Excessive Space Research Articles

A Clustering Scheduling Strategy for Space Debris Tracking

The number of spacecraft and space debris, particularly in low Earth orbit, is rapidly escalating, heightening the demand for rational and effective satellite allocation strategies to alleviate the strain on satellite monitoring and tracking systems. In this paper, a clustering scheduling strategy is proposed for satellites equipped with sensors tasked with monitoring a substantial amount of space debris. To address the challenge of excessive space debris, a dynamic and static clustering division scheme is introduced, leveraging target characteristics in conjunction with the nearest neighbor clustering algorithm. Furthermore, to enhance scheduling efficiency, four performance indicators are proposed to guide the generation of a resource allocation scheme using Genetic Algorithm and Simulated Annealing. Simulation results demonstrate that the clustering scheduling strategy effectively clusters a large number of space debris, while the scheduling scheme demonstrates better balance, comprehensiveness, and efficiency. The Genetic Algorithm and Simulated Annealing exhibit superior performance compared to their counterparts across various orbital inclinations and altitudes. Notably, in a specific simulation scenario related to constellations, the Genetic Algorithm and Simulated Annealing achieve a performance enhancement of 13.48% compared to the Genetic Algorithm and an improvement of 8.73% compared to Simulated Annealing.

Wind-Induced Vibration Control of High-Rise Buildings with Double-Skin Façades Using Distributed Multiple Tuned Façade-Dampers-Inerters

To address the shortcomings of tuned mass dampers (TMD), such as excessive internal space occupation and overlarge physical mass, this paper proposes a tuned façade damper inerter (TFDI) that utilizes parts of the outer façades of double-skin façades (DSF) as damping mass, capitalizing on the lightweight and efficient characteristics of inerters. The TFDI effectively resolves the challenge of multi-layer connections of inerters in high-rise buildings by utilizing corridor space. By vertically distributing TFDIs, a distributed multiple TFDI (d-MTFDI) system is formed. The configuration and motion of equations of this system are presented, and the control effectiveness is validated using wind tunnel test data. Two tuning modes are further proposed: unified tuning mode and distributed tuning mode. For the unified tuning mode, analytical expressions for optimal tuning frequency and damping ratio are derived; for the distributed tuning mode, numerical optimization methods are employed to determine the optimal tuning frequency range and damping ratio. Comparative results indicate that the distributed tuning mode achieves higher control efficiency than the unified tuning mode, with a significant reduction in the required optimal damping ratio. Furthermore, comparisons with d-MTMD demonstrate that d-MTFDI significantly enhances wind-induced vibration control performance.

Edge computing resource scheduling method based on container elastic scaling.

Edge computing is a crucial technology to solve the problem of computing resources and bandwidth required for extensive edge data processing, as well as for meeting the real-time demands of applications. Container virtualization technology has become the underlying technical basis for edge computing due to its efficient performance. Because the traditional container scaling strategy has issues such as long response times, low resource utilization, and unpredictable container application loads, this article proposes a method for scheduling edge computing resources based on the elastic scaling of containers. Firstly, a container load prediction model (Trend Enhanced-Temporal Convolutional Network, TE-TCN) is designed based on the temporal convolutional neural network, which features an encoder-decoder structure. The encoder extracts potential temporal relationship features from the historical data of the container load, while the decoder identifies the trend item of the container load through the trend enhancement module. Subsequently, the information extracted by the encoder and decoder is fed into the fully connected layer to facilitate container load prediction using the dual-input ResNet method. Secondly, Markov decision process (MDP) is used to model the elastic expansion problem of containers in multi-objective optimization. Utilizing the prediction outcomes of the TE-TCN load prediction model, a time-varying action space is formulated to address the issue of excessive action space in conventional reinforcement learning. Subsequently, a predictive container scaling strategy based on reinforcement learning is devised to align with the application load patterns in the container environment, enabling adaptation to the surge in traffic generated by the container environment. Finally, the experimental results on the WorldCup98 dataset and the real dataset show that the TE-TCN model can accurately predict the container load change. Experiments in the actual environment demonstrate that the proposed strategy reduces the average response time by 16.2% when the burst load arrives, and increases the average CPU utilization by 44.6% when the jitter load occurs.

Prosthodontic rehabilitation with all-on-four implant treatment combined CAD/CAM prosthesis in an oral cancer patient: a case report

BackgroundThe microvascular free fibula (MFF) flap is a reliable treatment modality for mandibular reconstruction and is suitable for dental implant placement after oncologic surgery. The most common issue with the MFF flap is its limited bone height, which typically results in excessive interarch space and complicates prosthodontic therapy. Overcoming the physical limitations from tumor excision and reducing the treatment time for prosthodontic rehabilitation to improve quality of life are critical clinical challenges.Case PresentationA 64-year-old male with lower left gum and bilateral buccal cancer received a single-layer microvascular MFF flap to reconstruct a mandibular defect post-tumor excision. He underwent a bilateral modiolus Z-plasty combined with a skin flap debulking procedure to relieve oral contracture, achieving adequate mouth opening for prosthodontic rehabilitation. Scar tissue bands on the bilateral cheeks significantly affected retention and stability, hampering dental impression performance. The patient sought prosthodontic rehabilitation to enhance his chewing function and quality of life promptly. Prosthodontic rehabilitation with all-on-4 implant therapy, utilizing computer-aided design and computer-assisted manufacturing (CAD/CAM), was completed within one month.ConclusionThis case utilized the all-on-4 implant system to address the insufficient fibular height for conventional dental implant placements. Dental CAD/CAM was employed to mill custom prosthetic abutments and a large titanium framework for the implant bar overdenture, compensating for the excessive interarch space between the grafted fibula and maxilla. This treatment approach successfully shortened the prosthodontic rehabilitation time and overcame anatomical limitations.

Growth characterization of mesophotic rhodoliths in the northern Gulf of Mexico using radiocarbon dating

Boxwork rhodoliths are characterized by a multitaxonomic composition and excessive void spaces, and they are distributed from intertidal to mesophotic depths. Growth rates and patterns in rhodoliths from shallower depths and simpler morphologies have been assessed using incremental and geochemical analyses, but rhodoliths from mesophotic depths have been harder to access and require more studies. Boxwork rhodoliths have proved too structurally complex for age and growth rate assessment techniques such as stable oxygen isotopes profiling or elemental mapping. We therefore utilized sequential radiocarbon analysis for growth characterization of mesophotic boxwork rhodoliths. Rhodoliths were collected from the Flower Garden Banks National Marine Sanctuary in the northern Gulf of Mexico from one location at 48–72 m and another at 103 m. Specimens were extensively sampled for radiocarbon concentrations from predicted nucleation points and along growth axes. From depths of 48–72 m, nucleation ages ranged from 795 ± 20 to 3270 ± 25 14C yr B·P. From depths of 103 m, nucleation ages ranged from 8960 ± 110 to 13,050 ± 150 14C yr B.P., dating back to the end of the Pleistocene in some cases. Several rhodoliths contained evidence of growth hiatuses, with deeper samples displaying age reversals that complicated the apparent growth history. We define age reversals to be non-sequential ages within a sampling sequence, typically a singular age before the return to normal age sequence. The research presented here includes the oldest known living rhodoliths, with the oldest rhodoliths grown before and after Holocene sea-level stabilization. These results have implications for the use of rhodoliths as paleoenvironmental indicators, and provide better understanding of the range and volume of carbonate production in these habitats.

Joint computation offloading and resource allocation for end-edge collaboration in internet of vehicles via multi-agent reinforcement learning

Vehicular edge computing (VEC), a promising paradigm for the development of emerging intelligent transportation systems, can provide lower service latency for vehicular applications. However, it is still a challenge to fulfill the requirements of such applications with stringent latency requirements in the VEC system with limited resources. In addition, existing methods focus on handling the offloading task in a certain time slot with statically allocated resources, but ignore the heterogeneous tasks’ different resource requirements, resulting in resource wastage. To solve the real-time task offloading and heterogeneous resource allocation problem in VEC system, we propose a decentralized solution based on the attention mechanism and recurrent neural networks (RNN) with a multi-agent distributed deep deterministic policy gradient (AR-MAD4PG). First, to address the partial observability of agents, we construct a shared agent graph and propose a periodic communication mechanism that enables edge nodes to aggregate information from other edge nodes. Second, to help agents better understand the current system state, we design an RNN-based feature extraction network to capture the historical state and resource allocation information of the VEC system. Thirdly, to tackle the challenges of excessive joint observation-action space and ineffective information interference, we adopt the multi-head attention mechanism to compress the dimension of the observation-action space of agents. Finally, we build a simulation model based on the actual vehicle trajectories, and the experimental results show that our proposed method outperforms the existing approaches.

Image Collage on Arbitrary Shape via Shape-Aware Slicing and Optimization.

Image collage is a very useful tool for visualizing an image collection. Most of the existing methods and commercial applications for generating image collages are designed on simple shapes, such as rectangular and circular layouts. This greatly limits the use of image collages in some artistic and creative settings. Although there are some methods that can generate irregularly-shaped image collages, they often suffer from severe image overlapping and excessive blank space. This prevents such methods from being effective information communication tools. In this article, we present a shape slicing algorithm and an optimization scheme that can create image collages of arbitrary shapes in an informative and visually pleasing manner given an input shape and an image collection. To overcome the challenge of irregular shapes, we propose a novel algorithm, called Shape-Aware Slicing, which partitions the input shape into cells based on medial axis and binary slicing tree. Shape-Aware Slicing,which is designed specifically for irregular shapes, takes human perception and shape structure into account to generate visually pleasing partitions. Then, the layout is optimized by analyzing input images with the goal of maximizing the total salient regions of the images. To evaluate our method, we conduct extensive experiments and compare our results against previous work. The evaluations show that our proposed algorithm can efficiently arrange image collections on irregular shapes and create visually superior results than prior work and existing commercial tools.

Designing movement space for elderly and disabled people in the construction law in selected countries.

Building law regulations determine designing the built environment recognising the needs of users of different ages and psychophysical abilities. Seniors and their spatial needs are covered there to a limited extent. The benchmark for design are wheelchair users. Their spatial requirements are greater in relation to independent walkers, including most older people. This makes it difficult to adapt the whole built environment to the needs of people with less mobility dysfunction. This can be considered in terms of spatial design and investment costs. The paper analyses the building regulations of 3 countries in terms of legal conditions to shape the architectural movement space of older people and disabled people. Analytical and comparative methods are used. Such research is becoming relevant and necessary. The analyses are conducted in the context of ageing populations. They are justified by statistical data on the age groups of Polish society. This is followed by design analyses of the legal requirements in the wheelchair movement space and proposals for alternatives, e.g., people walking with canes (case study). Their aim is to indicate methods to limit excessive communication spaces in buildings while maintaining functional values for all users. Research may show the possibility of greater diversification of regulations and alternatives to current laws. They are dedicated to participants in investment processes to shape accessible buildings. They can also be used in legislative work on amendments to the construction law. Changes in building regulations and a detailed approach to the mobility needs of older and disabled people (walking independently) are proposed. These decisions can provide benefits (spatial and economic savings). They fall into the "design for all" trend and sustainability of the built environment. These demands are based on no longer valid normative regulations. Med Pr Work Health Saf. 2024;75(3):189-197.

Sampling approaches to reduce very frequent seasonal time series

AbstractWith technological advancements, much data is being captured by sensors, smartphones, wearable devices, and so forth. These vast datasets are stored in data centres and utilized to forge data‐driven models for the condition monitoring of infrastructures and systems through future data mining tasks. However, these datasets often surpass the processing capabilities of traditional information systems and methodologies due to their significant size. Additionally, not all samples within these datasets contribute valuable information during the model training phase, leading to inefficiencies. The processing and training of Machine Learning algorithms become time‐consuming, and storing all the data demands excessive space, contributing to the Big Data challenge. In this paper, we propose two novel techniques to reduce large time‐series datasets into more compact versions without undermining the predictive performance of the resulting models. These methods also aim to decrease the time required for training the models and the storage space needed for the condensed datasets. We evaluated our techniques on five public datasets, employing three Machine Learning algorithms: Holt‐Winters, SARIMA, and LSTM. The outcomes indicate that for most of the datasets examined, our techniques maintain, and in several instances enhance, the forecasting accuracy of the models. Moreover, we significantly reduced the time required to train the Machine Learning algorithms employed.

Impact of H-Related Chemical Bonds on Physical Properties of SiNx:H Films Deposited via Plasma-Enhanced Chemical Vapor Deposition

SiNx:H film deposition via plasma-enhanced chemical vapor deposition has been widely used in semiconductor devices. However, the relationship between the chemical bonds and the physical and chemical properties has rarely been studied for films deposited using tools in terms of the actual volume production. In this study, we investigated the effects of the deposition conditions on the H-related chemical bonding, physical and chemical properties, yield, and quality of SiNx:H films used as passivation layers at the 28 nm technology node. The radiofrequency (RF) power, electrode plate spacing, temperature, chamber pressure, and SiH4:NH3 gas flow ratio were selected as the deposition parameters. The results show a clear relationship between the H-related chemical bonds and the examined film properties. The difference in the refractive index (RI) and breakdown field (EB) of the SiNx:H films is mainly attributed to the change in the Si–H:N–H ratio. As the Si–H:N–H ratio increased, the RI and EB showed linear growth and exponential downward trends, respectively. In addition, compared with the Si–H:N–H ratio, the total Si–H and N–H contents had a greater impact on the wet etching rates of the SiNx:H films, but the stress was not entirely dependent on the total Si–H and N–H contents. Notably, excessive electrode plate spacing can lead to a significant undesired increase in the non-uniformity and surface roughness of SiNx:H films. This study provides industry-level processing guidance for the development of advanced silicon nitride film deposition technology.

Influence of printing parameters on the thermal properties of 3D-printed construction structures

3D printing in the construction sector is a promising, sustainable approach with the potential to build energy-efficient buildings. However, there is a lack of studies on the thermal properties of 3D-printed structures, which are fundamental to their energy efficiency. This study investigates the impact of printing parameters on thermal conductivity and defect development in 3D-printed concrete structures. Heat flow meter, scanning electron microscope and infrared imaging techniques were employed to assess the thermal properties and microstructure of 3D-printed samples. The results demonstrate that printing parameters significantly influence the porosity, void formation, and microstructural characteristics of 3D-printed concrete, ultimately affecting thermal properties. An increase in extrusion rate leads to higher thermal conductivity, whereas an increase in speed and standoff distance reduces thermal conductivity. Excessive spacing between printed lines can induce voids and gaps, contributing to lower thermal conductivity, while minimal spacing can promote higher porosity and larger pores. Infrared imaging revealed nonuniform thermal distribution in 3D-printed structures caused by the layered structure and voids in interlayer gaps. This study provides valuable insights for designing and constructing energy-efficient 3D-printed buildings, contributing to the development of sustainable building practices and the advancement of additive manufacturing in construction.

Multi-agent deep reinforcement learning for hyperspectral band selection with hybrid teacher guide

Due to the presence of noisy and highly redundant bands in hyperspectral images (HSIs), band selection serves as a key preprocessing for downstream classification tasks. Recently, deep reinforcement learning (DRL) has been developed as a new trend for band selection of HSIs. Existing DRL-based methods often adopt single-agent, which are prone to fall into local optima due to an excessive action space. The multi-agent methods provide a feasible solution, but often require too much computation. To address these problems, a novel multi-agent DRL method with hybrid teacher guide (MH-DRL) is proposed for band selection of HSIs. In MH-DRL, each agent corresponding to a spectral band decides whether this band is selected. Moreover, a presentation-evaluation network (PE-Net) is constructed to design the reward by evaluating the candidate band subsets without any fine-tuning and represent the state by extracting the spatial-spectral features of HSIs. Then, three kinds of experienced band selection models are regarded as the teachers and designed to participate in the band exploration of DRL, which can improve the learning effectiveness and efficiency by accumulating the external knowledge from diverse teacher models. Finally, deep Q-learning algorithm is designed to update the agents and improve their self-learning ability from continuous exploration. Experimental results on three widely-used HSI data verify the performance of the proposed method better than some advanced band selection algorithms of HSIs. Our source code is available at https://github.com/jiefeng0109/MH-DRL.

Surface temperature and power generation efficiency of PV arrays with various row spacings: A full-scale outdoor experimental study

Quantifying the relationship between surface temperature and power generation efficiency of solar photovoltaics (PV) is critical to their practical implementation. Although empirical models have been developed on this, they were mainly based on indoor laboratory tests, ignoring a practically significant arrangement factor. Therefore, a combined outdoor experimental and empirical study on PV array systems considering various row spacings was undertaken through this study. Experimental results indicated that solar irradiance shows a relatively more prominent effect than row spacing and wind speed in outdoor environments. During the practical arrangement, it is only necessary to ensure that there is no shadow between the adjacent rows of PV arrays in practical applications, and the possible cooling effect brought by excessive row spacing may not be pursued. However, this does not mean that row spacing can be ignored when predicting surface temperature and power generation efficiency. Based on the data from our long-term experimental tests, empirical models to predict solar PV's surface temperature and power generation efficiency were developed, considering various row spacings. Comparison showed that the models developed in this study can reduce the mean relative error of the previous model's predictions from 20% to 5%, which proves the necessity of including row spacing in the models. This study's developed models and research outcomes pave the way for future large-scale practical analysis of solar PV arrays.

Investigation into designing VLSI of a flexible architecture for a deep neural network accelerator

Deep learning, a branch of AI, makes use of specialised neural networks. Computational acceleration of high-performance deep neural network algorithms continues to necessitate efficient architecture, high memory bandwidth, parallel processing, and resources, despite decades of study on such algorithms. Excessive space requirements are a problem for DNN implementations caused by resource-intensive components like activation functions (AF) and multiply-and-accumulate (MAC) units. In addition, edge-AI applications necessitate a densely packed, power-hungry, high-throughput DNN accelerator. If we want to build a DNN accelerator that uses little power and occupies little space, we need to optimise the MAC architecture, the AF, and the network complexity so that data flows efficiently. In addition, providing functional configurability while working with restricted chip surface is a difficulty for DNN hardware designs based on ASICs. This dissertation explores the efficient and low-power VLSI architecture of DNN accelerators, addressing the hardware implementation of DNN and targeting applications with limited resources. In order to assess MAC and non-linear AF operations, we investigate and enhance the CORDIC architecture. The poor throughput is a major downside of CORDIC-based architectures, even though they are area and power efficient. Consequently, we suggest a pipelined design for CORDIC-based MAC and AF that focuses on performance. Due to the increased hardware resource consumption that comes with pipeline stages, this study investigates the mutual exclusivity of CORDIC stages and investigates in depth the accuracy variation related to the number of stages needed to achieve high throughput.

Frequency tunable magnetostatic wave filters with zero static power magnetic biasing circuitry

A single tunable filter simplifies complexity, reduces insertion loss, and minimizes size compared to frequency switchable filter banks commonly used for radio frequency (RF) band selection. Magnetostatic wave (MSW) filters stand out for their wide, continuous frequency tuning and high-quality factor. However, MSW filters employing electromagnets for tuning consume excessive power and space, unsuitable for consumer wireless applications. Here, we demonstrate miniature and high selectivity MSW tunable filters with zero static power consumption, occupying less than 2 cc. The center frequency is continuously tunable from 3.4 GHz to 11.1 GHz via current pulses of sub-millisecond duration applied to a small and nonvolatile magnetic bias assembly. This assembly is limited in the area over which it can achieve a large and uniform magnetic field, necessitating filters realized from small resonant cavities micromachined in thin films of Yttrium Iron Garnet. Filter insertion loss of 3.2 dB to 5.1 dB and out-of-band third order input intercept point greater than 41 dBm are achieved. The filter’s broad frequency range, compact size, low insertion loss, high out-of-band linearity, and zero static power consumption are essential for protecting RF transceivers from interference, thus facilitating their use in mobile applications like IoT and 6 G networks.

Enhancing Aesthetic and Structural Integrity in Necrotic Permanent Incisor with Open Apex: A Case Report

Traumatic injuries in immature teeth often lead to complications in endodontic treatment due to difficulty in managing the open apex. Also, the restoration of these teeth poses significant challenges due to thin dentin walls, high configuration factors, and other biomechanical and adhesive issues. Standard fibre posts are often not suited for flared root canal preparations, leaving excessive cement space between the post and the tooth structure. To reduce postoperative complications (e.g. root fracture), reinforcement of radicular dentin becomes mandatory after endodontic treatment. The reinforcement of the canal can be done using a fibre post relined with composite resin. Direct anatomical posts offer a promising alternative that minimises the risk of debonding. This case report details the use of platelet rich fibrin as a internal matrix and how mineral trioxide aggregate can be used to form an immediate apical seal for apexification and a custom fiber post technique for reinforcing an immature maxillary central incisor in an 18-year-old patient. Keywords - Adhesive cement, Apical seal, Dentin reinforcement, Esthetics, Fiberglass, MTA, Post and core technique

Evaluating the impact of urban morphology on urban vitality: an exploratory study using big geo-data

ABSTRACT The United Nations has proposed Sustainable Development Goals (SDGs), of which SDG11 aims to “make cities and human settlements inclusive, safe, resilient and sustainable”. This is in line with Urban Vitality's objectives. This study proposes a quantitative framework to evaluate the impact of urban morphology on urban vitality. In this framework, a proxy that is more reflective of economic and human activities is proposed for depicting urban vitality based on geographic big data; then we use a Multi–scale Geographically Weighted Regression (MGWR) regression considering spatial heterogeneity to analyze the different effects of urban form on urban vitality. Taking the Beijing Sixth Ring Road area as the study region, the results indicate that tall, large–area, multi–functional buildings have a significantly positive impact on urban vitality. These areas exhibit single–function buildings and high levels of greenery, leading to inefficient space utilization. Increasing road network density, as well as the density of transportation and public facilities, also positively influences urban vitality. However, commercial density has a negative impact on workday vitality. Furthermore, excessive green space density, a high proportion of the largest green patches, and complex green space borders all contribute to a reduction in urban vitality.

Research on the Optimization of a Diesel Engine Intercooler Structure Based on Numerical Simulation

As a device for cooling charged air before it enters the cylinder, the intercooler is an indispensable part of the regular operation of a booster diesel engine. To solve the problem of the insufficient cooling performance of an intercooler for a high-power supercharged diesel engine, in this study, the flow field in the intercooler is simulated using the computational fluid dynamics (CFD) model of porous media, and the performance data measured using the steady flow test bench are used to provide boundary conditions for the calculation. The effects of the charged air mass flow rate and the tube bundle’s transverse spacing on the heat dissipation performance of the intercooler are analyzed and compared. The calculation results show that, under the condition of satisfying the regular operation of the diesel engine, the heat transfer coefficient of the intercooler heat dissipation belt increases with the increase in air mass flow and the spacing of cooling pipes, and the heat transfer coefficient can be increased by up to 57%. Still, excessive spacing of the cooling water pipes increases pressure loss in the charged air. Finally, the transverse spacing of the tube bundle is set to 17 mm, ensuring the pressure drop in the charged air, and the heat dissipation performance of the intercooler is increased by 6.04%. This paper provides a feasible solution for further optimizing the heat dissipation performance of intercoolers. Finally, grey correlation theory is used to study the correlation between air mass flow, cooling water pipe spacing, and intercooler heat dissipation performance. The correlation values are 0.8464 and 0.8497, respectively, indicating a significant relationship between air mass flow, cooling water pipe spacing, and intercooler heat dissipation performance.

A Multi Tuning-Frequency Passive/On-demand Active Pendulum Tuned Mass Damper

The use of tuned mass dampers (TMDs) in high-rise buildings, towers, and other tall structures for mitigating wind-induced vibration has resulted in significant improvements in serviceability of such structures. In applications where more than one mode of vibration with distinctly different natural frequencies are in need of damping, multiple passive TMDs each tuned to one of the natural frequencies are needed. In many such applications, the excessive cost, weight, and space requirement of multiple passive TMDs are not acceptable. This paper presents a novel passive pendulum tuned mass damper (PTMD) which on-demand can switch to an active PTMD. In its default passive state, the proposed device is tuned to the first mode of the structure and acts as a traditional passive PTMD. In its active state, while staying tuned to the first mode passively, the PTMD simultaneously tunes itself to multiple higher order modes adding tuned damping and providing a multi-directional vibration control.

Selective ensemble method for anomaly detection based on parallel learning

Anomaly detection is a highly important task in the field of data analysis. Traditional anomaly detection approaches often strongly depend on data size, structure and features, while introducing the idea of ensemble into anomaly detection can greatly improve the generalization ability. Ensemble-based anomaly detection methods still face some challenges, however, such as data imbalance, time and space demand and the selection of base detectors. To this end, we propose a selective ensemble method for anomaly detection based on parallel learning (SEAD-PL). First, a differentiated stratified sampling method is designed to alleviate the problem of data imbalance. Then, a distributed parallel training frame is built to address the problem of excessive time and space consumption for base detector training. Finally, a clustering-based ensemble selection strategy is introduced to balance the accuracy and diversity of base detectors. Experiments are performed on six datasets, which demonstrate that the proposed method has obvious advantages over four selected methods.