mg symmetric cells) and realized a stable cycle life of Mg–Mo6S8 full cells for more than 2300 hours with an average CE of 99.96%. The non-fluorinated weakly coordinating solvent THF suppressed uneven Mg growth and contributed to the formation of a polymeric SEI, which filled the cracks of the pristine SEI, preventing side reactions and passivation. This strategy revealed the critical role of weakly coordinating solvents in stabilizing the Mg anode for reversible Mg batteries.
More and more distributed energy resources are integrated into regional integrated energy systems (RIES), which poses great challenges to energy balance. RIES can coordinate power, gas, heat, and cooling systems jointly to enhance energy efficiency and explore flexibility for distributed renewable energy accommodation. Multi-energy demand forecasts are the basis of the flexible operation of RIES. However, the multi-energy demands are deeply coupled in RIES. Related researches utilize cross-sector information from different sectors to tackle the coupling relationship. Nevertheless, the unknown influence of cross-sector information offered by other sectors varies with the operating pattern, which is difficult to be evaluated. This work proposes an operating pattern recognition-based method for adaptive cross-sector information identification. Firstly, the K-means cluster algorithm is adopted to identify different operating patterns. After that, cross-sector information is selected based on the Pearson coefficient. Furthermore, two models, i.e., the local model and fine-tuned model, are modified with the assistance of selected cross-sector information. The proposed method is evaluated by a RIES with three energy types (electricity, chill water, and steam). The proposed two methods acquire better accuracy than the three benchmark models. Moreover, the Shapley value is applied to verify the contribution of selected cross-sector information. The result shows that all selected cross-sector information plays a significant role in load prediction.
Photonic RF transversal signal processors, which are equivalent to reconfigurable electrical digital signal processors but implemented with photonic technologies, are attractive for high-speed information processing. Optical microcombs are extremely powerful as sources for RF photonics since they can generate many wavelength channels from compact micro-resonators, offering greatly reduced size, power consumption, and complexity. Recently, a variety of signal processing functions have been demonstrated using microcomb-based photonic RF transversal signal processors. Here, we provide a detailed analysis for quantifying the processing accuracy of microcomb-based photonic RF transversal signal processors. First, we investigate the theoretical limitations of the processing accuracy determined by tap number, signal bandwidth, and pulse waveform. Next, we discuss the practical error sources from different experimental components of the signal processors. Finally, we assess the relative contributions of the two to the overall accuracy. We find that the overall accuracy is mainly limited by experimental factors when the processors are properly designed to minimize the theoretical limitations, and that these remaining errors can be further greatly reduced by introducing feedback control to calibrate the processors' impulse response. These results provide a useful guide for designing microcomb-based photonic RF transversal signal processors to optimize their accuracy.
The extensive conversion of carbon-rich coastal wetland to aquaculture ponds in the Asian Pacific region has caused significant changes to the sediment properties and carbon cycling. Using field sampling and incubation experiments, the sediment anaerobic CO2 production and CO2 emission flux were compared between a brackish marsh and the nearby constructed aquaculture ponds in the Min River Estuary in southeastern China over a three-year period. Marsh sediment had a higher total carbon and lower C:N ratio than aquaculture pond sediment, suggesting the importance of marsh vegetation in supplying labile organic carbon to the sediment. Conversion to aquaculture ponds significantly decreased sediment anaerobic CO2 production rates by 69.2% compared to the brackish marsh, but increased CO2 emission, turning the CO2 sink (−490.8 ± 42.0 mg m−2 h−1 in brackish marsh) into a source (6.2 ± 3.9 mg m−2 h−1 in aquaculture pond). Clipping the marsh vegetation resulted in the highest CO2 emission flux (382.6 ± 46.7 mg m−2 h−1), highlighting the critical role of marsh vegetation in capturing and sequestering carbon. Sediment anaerobic CO2 production and CO2 uptake (in brackish marsh) and emission (in aquaculture ponds) were highest in the summer, followed by autumn, spring and winter. Redundancy analysis and structural equation modeling showed that the changes of sediment temperature, salinity and total carbon content accounted for more than 50% of the variance in CO2 production and emission. Overall, the results indicate that vegetation clearing was the main cause of change in CO2 production and emission in the land conversion, and marsh replantation should be a primary strategy to mitigate the climate impact of the aquaculture sector.
Sub-harmonic control (SHC) is an excellent method for wireless charging system because it is easy to achieve full-load range soft switching and unity power factor. However, in existing works, SHC is only applied to the system with primary series compensation topology (PSCT), which has the drawbacks of low degrees of design freedom, high sensitivity to misalignments and no open-circuit protection. To overcome these issues, in this paper, systematic synthesis of primary high-order compensation topology (PHCT) for WPT is identified in terms of general T-branch structure. Afterwards, the applicability of PHCT to SHC is explored by gaining insight into their filter characteristics. It reveals that when SHC is applied, four kinds of PHCTs are superior over other candidates in performance because the unity power factor and soft switching can be achieved without limitation on pulse distribution. Then, six kinds of PHCTs exhibit a moderate performance as they may nullify the strengths of SHC under some specific pulse distributions. The rest of PHCTs are not applicable to SHC no matter what the pulse distribution. Finally, the experimental results on prototype with jointly applying SHC and PHCT are examined.
To reduce carbon emissions and pursue sustainable economic development, China's central government formulated the low-carbon city pilot (LCCP) policy. Current studies focus primarily on the impact of the policy at the macro level (provinces and cities). So far, no study has looked at the impact of the LCCP policy on companies' environmental expenditures. Besides, as the LCCP policy is a weak-constraining central policy, it is interesting to see how it works at the company level. We employ company-level empirical data and the Propensity Score Matching-Difference in Differences (PSM-DID) method, which outperforms the traditional DID model in avoiding sample selection bias, to address the above issues. We concentrate on the second phase of the LCCP policy from 2010 to 2016, encompassing 197 listed companies in China's secondary and transportation industries. Our statistical results show that if the listed company's host city has piloted the LCCP policy, the company's environmental expenditures are reduced by 0.91 points at the 1% significance level. The above finding calls attention to the policy-implementation gap between the central and the local governments in China, which may make those weak-constraining central policies like the LCCP policy have purpose-defeating outcomes at the company level.
A multi-motor system supplied from the reduced-switch-count voltage source inverter (VSI) is considered in this paper. Firstly, an improved multiple vectors model predictive control (MV-MPC) scheme is proposed for driving n three-phase permanent magnet synchronous motors (PMSMs) by (2 n +1)-leg VSI. To realize independent control of multi-motor, a control period is partitioned into multi-interval, one per motor. In the proposed MV-MPC scheme, the quasi-optimal voltage vectors (VVs) are selected to replace zero VVs under the restriction of identical common leg switching states. Thus, a whole control period can be fully utilized to generate multiple active VVs for a motor. Moreover, the overcurrent in the common leg is discussed and addressed. Take the five-leg VSI as an example, the expression of overcurrent in the common leg is first derived. Next, a general overcurrent elimination method based on current phase regulation is designed for the multi-motor system operating at an identical speed. Finally, the proposed MV-MPC scheme combined with the overcurrent elimination strategy is experimentally carried out. The test results indicate that the MV-MPC method can effectively reduce current ripple and achieve a faster dynamic response. The overcurrent elimination method can minimize the overcurrent regardless of the load conditions of the motors.
Even though conventional wisdom claims that in order to improve viewer engagement, the cloud-edge providers should serve the viewers with the nearest edge nodes, however, we show that doing this for crowdsourced live streaming (CLS) services can introduce significant costs inefficiency. In this paper, we first carry out large-scale measurement analysis by using the real-world service data from Huawei Cloud, a representative cloud-edge provider in China. We observe that the massive number of channels has proposed great burdens to the operating expenditure of the cloud-edge providers, and most importantly, unbalanced viewer distribution makes the edge nodes suffer significant costs inefficiency. To tackle the above concerns, we propose AggCast , a novel CLS scheduling framework to optimize the edge node utilization for the cloud-edge provider. The core idea of AggCast is to aggregate some viewers that are initially scattered on different regions, and assign them to fewer pre-selected nodes, thereby reducing bandwidth costs. In particular, by integrating the useful insights obtained from our large-scale measurement, AggCast can not only ensure that quality of experience (QoS) does not suffer degradation, but also satisfy the systematic requirements of CLS services. AggCast has been A/B tested and fully deployed. The online and trace-driven experiments show that, compared to the most prevalent method, AggCast saves over 16.3% back-to-source (BTS) bandwidth costs while significantly improving QoS (startup latency, stall frequency and stall time are reduced over 12.3%, 4.57% and 3.91%, respectively)
Nervous system disease (NSD) is a global health burden with increasing prevalence in the last 30 years. There is evidence that greenness can improve nervous system health through a variety of mechanisms; however, the evidence is inconsistent. In the present systematic review and meta-analysis, we examined the relationship between greenness exposure and NSD outcomes. Studies on the relationship between greenness and NSD health outcomes published till July 2022 were searched in PubMed, Cochrane, Embase, Scopus, and Web of Science. In addition, we searched the cited literature and updated our search on Jan 20, 2023, to identify any new studies. We included human epidemiological studies that assess the association of greenness exposure with the risk of NSD. Greenness exposure was measured using NDVI (the normalized difference vegetation index) and the outcome was the mortality or morbidity of NSD. The pooled relative risks (RRs) were estimated using a random effects model. Of 2059 identified studies, 15 studies were included in our quantitative evaluation, in which 11 studies found a significant inverse relationship between the risk of NSD mortality or incidence/prevalence and an increase in surrounding greenness. The pooled RRs for cerebrovascular diseases (CBVD), neurodegenerative diseases (ND), and stroke mortality were 0.98 (95 % CI: 0.97, 1.00), 0.98 (95 % CI: 0.98, 0.99), and 0.96 (95 % CI: 0.93, 1.00), respectively. The pooled RRs for PD incidence and stroke prevalence/incidence were 0.89 (95 % CI: 0.78, 1.02) and 0.98 (95 % CI: 0.97, 0.99), respectively. The confidence of evidence for ND mortality, stroke mortality, and stroke prevalence/incidence was downgraded to "low", while CBVD mortality and PD incidence were downgraded to "very low" due to inconsistency. We found no evidence of publication bias and the sensitivity analysis results of all subgroups are robust except for the stroke mortality subgroup. This is the first comprehensive meta-analysis of greenness exposure and NSD outcomes in which an inverse relationship was observed. It is necessary to conduct further research to ascertain the role greenness exposure plays in various NSDs and the management of greenness should be considered a public health strategy.
Wetland sediment is an important nitrogen pool and a source of the greenhouse gas nitrous oxide (N2O). Modification of coastal wetland landscape due to plant invasion and aquaculture activities may drastically change this N pool and the related dynamics of N2O. This study measured the sediment properties, N2O production and relevant functional gene abundances in 21 coastal wetlands across five provinces along the tropical-subtropical gradient in China, which all had experienced the same sequence of habitat transformation from native mudflats (MFs) to invasive Spartina alterniflora marshes (SAs) and subsequently to aquaculture ponds (APs). Our results showed that change from MFs to SAs increased the availability of NH4+-N and NO3--N and the abundance of functional genes related to N2O production (amoA, nirK, nosZ Ⅰ, and nosZ Ⅱ), whereas conversion of SAs to APs resulted in the opposite changes. Invasion of MFs by S. alterniflora increased N2O production potential by 127.9%, whereas converting SAs to APs decreased it by 30.4%. Based on structural equation modelling, nitrogen substrate availability and abundance of ammonia oxidizers were the key factors driving the change in sediment N2O production potential in these wetlands. This study revealed the main effect patterns of habitat modification on sediment biogeochemistry and N2O production across a broad geographical and climate gradient. These findings will help large-scale mapping and assessing landscape change effects on sediment properties and greenhouse gas emissions along the coast.
Recently proposed Split Learning (SL) is a promising distributed machine learning paradigm that enables machine learning without accessing the raw data of the clients. SL can be viewed as one specific type of serial federation learning. However, deploying SL on resource-constrained IoT devices still has some limitations, including high communication costs and catastrophic forgetting problems caused by imbalanced data distribution of devices. In this paper, we design and implement IoTSL, which is an efficient distributed learning framework for efficient cloudedge collaboration in IoT systems. IoTSL combines generative adversarial networks (GANs) and differential privacy techniques to train local data-based generators on participating devices, and generate data with privacy protection. On the one hand, IoTSL pre-trains the global model using the generative data, and then fine-tunes the model using the local data to lower the communication cost. On the other hand, the generated data is used to impute the missing classes of devices to alleviate the commonly seen catastrophic forgetting phenomenon. We use three common datasets to verify the proposed framework. Extensive experimental results show that compared to the conventional SL, IoTSL significantly reduces communication costs, and efficiently alleviates the catastrophic forgetting phenomenon.
This work deals with the optimal epidemics policy-seeking problem on networks-of-networks (NoN) in the presence of unknown malicious adding-edge attacks. This problem is investigated in a framework of games-of-games (GoG), in which the conflicts between each network policymaker and the attacker are captured by a series of the Stackelberg games, while all network policymakers together compose a Nash game. First, the tolerable maximum attack magnitude is investigated and given implicitly. Then, we prove the existence of the gestalt Nash equilibrium (GNE) under mild attacks bounded by the above magnitude. A Heuristic algorithm based on iterative geometric programming is proposed to seek the GNE of the above GoG, whose asymptotical convergence is verified. Correspondingly, a greedy Heuristic strategy for the malicious attacker to compromise the NoN topology is developed. The practicability and validity of the above theoretical results and algorithms are illustrated via a simulation example.
We resolve three long-standing open problems, namely the (algorithmic) decidability of network coding, the decidability of conditional information inequalities, and the decidability of conditional independence implication among random variables, by showing that these problems are undecidable. The proof utilizes a construction inspired by Herrmann’s arguments on embedded multivalued database dependencies, a network studied by Dougherty, Freiling and Zeger, together with a novel construction to represent group automorphisms on top of the network.
In this paper, we propose a novel framework for tactile-based dexterous manipulation learning with a blind anthropomorphic robotic hand, i.e. without visual sensing. First, object-related states were extracted from the raw tactile signals by a graph-based perception model - TacGNN. The resulting tactile features were then utilized in the policy learning of an in-hand manipulation task in the second stage. This method was examined by a Baoding ball task - simultaneously manipulating two spheres around each other by 180 degrees in hand. We conducted experiments on object states prediction and in-hand manipulation using a reinforcement learning algorithm (PPO). Results show that TacGNN is effective in predicting object-related states during manipulation by decreasing the RMSE of prediction to 0.096cm comparing to other methods, such as MLP, CNN, and GCN. Finally, the robot hand could finish an in-hand manipulation task solely relying on the robotic own perception - tactile sensing and proprioception. In addition, our methods are tested on three tasks with different difficulty levels and transferred to the real robot without further training. https://sites.google.com/view/tacgnn</uri
The automation of unmanned aerial vehicles (UAVs) has been greatly promoted by visual object tracking methods with onboard cameras. However, the random and complicated real noise produced by the cameras seriously hinders the performance of state-of-the-art (SOTA) UAV trackers, especially in low-illumination environments. To address this issue, this work proposes an efficient plug-and-play cascaded denoising Transformer (CDT) to suppress cluttered and complex real noise, thereby boosting UAV tracking performance. Specifically, the novel U -shaped cascaded denoising network is designed with a streamlined structure for efficient computation. Additionally, shallow feature deepening (SFD) encoder and multi-feature collaboration (MFC) decoder are constructed based on multi-head transposed self-attention (MTSA) and multi-head transposed cross-attention (MTCA), respectively. A nested residual feed-forward network (NRFN) is developed to focus more on high-frequency information represented by noise. Extensive evaluation and test experiments demonstrate that the proposed CDT has a remarkable denoising effect and improves UAV nighttime tracking performance.
Event cameras that asynchronously output low-latency event streams provide great opportunities for state estimation under challenging situations. Despite event-based visual odometry having been extensively studied in recent years, most of them are based on the monocular, while few research on stereo event vision. In this paper, we present ESVIO, the first event-based stereo visual-inertial odometry, which leverages the complementary advantages of event streams, standard images, and inertial measurements. Our proposed pipeline includes the ESIO (purely event-based) and ESVIO (event with image-aided), which achieves spatial and temporal associations between consecutive stereo event streams. A well-design back-end tightly-coupled fused the multi-sensor measurement to obtain robust state estimation. We validate that both ESIO and ESVIO have superior performance compared with other image-based and event-based baseline methods on public and self-collected datasets. Furthermore, we use our pipeline to perform onboard quadrotor flights under low-light environments. Autonomous driving data sequences and real-world large-scale experiments are also conducted to demonstrate long-term effectiveness. We highlight that this work is a real-time, accurate system that is aimed at robust state estimation under challenging environments.
Stretchable power devices and self-powered sensors have become increasingly desired for wearable electronics and artificial intelligence. In this study, an all-solid-state triboelectric nanogenerator (TENG) is reported, whose one solid-state structure prevents delamination during stretch and release cycles and increasing the patch adhesive force (3.5 N) and strain (586% elongation at break). Through the synergetic virtues of stretchability, ionic conductivity, and excellent adhesion to the tribo-layer, reproducible open-circuit voltage (VOC ) of 84 V, charge (QSC ) of 27.5 nC, and short-circuit current (ISC ) of 3.1 µA after drying at 60°C or 20,000 contact-separation cycles are obtained. Apart from contact-separation, this device shows unprecedented electricity generation through stretch-release of solid materials leading to a linear relationship between VOC and strain. For the first time, this work provides a clear explanation of the working mechanism of contact-free stretching-releasing and investigates the relationships of exerted force, strain, thickness of the device, and electric output. Benefitting from the one solid-state structure, this contact-free device remains stable even after repeated stretch-release cycling, maintaining 100% of its VOC after 2500 stretch-release cycles. These findings provide a strategy toward highly conductive and stretchable electrodes for harvesting mechanical energy and health monitoring.
Intrinsic plasticity of neurons, such as spontaneous threshold lowering (STL) to modulate neuronal excitability, is key to spatial attention of biological neural systems. In‐memory computing with emerging memristors is expected to solve the memory bottleneck of the von Neumann architecture commonly used in conventional digital computers and is deemed a promising solution to this bioinspired computing paradigm. Nonetheless, conventional memristors are incapable of implementing the STL plasticity of neurons due to their first‐order dynamics. Here, a second‐order memristor is experimentally demonstrated using yttria‐stabilized zirconia with Ag doping (YSZ:Ag) that exhibits STL functionality. The physical origin of the second‐order dynamics, i.e., the size evolution of Ag nanoclusters, is uncovered through transmission electron microscopy (TEM), which is leveraged to model the STL neuron. STL‐based spatial attention in a spiking convolutional neural network (SCNN) is demonstrated, improving the accuracy of a multiobject detection task from 70% (20%) to 90% (80%) for the object within (outside) the area receiving attention. This second‐order memristor with intrinsic STL dynamics paves the way for future machine intelligence, enabling high‐efficiency, compact footprint, and hardware‐encoded plasticity.
Current attempts in vaccine delivery systems concentrate on replicating the natural dissemination of live pathogens, but neglect that pathogens evolve to evade the immune system rather than to provoke it. In the case of enveloped RNA viruses, it is the natural dissemination of nucleocapsid protein (NP, core antigen) and surface antigen that delays NP exposure to immune surveillance. Here, we report a multi-layered aluminum hydroxide-stabilized emulsion (MASE) to dictate the delivery sequence of the antigens. In this manner, the receptor-binding domain (RBD, surface antigen) of the spike protein was trapped inside the nanocavity, while NP was absorbed on the outside of the droplets, enabling the burst release of NP before RBD. Compared with the natural packaging strategy, the inside-out strategy induced potent type I interferon-mediated innate immune responses and triggered an immune-potentiated environment in advance, which subsequently boosted CD40 ⁺ DC activations and the engagement of the lymph nodes. In both H1N1 influenza and SARS-CoV-2 vaccines, rMASE significantly increased antigen-specific antibody secretion, memory T cell engagement, and Th1-biased immune response, which diminished viral loads after lethal challenge. By simply reversing the delivery sequence of the surface antigen and core antigen, the inside-out strategy may offer major implications for enhanced vaccinations against the enveloped RNA virus.
The knowledge of osteoarthritis (OA) has nowadays been extended from a focalized cartilage disorder to a multifactorial disease. Although recent investigations have reported that infrapatellar fat pad (IPFP) can trigger inflammation in the knee joint, the mechanisms behind the role of IPFP on knee OA progression remain to be defined. Here, dysregulated osteopontin (OPN) and integrin β3 signaling are found in the OA specimens of both human and mice. It is further demonstrated that IPFP‐derived OPN participates in OA progression, including activated matrix metallopeptidase 9 in chondrocyte hypertrophy and integrin β3 in IPFP fibrosis. Motivated by these findings, an injectable nanogel is fabricated to provide sustained release of siRNA Cd61 (RGD−Nanogel/siRNA Cd61) that targets integrins. The RGD−Nanogel possesses excellent biocompatibility and desired targeting abilities both in vitro and in vivo. Local injection of RGD−Nanogel/siRNA Cd61 robustly alleviates the cartilage degeneration, suppresses the advancement of tidemark, and reduces the subchondral trabecular bone mass in OA mice. Taken together, this study provides an avenue for developing RGD−Nanogel/siRNA Cd61 therapy to mitigate OA progression via blocking OPN‐integrin β3 signaling in IPFP.
Institution pages aggregate content on ResearchGate related to an institution. The members listed on this page have self-identified as being affiliated with this institution. Publications listed on this page were identified by our algorithms as relating to this institution. This page was not created or approved by the institution. If you represent an institution and have questions about these pages or wish to report inaccurate content, you can contact us here.
Pokfulam Road, Hong Kong, Hong Kong
Head of institution
Prof. Peter Mathieson