Search

Development of urban resilience is largely determined by the planning and design strategies proposed for managing the risk of vulnerable places exposed to different hazards. These places are commonly occupied by poor dwellers or known as deprived areas. However, efficient planning and design strategies are difficult to identify and evaluate even the exposure of urban deprivation is known. It is mainly because the knowledge about the exposure of vulnerable urban elements in deprived areas is not explicitly interpretable in the context of planning and design practice. This work is set upon the rationale that environmental and ecological patterns within urban areas can be interpreted by the morphology of cities, so that the exposure of different urban areas including those deprived can be interpreted by explicit measurements of urban forms, which can potentially inform planning and design practitioners. A workflow is proposed and exemplified for the entire city of Nairobi, Kenya, where several technical steps are involved including mapping and measuring the morphology of basic urban elements such as buildings, relating the measurements to potential hazards, and finally obtain an explicit characterization of multi-hazards exposure of urban deprivation.

Graph Neural Networks (GNNs) has been widely used to address the sparsity and cold start problems in recommendation system. By propagating embeddings from multi-hop neighbor nodes among the interaction graph and update target user and item embeddings, GNNs-based methods can achieve better recommendation performance. But those methods directly concatenate the output of each layer and ignore the different influences between different layers, and they simply use the inner product of the user and item’s embeddings to calculate the similarity and make recommendation based on it, which is insufficient to reveal the complex and nonlinear interactions.In this work, we propose to learn multi-order interactions between users and items and capture correlations between different-order information. We design a new recommendation framework MCCR, which treats each layer’s output as differentorder feature, and propose a multi-order interaction module to represent feature interactions. We adopt a multi-layer 3D CNN module to learn high-order interaction signals between users and items in an explicit approach. Through extensive experiments on three real-world datasets, which shows that MCCR evidently outperforms the state-of-the-art methods consistently.

Obtaining demographic attributes of online users is of great significance for retail marketing, targeted advertisement and many other scenarios. Users' wanderings on various websites and applications contains user preference on different items, and can be leveraged to infer one's private attributes. Existing studies usually focus on manually defined features, relationships in online social networks, or modeling global user preferences. However, attribute inference from the most common behavioral data (e.g., browsing history, shopping cart) is recently overlooked, and still requires further research. In this work, we propose a Topic-Enhanced Attentive Recurrent Neural Network (TEA-RNN) model to capture both local neighborhood-based features (with attentive RNN) and global patterns (with topic model) within user behaviors, and apply multi-task learning mechanism with weighted losses to further leverage the latent relationships within demographics. Experimental results on real-world datasets demonstrates the effectiveness of TEA-RNN by comparing with several commonly used baselines.

In order to study the dynamic characteristic of an electrothermal actuator, a high-speed camera-based optical measurement system is first established. Then, a combination of the background differencing and canny edge detection algorithm is utilized to deal with the images from the measurement system above. Hence, the picture that represents the starting point of the motion of the actuator is distinguished between these pictures and the transient displacement can also be calculated. The time that the actuator reach steady state is no more than 100 ms when a constant voltage is applied on it. Finally, the dynamic behavior of the actuator is further studied under changeable voltage. For the sine voltage, an obvious hysteresis appears at the beginning and gradually vanishes with the time growing. When applied to a periodic voltage such as square or sine, the motion of the actuator is also gradually cyclical and has the same time period.

Predicting users' preferences based on historical interactions plays a critical role in recommender systems. Conventionally, users' recent sequence behaviors are fed into a model to predict the next item that a user is most interested in. However, most of the existing methods simply utilize the user's recent behavior sequences to model short-term preferences. This is problematic, because users may pay more attention to partial features in the sequences due to their long-term preferences. To this end, we propose a user Preference Excitation network to more accurately model user sequence information. Specifically, a personalized gating layer is utilized to model the user's specific preferences and a preference excitation self-attention layer learns the complicated sequential patterns. Furthermore, the gate mechanism is used to adaptively fuse the information from these two aspects. Empirically, we evaluate our model on several real-world datasets and the comprehensive experimental results show that our model consistently outperforms the selective state-of-the-art methods. Source code of the paper is available at https://github.com/strange-jiong/PERec.

Traditional attributed network embedding methods are designed to map structural and attribute information of networks jointly into a continuous Euclidean space, while recently a novel branch of them named binarized attributed network embedding has emerged to learn binary codes in Hamming space, aiming to save time and memory costs and to naturally fit node retrieval task. However, current binarized attributed network embedding methods are scarce and mostly ignore the local attribute similarity between each pair of nodes. Besides, none of them attempt to control the independency of each dimension(bit) of the learned binary representation vectors. As existing methods still need improving, we propose an unsupervised Neural-based Binarized Attributed Network Embedding (NBANE) approach. Firstly, we inherit the Weisfeiler-Lehman proximity matrix from predecessors to aggregate high-order features for each node. Secondly, we feed the aggregated features into an autoencoder with the attribute similarity penalizing term and the orthogonality term to make further dimension reduction. To solve the problem of integer optimization we adopt the relaxation-quantization method during the process of training neural networks. Empirically, we evaluate the performance of NBANE through node classification and clustering tasks on three real-world datasets and study a case on fast retrieval in academic networks. Our method achieves better performance over state- of-the-art baselines methods of various types.

Recently, network embedding has attracted a surge of attention due to its ability to automatically extract features from graph-structured data. Though network embedding method has been intensively studied, most of the existing approaches pay attention to either structures or attributes. In this paper, we propose a novel attributed network embedding method based on a Siamese neural network, named SANE, to capture both the network structure and node attribute information in a principled way. Specifically, to preserve local semantic proximity, we adopt a Siamese neural network, which can directly learn the similarity of paired nodes with their attributes as input. Then, a skip-gram module is connected with the final shared hidden layer to capture high-order proximity based on the latent representation of node attributes. Thus, we can learn the complex interrelations between nodes. Empirically, we evaluate our model on several real-world datasets and the experimental results have verified the effectiveness of our proposed approach.

Dynamic graph link prediction has attracted increasing attention in various fields such as social networks, paper citation networks and knowledge graphs. Many models have been developed to predict the future graph structure. In this paper, we propose a link prediction model with semantic evolution (LISE), to predict links in a sequence of graph over time. Our approach is based on the discovery of non-random initialization dynamic word embedding which is a kind of method to study semantic evolution. It can help us train node embedding in the same space and introduce temporal context into the embedding training of nodes. Based on node embedding in the same space, LISE can unify historical behavior, graph snapshots structure information and dynamic attributes into a frame. We evaluate our proposed method and various comparing methods on two real-world datasets. The experimental results prove the effectiveness of the link prediction made by LISE model.

Attributed network embedding, which aims to map the structural and attribute information into a latent vector space jointly, has attracted a surge of research attention in recent years. However, a vast majority of existing work explores the correlation between node structure and attribute values whereas the attribute type information which can be potentially complementary is ignored. How to effectively model the nodes, attribute types and attribute values as well as their relations in a unified framework is an open yet challenging problem. To this end, we propose a translation-based attributed network embedding method named TransANE. In our approach, the whole attributed network is considered as a coupled network which consists of two components, i.e., node relation network and attribute correlation network. We construct attribute correlation network by the co-occurrence of attribute values. Each node-attribute relation is regarded as an attributional triple, e.g., (Tom, Gender, Male). We introduce knowledge representation method to model the mapping between nodes, attribute types and attribute values. Empirically, experiments on two real-world datasets including node multi-class classification and network visualization are conducted to evaluate the effectiveness of our method TransANE in this paper. Our method achieves significant performance compared with state-of-the-art baselines.

In this paper, we studied the multiple soft breakdown phenomena of gate oxide during TDDB (Time Dependent Dielectric Breakdown). After the soft breakdown, the parameters of the device had been greatly degraded (as a result, the device lost its original functions), although the multiple soft breakdown did not lead the oxide catastrophical failure. Further analysis also verified that the gate oxide had been damaged after the first soft breakdown. We found that there is a very good match between the times of soft breakdown and the failure points fixed using EMMI-OBIRCH system. Moreover, such correspondence was verified using the in-situ failure analysis method. Finally, we proposed an explanation concerning the oxide breakdown based on the phenomenon.

Travelling by taxi is more convenient and effective. With an overcrowding population and a much terrible traffic, the traditional way of hailing a taxi encounters many challenges like where to pick-up/drop-off passengers reasonably and where to find potential passengers quickly. More cities have established taxi stands to advocate and to guide passengers to hail a taxi. However, most of them have low rate of usage. The reason lies in that to determine where to establish reasonably is a big problem. In this paper, we are the first to propose a DFA to identify data signifying pick-up/drop-off events. We propose a DBH-CLUS method to identify pick-up/drop-off hotspots. The method applies hierarchal clustering based on agglomerative clustering analysis method. We have conducted three experiments to verify the DFA, to analyze the region agglomeration and to analyze the accuracy. The experimental results manifest that our method can precisely identify hotspots from the original GPS data and provide an excellent tool to facilitate taxi stand planning.

In single-photon emission computed tomography (SPECT) imaging, the attenuation map can be estimated from cone-beam computed tomography (CBCT) images for attenuation correction. However, radiation dose involved in CBCT scans constitutes a potential health concern to the imaged subject. An approach to reducing CBCT-imaging dose is to decrease the number of views at which projections are acquired. Analytic-based algorithms may not be suitable for this type of data because they generally require data collected at a large number of densely distributed views. On the other hand, optimization-based reconstruction algorithms based on a discrete linear model have been demonstrated to be capable of reconstructing accurate images from projections fewer than those required by analytic-based algorithms. In the work, we investigate low-dose CT imaging in SPECT/CBCT attenuation correction by lowering number of projection views. The results demonstrated that CT-imaging dose in SPECT/CBCT may be reduced significantly by use of optimization-based reconstruction algorithms.

The prediction of water influx intensity is much important to the development of aquifer gas reservoir. In this paper, according to water activity and categorizing the type of gas reservoirs with reservoir types, we can use geologic and producing test information to study on the Prediction methods of the water influx intensity. Further confirmed the results in the producing test; through the producing test information we can use a variety of methods to find the display about the edge and bottom water activity, and thus deduce the water influx intensity, provide the basis for gas reservoir development programming.

In recent years, combined SPECT/CBCT systems have been developed and become commercially available for yielding spatially registered SPECT and CT images. The CBCT unit of the Philips SPECT/CBCT system (BrightView XCT) adopts a flat-panel detector for detecting X-rays. Because of heart motion and relatively slow data acquisition speed of the CBCT unit, the reconstruction of a beating heart is challenging for CBCT. Gating is often used in cardiac imaging for acquiring CBCT projection data of the heart. In order to acquire projection data of a certain phase over 360 degree, multiple turns of projection data of the heart have to be acquired, after which the data are sorted to form subsets for different phases. For cardiac data within one source rotation, projection data for each phase are sparsely and non-uniformly distributed over the scanning angular range. In this work, we investigate image reconstruction for different heart phases from both simulation and patient cardiac data sets within one source rotation. The reconstruction task is formulated as a constrained minimization of image total variation (TV) problem, and an ASD-POCS algorithm is used for solving the constrained TV-minimization problem. The approach proposed to image reconstruction can potentially reduce scanning time and imaging dose to the patient. More important for cardiac imaging, this approach may reduce artifacts caused by heart motion. For the cases studied, our results suggest that images of potential practical utility can be reconstructed from both simulation and patient data within a single phase.

This paper presents a new approach called parallel inductor synchronized switch harvesting to harvest electric energy from a mechanical excited piezoelectric element. Based on the criteria energy harvesting circuit, this new parallel inductor synchronized switch harvesting circuit was constructed and explained. The theoretical model of output electric energy for this new circuit was derived and a MULTISIM simulation study based on ANSYS piezoelectric coupling analysis was conducted to evaluate the ability of harvesting electric energy. The simulation results show that use of the parallel inductor synchronized switch energy harvesting circuit increases power transfer by over 400% as compared to when the criteria energy harvesting circuit is used.

Advanced single-photon emission computed tomography (SPECT) includes a cone-beam computed tomography (CBCT) unit for acquisition of anatomic information about the imaged subject. CT images obtained can be used for attenuation correction in SPECT and localization determination. Hardware considerations have led to the use of an imaging configuration in which the CBCT field of view (FOV) is extended through the use of an offset detector. In general, current clinical CBCT applications involve data acquired at a large number of views. Therefore, the accumulated radiation dose integrated over a large number of views can be of a health concern. In this work, we report our preliminary investigation of image reconstruction by using the recently developed adaptive steepest descent-projection onto convex sets (ASD-POCS) algorithm from real patient data collected at a number of views much fewer than what are used currently in CBCT imaging in clinical SPECT/CT studies.

In this paper, using Defrise Phantoms, we present our investigations of the axial limitations of an analytical cone-beam reconstruction algorithm (FDK), and the iterative ordered-subset transmission reconstruction (OSTR) iterative algorithm, for an axially extended version of the Philips Brightview XCT cone beam CT (CBCT) geometry. Simulations were preformed for head size and body size Defrise Phantoms of different axial dimensions to investigate limitations on axial extent as a function of these. OSTR yielded overall better axial performance than FDK. It may be possible to reconstruct a larger axial extent for brain studies; however, patients with larger body sizes pose more of a problem.

With the development of science and technology, its performance of delay arming device has been continually improved. However, bore premature, premature detonation, failure to fire and other grave accidents have still often happened in recent years. The reason is that delay arming device prematurely rescinds fuse safety mechanism or its fail of delaying arming mechanism. It has been widely focused on how to improve its safety and reliability of delay arming device at home and abroad, but there have only been few researches on its self-regulation of delay arming time. Based on its principle of MR fluid damp controlling technology and its yield stress of MR fluid will be changed with its changes of magnetic field strength, design and simulation of magnetic circuit for MR fluid fuze has been done in this paper. And this model among input current of coil, discharge aperture and time of discharge has been obtained by test and some useful results are made by numerical simulation. The experimental results prove that its delay arming time of MR fuze meets the design requirements and indicate that input current of coil has an obvious effect on discharge time.

Virtual testing (VT) technology has been widely applied into current mechanical engineering field, it has many advantages, including comparative low requirement of hardware, simplification experiment process, testing cost reduction, etc. As an important subsystem of ammunition, fuze is the key element for its efficiency performance. It would shorten the design and test cycle, ameliorate the traditional methods if VT technology was brought into fuze design, which also has both significant military and economic benefits. Based on the actual product investigation, the precisely model of fuze S&A mechanism can be automatically generated, and its kinematics simulation can also be finished by using CAD, CAE and UG customization technology. At the same time, a VT platform for fuze S&A mechanism is constructed in this paper.

Wireless Sensor Networks (WSNs) consist of spatially-distributed autonomous sensors that can cooperatively monitor physical and environmental conditions. Because of sensors' resource-constraints in terms of size, power, and bandwidth, one of the fundamental objectives in WSNs is improving energy-efficiency. In this paper, utilizing sensors with variable sensing radii, we propose a group-based technique to obtain energy-efficient k-coverage based on our previous work with the Delaunay-Triangulation-based 1-coverage algorithm. Our sensing-radii optimization technique ensures full coverage and attains nearly-optimal energy consumption in sensing. Furthermore, our ns-2 simulations confirm that the group-based k- coverage reduces sensing energy consumption and maintains a sound coverage ratio for reliable surveillance.

In this paper we have modeled the spatially-varying resolution of a fan-beam collimator as a function of both distance and lateral position across the imaging plane within iterative reconstruction. We demonstrate that an incremental blurring technique can be used to more efficiently model such non-stationary resolution than the straightforward non-stationary convolution methodology with a gain of 1.5X in speed and 99% less memory usage. This method can therefore be used for efficient variable resolution compensation in 3D for collimator geometries in which the detector response function varies across the imaging plane as well as with distance from the face of the collimator†.

This research was developed and aimed at an embedded controller to achieve an active variable damping control to magnetorheological fluid damper under impact load. Its character of impact load is firstly analyzed and its working principle of MR damper applied into anti-shock engineering design is also introduced. Then, an impact test has been done by using the developed test rig in order to investigate response time of MR dampers under impact load. Based on the measuring results, some factors effecting time delay of MR damper under impact load are analyzed. And simulation results prove that current response time can be gradually reduced with its resistance increase and can be gradually increased with its equivalent inductance increase, yet its frequency of PWM wave and temperature has little influence on it. Considering its strong nonlinear and uncertainty of a MR damper under impact load, and a large of signal processing and controlling operation, in order to improve its real-time, reliability and stability of control system, an embedded measuring-controlling system based on ARM and DSP processor is designed and developed for its engineering application of semi-active vibration control system. The embedded system design scheme, its each part design and the system software design are all introduced. Finally, experimental results show that the developed embedded system can effectively reduce time delay of MR damper and achieve a good semi-active vibration control effect.

This research is developed and aims at a new microactuator, which is used in the fuze Safety and Arming device (S&A) with a planar structure fabricated by the MEMS technology. In order to satisfy its engineering application requirements of the fuze service, which are small volume with S&A barrier dimension of 25mm×25mm×2mm, large force, long stroke as well as non-contactable, and a complex fuze system environment, a disk-type electromagnet microactuator prototype was designed and developed. Then, some main parameters of the designed electromagnetism microactuator had been estimated and optimized. Numerical simulation for three typical of microactuator had been also done, including their outputting force and generating heats, and their anti-interference performances in magnetic field. In order to test their movements, driving forces and performances of the developed microactuators, a suit of testing platform was designed and developed. The experimental results showed their overall performances of the developed microactuators under operating conditions could meet the engineering application requirements of a fuze system.

Grassland is an essential part of terrestrial ecosystems. It has a significant impact on the carbon cycle, as well as on climate and on regional economies. Till now, vegetation indices are the most popular remote sensed detecting method of grassland degradation. Although vegetation indices are useful for estimating the biomass, but detecting changes of vegetation indices are not always effective, as grassland vegetation with different characteristics may still produce similar vegetation index values. The development of hyperspectral sensors provides a new approach to solve this problem. The Hulunbeier grassland was chosen as a study object. Reflectance spectra of leaves and pure canopies of some dominant grassland species, as well as reflectance spectra of mixed grass community were measured. Using spectral feature parameterization methods such as spectral slope, spectral derivative, spectral integration, and spectral index, the spectral feather of leaves and pure canopies had been extracted. So the typical grassland vegetation species can be distinguished. Then the spectra of mixed grass community were unmixed using linear mixing models, and the proportion of all the components had been calculated. The field validation proved spectral feature parameterization and pixel unmixing methods in this research are effective.

In recoil mechanisms applications, the adjustability of Magnetorheological (MR) damper is desired to be used in closed loop control system to provide the optimal damping force and the recoil dynamics can be controlled. In this paper, the damping control system of MR dampers subject to impact load is studied. Since the recoil cycle is very fast, about only several hundreds milliseconds, the response time of MR damping force is an important characteristic for MR dampers’ recoil applications. First, the special designed MR damping force response time was tested and a simplified nonlinear model of MR damping force control system was established. Second, since the response time tested above can not mesh with the application demands, PI control and adaptive algorithms were used to improve the response time and the tracking ability of the MR damping force by numerical simulation analysis. Finally, impact experiment validation was done. The conclusion is achieved that, it’s indeed possible that with good control method the adjustability of MR dampers is able to be used in a closed-loop system to avoid the large peak of the MR damping force, and the optimal recoil dynamics is able to be achieved.

Wireless sensor networks (WSNs) consist of spatially-distributed autonomous sensors that can cooperatively monitor physical and environmental conditions. Because of sensors' resource-constraints in terms of size, power, and bandwidth, one of the fundamental objectives in WSNs is improving energy-efficiency. In this paper, we propose a mesh-based technique to obtain energy-efficiency in sensing and improve the quality of coverage. Our technique dynamically schedules a set of active sensors based on a square mesh or an equilateral-triangular mesh. The purpose of our technique is to quickly build an energy-efficient mesh that is self-adaptive to the local topology and allows energy balancing. Furthermore, a hole detection and recovery mechanism is discussed to guarantee full coverage. The simulation results confirm that our mesh-based coverage technique reduces sensing energy consumption and maintains a sound coverage ratio for reliable surveillance.

Wireless sensor networks (WSNs) are generally used for harsh environments involving military surveillance, emergency response, and habitat monitoring. Due to severe resource constraints in sensor nodes, including memory space, energy storage, and communication bandwidth, a need arises for an in-network aggregation of sensory data. We propose a sensor- to-sink transport protocol, which is suitable for data aggregation and provides reliable upstream packet delivery by dynamically configuring inactive nodes as "monitors" to assist in quick loss detection and recovery. To ensure energy efficiency, a minimum set of monitors is preferred; however, the problem of finding a minimum set of monitors is NP-complete. Thus, we propose a distributed greedy heuristic to solve this problem efficiently. Our ns-2-based simulations show significant performance improvements over other transport schemes in terms of throughput and data delivery rate under scenarios with intermittent traffic load and unpredictable node failures.

In recent years, Mobile Ad-hoc Networks (MANETs) have received tremendous attention because of their self-configuration and self-maintenance capabilities. With open network architecture, frequent changes in topology, and shared wireless media, problems arise in the MANET routing protocol design. Such problems can include Broadcast Storm, stale route, faulty nodes, and latency. This paper proposes a new MANET routing protocol called Density-first Ad-hoc Routing Protocol (DARP). This new protocol considers node density, node mobility, and route length in order to select new routes with longer lifetimes and better throughput. Also, it uses a local self-recovery technique during the route maintenance phase in order to avoid the overhead and delay often caused by re-initiating route discovery. Our ns-2-based simulation shows that DARPpsilas performance is improved significantly in terms of throughput and routing overhead. Furthermore, DARP shows better scalability than other routing protocols in more demanding environments with high mobility and heavy traffic loads.

In a Mobile Ad hoc NETwork (MANET), mobile devices cooperate to forward packets for each other without the assistance of dedicated routing infrastructures. Due to its networking flexibility and node mobility, MANET routing presents a new set of nontrivial challenges such as broadcast overhead, frequent topology changes, and slow convergence. Additionally, many proposed MANET routing protocols assume a friendly and cooperative environment, and hence are vulnerable to various attacks by misbehaving nodes. Recently, significant attention has been devoted to developing the fault resilient routing protocol for MANET, which requires routing stability and scalability in hostile and demanding environments. In this paper, we propose a fault resilient ad hoc routing protocol called local self-recovery routing (LSR), that is geared to adapt to topology dynamics and selfish behaviors in the shortest possible time. During the route recovery, the shortest detour is found to repair the faulty links locally using a cone-shaped recovery zone, and the broadcast area is efficiently chosen to minimize overhead while achieving a good recovery rate. Our ns-2 based simulation shows significant performance enhancements of LSR in throughput and routing overhead. Also, based on the simulation, the optimal angle of a cone-shaped recovery zone is studied under various scenarios.

Wireless sensor networks (WSN) consist of spatially distributed autonomous devices using sensors to cooperatively monitor physical or environmental conditions. Generally, sensor nodes are resource-constraint in terms of size, power and bandwidth, therefore energy-efficient design becomes one of the fundamental goals in WSN. To provide reliable service in surveillance as well as energy efficiency, significant attention has recently been devoted to dynamic coverage algorithms based on sensors which are capable of adjusting its transmission and sensing range. In this paper, we propose two local sensing radii optimization schemes based on one-hop approximation of Delaunay Triangulation to minimize the energy consumption and extend the lifetime of networks. Furthermore, we prove that our approximation of Delaunay triangulation guarantees complete coverage and generates the same radii assignment as traditional Delaunay triangulation. Our ns-2 based simulation shows performance improvements in many aspects such as coverage ratio, sensing energy consumption, node failure rate, and lifetime of networks.

Global manufacturing network (GMN) is a new manufacturing system that has super manufacturing and servicing abilities. It is challenging traditional manufacturing conceptions. This paper describes and analyzes the driving forces of emergence of the concept of GMN, its evolution, members, and its theoretical bases. This paper also explores the knowledge transfer in the GMN, which consists of the mechanism and process of knowledge transfer in the GMN. Finally, the factors which contribute the knowledge transfer in the GMN are discussed.

Global Manufacturing Network (GMN) is a new manufacturing system that has super manufacturing and servicing abilities. It is challenging traditional manufacturing conceptions. This paper describes and analyzes evolution of the concept of GMN, its members, and its theoretical bases. This paper also explores the reasons, of knowledge transfer in the GMN, as well as mechanism and process of knowledge transfer in the GMN. Finally, the implication of GMN for Chinese local firms is discussed.

This paper addresses a parametric method based on high-order ambiguity function (HAF) to solve the problem of phase contamination of HF skywave radar signals corrupted by the ionosphere. When signal-to-noise ratio and data sequence available satisfy the predefined conditions, the ionospheric phase contamination may be modeled by the polynomial phase signal. As a new parametric tool for analyzing polynomial phase signal, HAF is applied to estimate polynomial phase model parameters and reconstruct the disturbance signal. Using the estimated reconstructed signal, compensation can be performed before coherent integration and the original radar return spectrum can be restored. A piecewise scheme is proposed to track rapid variation of the phase contamination in HAF method, and it can remove the Doppler spread effect caused by the ionosphere nonstationarity. Simulation is used to demonstrate the efficiency of the proposed method.