Search

In this paper, a contour error control strategy based on spatial iterative learning control (sILC) is developed for repetitive processing of five-axis computer numerical control (CNC) machine tools. A curve approximation method with an adaptive moving window is developed to achieve accurate tool position and orientation contour error estimation, and a five-axis contour error control strategy based on sILC is proposed. The compensation method is derived using an sILC algorithm to modify the geometric reference path instead of modifying the controller. The experimental results show that the proposed control strategy reduces contour errors of five-axis CNC machine tools, and it outperforms the traditional tracking error control.

Currently, depression has become a common mental disorder, especially among postgraduates. It is reported that postgraduates have a higher risk of depression than the general public, and they are more sensitive to contact with others. Thus, a non-contact and effective method for detecting people at risk of depression becomes an urgent demand. In order to make the recognition of depression more reliable and convenient, we propose a multi-modal gait analysis-based depression detection method that combines skeleton modality and silhouette modality. Firstly, we propose a skeleton feature set to describe depression and train a Long Short-Term Memory (LSTM) model to conduct sequence strategy. Secondly, we generate Gait Energy Image (GEI) as silhouette features from RGB videos, and design two Convolutional Neural Network (CNN) models with a new loss function to extract silhouette features from front and side perspectives. Then, we construct a multi-modal fusion model consisting of fusing silhouettes from the front and side views at the feature level and the classification results of different modalities at the decision level. The proposed multi-modal model achieved accuracy at 85.45% in the dataset consisting of 200 postgraduate students (including 86 depressive ones), 5.17% higher than the best single-mode model. The multi-modal method also shows improved generalization by reducing the gender differences. Furthermore, we design a vivid 3D visualization of the gait skeletons, and our results imply that gait is a potent biometric for depression detection.

This dissertation contains three chapters. The first chapter evaluates the effect of a gun control act in California. State legislators in the U.S. are striving to curb gun violence. A common approach is to extend the existing firearms ban list. This paper examines the effect of legislation restricting sales of selected firearms in California using the synthetic control method. This case study method forms a synthetic unit using a linear combination of other states in the U.S. as the control group. The results show substantial increases in firearm sales in California from the point of passage until the law becomes effective. After the surge ends when the law becomes effective, the sale of firearms is only moderately affected thereafter. This paper also creates robustness checks to confirm that the synthetic control method is working properly with low firearm density in California, which calls into question some of the assumptions underlying the synthetic control method. The Difference-in-Difference regression reaches the same conclusion. The second chapter focuses on immigrant assimilation in the U.S. Assimilation is the process in which immigrants improve earnings as they become more adapted to the host country society. Cross-sectional studies show that immigrants have lower earnings upon arrival and faster earnings growth compared to natives. Longitudinal studies conclude that estimates based on cross-sectional data are positively biased due to decreasing cohort quality and negatively selected outmigration. I reproduce such estimates with recent U.S. data. The estimates would appear to show "bias," as inclusion of cohort fixed effects alter estimates. However, in contrast to expectations based on the current literature, decreasing cohort quality and outmigration do not explain the difference. Next, I apply a non-parametric method to make the wage distributions visually comparable across cohorts and time. I find that the linear specification of assimilation is misleading. Finally, I revisit the classic model with a quadratic assimilation term and expand it to explore the assimilation process's heterogeneity. I find that the "bias" disappears with a quadratic assimilation effect. The assimilation effect is sensitive to age at arrival and country of origin. The third chapter considers an unexplained puzzle in one of the most widely used public datasets in the U.S. The American Community Survey (ACS) replaced the Decennial Census as the primary data source for identifying immigrants' socioeconomic characteristics. This paper focuses on cohort analysis, in which a cohort combines immigrants arriving in a given year from surveys in multiple years. Tracking the sizes of cohorts from 2006 to 2019 using the ACS, we observe an abnormal increase in cohort size in the 10th and 20th years since arrival. Two hypotheses are tested, population estimate structural break and the renewal of green card. Neither appears to explain the puzzle.

Alzheimer's disease(AD) is a neurodegenerative disease that progresses slowly but worsens gradually, also, the most common kinds of dementia. Clinically, the diagnosis of AD is mainly based on rating scales and neuroimaging technology which is invasive, costly and time-consuming. Other than that, the clinical pathology has become irreversible when neuroimaging characteristics appear. It is imperative to develop new noninvasive methods for early diagnosis of AD. Several studies indicated the probable association of cognitive decline with gait changes might shed light on potential features for distinction of AD. This paper aims to exploit the feasibility of gait features for early diagnosis of mild cognitive impairment(MCI) and AD by using machine learning methods. A device-free AD detection system is built, with a natural undisturbed gait collecting system and a well-performed Long Short-Term Memory(LSTM) based model, in this article. Moreover, it can serve as a simplified, non-invasive, and highly accurate clinical auxiliary tool for early diagnosis and distinction of AD. Experimental results showed a 90.48%, 92.00%, and 88.24% in accuracy, sensitivity, and specificity respectively for distinguishing AD by using the method with LSTM based model. Furthermore, the gait cycle and stride length in MCI or AD were more variable than in healthy controls through redefining and calculating the gait features with skeleton data obtained by Kinect devices.

Classification of Alzheimer's Disease (AD) has been becoming a hot issue along with the rapidly increasing number of patients. This task remains tremendously challenging due to the limited data and the difficulties in detecting mild cognitive impairment (MCI). Existing methods use gait [or EEG (electroencephalogram)] data only to tackle this task. Although the gait data acquisition procedure is cheap and simple, the methods relying on gait data often fail to detect the slight difference between MCI and AD. The methods that use EEG data can detect the difference more precisely, but collecting EEG data from both HC (health controls) and patients is very time-consuming. More critically, these methods often convert EEG records into the frequency domain and thus inevitably lose the spatial and temporal information, which is essential to capture the connectivity and synchronization among different brain regions. This paper proposes a cascade neural network with two steps to achieve a faster and more accurate AD classification by exploiting gait and EEG data simultaneously. In the first step, we propose attention-based spatial temporal graph convolutional networks to extract the features from the skeleton sequences (i.e., gait) captured by Kinect (a commonly used sensor) to distinguish between HC and patients. In the second step, we propose spatial temporal convolutional networks to fully exploit the spatial and temporal information of EEG data and classify the patients into MCI or AD eventually. We collect gait and EEG data from 35 cognitively health controls, 35 MCI, and 17 AD patients to evaluate our proposed method. Experimental results show that our method significantly outperforms other AD diagnosis methods (91.07 vs. 68.18%) in the three-way AD classification task (HC, MCI, and AD). Moreover, we empirically found that the lower body and right upper limb are more important for the early diagnosis of AD than other body parts. We believe this interesting finding can be helpful for clinical researches.

The paper presents a formal proof of a machine closed theorem of TLA+in the theorem proving system Coq. A shallow embedding scheme is employed for the proof which is independent of concrete syntax. Fundamental concepts need to state that the machine closed theorems are addressed in the proof platform. A useful proof pattern of constructing a trace with desired properties is devised. A number of Coq reusable libraries are established.

Mobile ad hoc network (MANET) is a dynamic wireless communication network. Because of the dynamic and infrastructureless characteristics, MANET is vulnerable in reliability. This paper presents a novel reliability analysis for MANET. The node mobility effect and the node reliability based on a real MANET platform are modeled and analyzed. An effective Monte Carlo method for reliability analysis is proposed. A detailed evaluation is performed in terms of the experiment results.

This paper considers a novel fault diagnosis mechanism for wireless sensor networks (WSNs). Without additional agents, the built-in and self-organized diagnosis mechanism can monitor each node in real time and identify faulty nodes. As the diagnosis is operated within a cluster of nodes, it can reduce power consumption and communication traffic. We present a modeling of the diagnosis algorithm for WSNs, with a probabilistic analysis of the local and global performances of our approach. Extensive experiments demonstrate the effectiveness of the proposed method.

Since the market competition of manufacturing becomes intense, improving the productive efficiency is of the most importance. Manufacturing Execution System (MES) is considered to be a good solution. But the communication network, which is the basis of MES, faces with the serious environment of the shop in manufacturing. Traditional methods, such as cable network and AP based wireless network, cannot meet the requirements of this environment. MANET is regard as a reasonable solution, but weak in the reliability. There are many researches focus on the reliability of MANET. But most of them aim at a general MANET's environment. So these works cannot provide guidance for the real environment. Based on a deep analysis of the application and environment, this paper proposes a reliability model for an assembly-line production oriented MANET. And the analysis of this reliability model is presented.

International audience; Various forms of pervasive computing environments are being deployed in an increasing number of areas including hospitals, homes and military settings. Entities in this environment provide rich functionalities (i.e. services). How to organize these heterogeneous and distributed entities to deliver user-defined services is challenging. Pantagruel is an approach to integrate a taxonomical description of a pervasive computing environment into a visual programming language. A taxonomy describes the relevant entities of a given pervasive computing area and serves as a parameter to a sensor-controller-actuator develop- ment paradigm. The orchestration of area-specific entities is supported by high-level constructs, customized with respect to taxonomical information. Pantagruel is also a language that describes and manages services. Further more, Pantagruel can be viewed as a high level service contract between the service designer and the program implementer. This paper presents a for- malization of Pantagruel, both its syntax and semantics. Four kinds of static properties are stated based on the formalization. Predicate abstraction based algorithms are designed to verify the properties.

Robust scheduling is an effective way to solve job-shop scheduling problem (JSP) under a dynamic and stochastic environment with uncertainties, which may lead to numerous schedule disruptions. Maintaining the optimization obtained, the robust schedule is more stable in execution. The vast majority of the researches on robust scheduling concentrate on the development of the robustness and stabilization with an optimization target that minimizes makespan. However, other optimization targets are needed in reality, and in this paper, we introduce another optimization target that maximizes the payload of the machine, which can reflect the capacity of machine. This work applies a two-step way to achieve a robust schedule. A baseline schedule in the deterministic environment is generated first, which pursues standard optimization criterions, and then the temporal buffers are inserted into the baseline schedule which make the schedule absorb some level of unexpected disruption without rescheduling. According to the payload of each machine which should be close to the maximum at any time in execution, we can get each size of the temporal buffers inserted into the different operations. An extensive experiment is used to show the relative performance of the proposed robust schedule. It is shown that the schedule is robust and stable under the stochastic disturbance coupling with minimal makespan tardiness and symmetry payload.

MANET (Mobile ad hoc network) is vulnerable to security attack due to its characteristics such as open network environment, lack of centralized control, dynamic topology, and so on, while traditional security mechanisms cannot meet the security requirements of MANET due to the limited communication bandwidth, computation power, memory and battery capacity, and dynamic deployment environment. Many approaches have been suggested to trade-off between the security strength and efficiency, and most of them focus on the confidentiality and integrity of MANET while pay less attention to the availability issue. Based on the analysis of availability related factors in MANET, we present an availability based trust model in this paper. Furthermore, a cluster based trust model (ABTMC, Availability Based Trust Model of Clusters) is presented with the consideration with the resource constrain of node, dynamic topology, and the unbalance of computing capability and resource capability among MANET nodes. Many availability oriented security service can be obtained based on this model.

Node misbehavior due to selfish or malicious intention could significantly degrade the performance of MANET because most existing routing protocols in MANET are aiming at finding most efficiency path. To deal with misbehavior in MANET, an incentive mechanism should be integrated into routing decision-making. In this paper firstly we review existing techniques for secure routing, and then propose to use trust vector model based routing protocols. Each node would evaluate its own trust vector parameters about neighbors through monitoring neighbors' pattern of traffic in network. At the same time, trust dynamics is included in term of robustness. Then we evaluate the performance of the proposed mechanism by modifying Dynamic Source Routing (DSR) so that each node has a dynamic changing "trust vector" for its neighbors' behaviors. We also compare the simulation results of with and without the proposed mechanism. The simulation results demonstrate that the proposed mechanisms can effectively detect malicious nodes and mitigate black hole attacks.

Biometric authentication and privacy protection are conflicting issues in a practical system. Since biometrics cannot be revoked or canceled if compromised duo to the permanent association with the user, privacy-preserving biometric recognition is desired. However, the recently proposed template protection schemes are not yet sufficiently mature. Specially, the popular noninvertible transform approach will result in an obvious decrease of GAR for a fixed FAR. In this paper, we put forward a novel anonymous fingerprint recognition scheme, Biomapping, as the first approach to integrate the feature extraction, noninvertible transform, and anonymous query as a whole. Biomapping extracts the fingerprint feature utilizing a minutiae-centered region encoding, then performs anonymous enrollment and verification using the noninvertible and discriminable constructions. Experiments on the public domain database show Biomapping can provide better recognition accuracy along with the ability to protect the biometric template, thus becomes a promising solution for privacy trustworthy biometric applications.

Purpose To optimize timing parameters in an intermolecular double-quantum coherence (iDQC) imaging pulse sequence for overall image signal-to-noise ratio (SNR) and blood oxygenation level-dependent (BOLD) sensitivity for brain functional imaging. Material and Methods Fresh human blood was measured under different oxygenation conditions, and human brain functional magnetic resonance (fMR) images in three normal volunteers were obtained, using iDQC techniques at 1.5 T. The dependence of SNR and BOLD sensitivity was measured as a function of time delays after the iDQC evolution period. Results A time delay after the iDQC evolution period τ can be adjusted either to refocus the dephasing accumulated during τ, thus increasing SNR, with full rephasing occurring at delay = ±2τ (for iDQC order n = ±2), or to enhance BOLD effects with consequent reduced image SNR at delay = 0. Conclusion Image SNR and BOLD sensitivity often impose different requirements for iDQC image sequence design and timing parameter selections. It is therefore important to select properly relevant parameters for different applications. J. Magn. Reson. Imaging 2002;16:733–740. © 2002 Wiley-Liss, Inc.