Fault diagnosis is essential for the reliable, safe, and efficient operation of the plant and for maintaining
quality of the products in industrial system. This paper presents an ensemble fault diagnosis algorithm based on fuzzy
c-means algorithm (FCM) with the Optimal Number of Clusters (ONC) and probabilistic neural network (PNN), called
FCM-ONC-PNN. In clustering methods, the estimation of the optimal number of clusters is significant for subsequent
analysis. As a simple clustering method, FCM has been widely discussed and applied in pattern recognition and
machine learning, but FCM could not guarantee unique clustering result because initial cluster number is chosen
randomly. As the number of clusters is randomly chosen, the iterative amount is large and the result of the classification
is unstable. In this paper, firstly subtractive clustering is proposed to find the optimal number of clusters and the
clustering results of the FCM are compared with random initialization method, and then PNN is used to classify the
clustering data of FCM. The experiments show that the modified initial cluster number of FCM algorithm can improve
the speed, and reduce the iterative amount. At the same time, FCM-ONC-PNN approach can make classification more
stable and have higher precision.
Different from terrestrial networks, satellite network topology is time-varying. Moreover, satellite networks
communication resources are scare and expensive. So it is important to research on the fault management for
satellite networks to insure it work safely and smoothly. However, the traditional fault management method for terrestrial
networks is difficult to use for satellite networks. In this paper, the satellite networks topology is studied and
the satellite networks management architecture is established. Based on these the three level management mechanism
is proposed and put forward a new fault management method for satellite networks. When a satellite agent could
respond to the network management instruction received from the management station, the traditional fault management
method was used through network management technique. If the satellite could not respond to the network
management demand, the intra-domain cooperation or inter-domain cooperation would be activated. The suspected
fault satellite could be tested through cooperation among the satellite agents in satellites. The simulation results shows
that in the circumstance of the low faulty frequency, the new method could be effectively used in satellite network
with short cooperative time and low throughput.
The representation and organization of knowledge is the core of key technologies of artificial intelligence,
and the modeling of design knowledge is an important and difficult step in developing intelligent
designing system. The Design-Mode-Oriented Model (DMOM) used in representation of design knowledge
was proposed in this paper. And corresponding to DMOM, three object design models, named as function
model, parametric model and case model were also proposed to represent the knowledge of object design.
The information gathered from users can be classified into three kinds and each of which represents functions,
technical specifications and types of product respectively. Three design modes were also established
by mapping the three kinds of information to corresponding object design models. By using model transformation,
design information can be processed and the solution of design can be found. This modeling
method can be used in the applications of designing for most of electrical and mechanical products.
Based on a small, lightweight, low-cost high performance inertial Measurement Units(IMU), an effective
calibration method is implemented to evaluate the performance of Micro-Electro-Mechanical Systems(MEMS)
sensors suffering from various errors to get acceptable navigation results. A prototype development board based on
FPGA, dual core processor’s configuration for INS/GPS integrated navigation system is designed for experimental
testing. The significant error sources of IMU such as bias, scale factor, and misalignment are estimated in virtue of
static tests, rate tests, and thermal tests. Moreover, an effective intelligent calibration method combining with Kalman
filter is proposed to estimate parameters and reduce the effect of IMU dynamic errors that can degrade the system
performance. The efficiency of proposed approach is demonstrated by various experimental scenarios.
In variant design of mass customization, sometimes it’s necessary to modify some dimensions of product
modules. This paper introduces a methodology for dimension parameter transfer related to variant design that based
on constraint satisfaction priority of part and assembly connection. In order to research parameter transfer path and
build the parameter transfer structure, parts and its variant parameters are classified. Parameter association relations
between parts are analyzed based on assembly connection in product. Parameter constraint relations also presented
for variant design. And, parameter transfer structure and model is put forward. Constraint forms between dimension
parameter are listed and described, and then a reconstructed network structure for dimension constraints relations is
built in which every edge has only one direction with purpose of making the network simple and easy to solve the
constraints. Finally, using Pro-E as a 3D environment and VC++ as a development tool, all methods are applied to a
shaft-hub link of speed reducer product, thus verifying the proposed method.
This paper presents the procedures of face segmentation in the color image based on skin detection through
the establishment of skin model and the segmentation of skin region. Firstly, in order to reduce the effect of factors
on the segmentation of face region, a method for compensating the color of input images is used to alleviate the
interferences from bad illuminating conditions. Secondly, Gaussian model about skin information is established which
can be used to detect skin pixels in color images and transform color images to gray-scale images. Thirdly, a new
algorithm of Otsu is used to find out the skin regions in binary images. Finally, mathematical morphology operator
and prior knowledge are used to find out the face regions and discard regions that are similar to the skin in color.
This method can deal with various sizes of faces, different illumination conditions, diverse poses and changeable
expressions. In particular, the scheme significantly increases the execution speed of the face segmentation algorithm
in the case of complex backgrounds. The experiments show that this method reduces the computation of the procedure,
and at the same time improves the detection speed and efficiency.
In this paper, the correlation tracking algorithm based on the function of minimum average absolute
difference is analyzed. In addition, a kind of correlation tracking algorithm based on particle filtering is used to avoid
the target shaded by object. And last, Kalman filtering approach is used as the motion model of the multi-target
tracking, and the recursive filtering method is adopted to calculate and predict the location of moving targets. The
tracking results show that correlation tracking algorithm based on particle filtering can overcome the obstacle occlusion
problem perfectly, which can also improve the robustness of target tracking. Besides, target tracking algorithm based
on Kalman filtering works well for multi-target tracking, which has great practical significance on tracking technology
for moving targets.
It is a fundamental problem to calculate target area in pattern recognition. This paper compares the
common algorithms for calculating target area, and analyzes the advantages and disadvantages of these algorithms.
In order to overcome the shortcomings of common algorithms and obtain a more simple and effective algorithm, the
knowledge of mathematical morphology is studied. Then, a new algorithm based on an improved algorithm of contour
extracting for calculating target area is proposed. The algorithm will obtain the coordinate and type information of
the contour points by extracting the contour of the target region and calculate the target area based on weighting
summation of the abscissa of contour point. At last, detect whether the algorithm is effective and practical or not by
comparing with several common algorithms. The result of experiment shows that this algorithm is simple, effective
and easy to implement.
MIMO (Multiple-Input Multiple-Out-put) system is a core technology used in 802.11n which used multiple
antennas to suppress fading. In the field of WSN network, routing protocols must be designed to minimize
energy consumption for the nodes’ energy supply is constrained. Using MIMO technique in WSN could make use of
the energy accumulation in the receiving nodes to increase the receiver signal intensity, thus reducing bit error rate
and improve the reliability of data communication. As the MIMO-based WSN could supply higher data transfer rate
for the same transmitting power and bit error rate, it also efficiently reduces the time of the node sending data and
improves communication performance of the whole network. In this paper, based on dynamic collaborative virtual
antenna MIMO technology, a new energy efficient routing protocol is proposed. The simulation results indicate that
our protocol can obviously reduce the energy consumption in the specific application environment.
In signal processing circuits of a RDIF tag, the relaxation of restriction concerning operating voltage range
enables us to develop various applications. In order to provide higher voltage to the signal processing circuit, two types
of step-up AC-DC converters for RFID tags are proposed in this paper: the
step-up converter and the
converter. To reduce ripple noise of output voltage, the proposed converters consist of 2 charge-pump type AC-DC
converters with opposite polarities. Unlike conventional converters, the proposed converters can offer higher output
voltages by realizing /
step-up conversion. Furthermore, as for the
step-up converter, the circuit size of the
proposed converter is almost the same as that of the Cockcroft-Walton type AC-DC converter. The properties of the
proposed converters are clarified by theoretical analyses. The theoretical formulas can derive the equivalent circuits,
output voltages, and power efficiency. Furthermore, SPICE simulations and the experiments showed the validity of
theoretical formulas and circuit design.
We performed statistical analysis on the total PTR resource record (RR) based DNS query packet traffic
from a university campus network to the top domain DNS server through March 14th, 2009, when the network servers
in the campus network were under inbound SSH dictionary attack. The interesting results are obtained, as follows:
(1) the network servers, especially those providing SSH services, generated the significant PTR RR based DNS query
request packet traffic through 07:30-08:30 in March 14th, 2009, (2) we calculated sample variance for the DNS query
request packet traffic, (3) the variance can change in a sharp manner through 07:30-08:30, (4) we developed a couple
of DNS based SSH detection technologies by employing the PTR RR DNS query request packet traffic variance- and
the DNS query keywords Euclid distance based methods, and (5) we evaluated and compared the both detection rates.
As a result, although the both detection technologies take high detection rates, the Euclid distance based detection
technology can take a low false positive rate than that of the variance based one, indicating that we can detect the
inbound SSH dictionary attack to the network server in the campus network by observing the total PTR RR DNS
query request packet traffic from the campus network.
This paper discusses the fundamental matrix that is recovered by the use of the characteristic points of
the images collected by camera. Firstly, the corners are detected by Harris Operator, but there are error matching
points in the set of characteristic points. Secondly, the relaxation matching is used to search fine matching point pairs.
Thirdly, in order to gain better points, the objective function is founded by interrelated geometry theory. Lastly, the
fundamental matrix will be solved with corresponding matching points best matching points are gained. The PSO
algorithm is used to optimize the fundamental matrix. And the epipolar line is recovered by the fundamental matrix.
The results of experiments show that better effects are obtained.
In wireless sensor network the bad distribution for cluster head was not taken into account in traditional
clustering algorithm such as LEACH algorithm. However the distribution uniformity for the cluster head in wireless
sensor networks is very critical, for the better cluster head distribution could effectively save the communication cost
and prolong the the lifespan of the network. The math programming approach that we use in this paper is based
on a variation of a maximal expected covering location model due to Daskin. Daskin introduced a variant of the
MCLP that considers the possibility that facilities may be unable to respond to demand at all times. The resultant
model was labeled as MEXCLP (Maximum Expected Covering Location Problem). So we put forward the MEXCLP
algorithm for clustering according to the greatest expectations of coverage principle, in this algorithm the cluster
head was taken regarded as the provider of services, the cluster head could provide service to the member nodes in
the cluster. The algorithm considers the process of service failure (link failure, and so on), and it is based on the
assumption of probability of the services failure, a reasonable choice through the head cluster node enables network
nodes be served with the largest request number. From the simulation results can be see that the MEXCLP algorithm
is more uniformed in finding head cluster, and under the same failure rate circumstances, MEXCLP algorithm may
provide more efficient service and consume less energy than the LEACH algorithm. So this clustering algorithm could
effectively and efficiently be suitable to the wireless sensor network and provide more reliable service for the sensor
nodes in the cluster.
There are two basic methods of fuzzy pattern recognition: one is direct method based on membership
degree, and another is indirect method based on closeness degree. In this paper, different algorithms of closeness
degree have been studied and a new method for calculating the closeness degree has been introduced. By combining
the concepts of membership degree and closeness degree, the pattern characteristic values derived from both real
numbers and fuzzy set expressions can be handled, then soft transition from direct method to indirect method can be
realized by the new algorithm; It also could adjust every kind of weights of closeness degree to adapt itself based on the
characteristics of classical model and be recognized model. And some case studies would be provided to demonstrate
the effectiveness and availability for the new algorithm. The algorithm could also be used in different circumstances
to deal with closeness degree other than pattern recognition.
In the past years there has been an enormous growth in the use of Internet, and new real-time connectionoriented
services like streaming technologies and mission-critical transaction-oriented services are in use and new
ones are currently emerging. The research on more reliable network becomes an inevitable trend presently. MPLS
is a next generation backbone architecture, which can speed up packet forwarding to destination by label switching.
However, if there is not a backup LSP when the primary LSP fails, MPLS frames cannot be forwarded to destination.
Therefore, fault recovery has become an important research area in MPLS Traffic Engineering. At present, Protection
Switching can be approached by two famous methods, Makam and Haskin based on which other methods basically
come into being. When the place of failure is away from ingress node, a problem will come out. These two famous
methods both have their disadvantage. In order to minimize or eliminate their drawback, this thesis tries to do some
exploration on the MPLS-based recovery model. The model in the thesis using the Reverse Backup Path to solve
the loop of data back to the path is too long, and the simulation experiment shows that new method of MPLS-based
recovery has less packet disorder and much lower delay and packet losses.
This paper proposes a connection weighting scheme of a complex-valued Hopfield neural network for
associative memory constrained by given attractive domain. Both equilibrium conditions and stability analysis results
are used in the synthesis procedure. We solve the equilibrium equation by singular value decomposition technique
and obtain a general solution of the connection weight matrix with a free sub-matrix. Such general solution and the
parameter matrix corresponding to the given attractive domain are contained in the inequations which are derived
from stability analysis and can be represented as linear matrix inequations (LMIs). The connection weighting solution
of such LMIs can guarantee stability and attractability of the network simultaneously. A simulation example of a
3-dimension complex-valued Hopfield neural network shows the proposed synthesis method. The simulation results
demonstrate the attractive ability of two complex-valued vectors in the prespecified attractive domain.
SRM has become an attractive alternative in adjustable speed drive due to its simplicity in both motor
construction and power converter, and vibration that causes acoustic noise in SRM have generated intense interest
from the time it became commercially competitive with dc and ac drives. Basically, there are two conventional ways
to reduce the noise: optimize the physical structure of the motor, or, choose appropriate control strategies. In this paper,
a new method is proposed to combine these two optimization ways together: harmony optimization of SRM. Physical
design is mainly aiming at the back iron thickness and the air gap length of the motor. Control optimization is about
the control mode, turn-on and the turn-off angle which affects the power factor strongly. Based on the same cooper
losses, the mathematics analysis and finite element simulation are carried out. The results show that the harmony
optimization will reduce the vibration and torque ripple in a large extent without extra power losses.
This paper discusses about the issue of ”soft control” for swarms system in Euclidean space, which
coordinates the collective behavior of the group by adding a few controlled intelligent agents under the condition
of keeping the local rules of the existing agents in the system. It shows that the swarm center will be effectively
transferred into an expectant position by putting a few controlled intelligent agents and controlling their initial position
or controlling their position for a short time, and the swarm members will converge to a bounded region around the
expectant position in a finite time. This paper gives a controlled law for controlled intelligent agents. Simulation
testing shows the feasibility of soft control for swarm system.
The electric care bed control system is divided into three modules, which are physiologic parameter
detection module, human-computer interaction module and motor drive module (MDM). Those modules communicate
with each other by controller area network (CAN) bus. Under the idea of modular design, a plug and play (PnP)
multifunctional motor drive module (MDM) is developed. This MDM has a CAN interface which can make it possible
to communicate with human-computer interaction module, and then control the position closed-loop and speed closedloop.
The MDM based on CAN bus will be more efficient in communication and system expansion and make the
electric care bed function perfectly.
This note provides new stability criteria for a class of uncertain singular systems with multiple-state
delays. By introducing the state transformation, the study of the robust stability for original systems is changed
into this study for the equivalent systems. Based on the Lyapunov-Krasovskii functional combination with LMI
techniques, a delay-dependent robust stability criterion for the nominal systems of a class of uncertain singular systems
is established, which ensures the nominal systems are asymptotically stable. Furthermore, the delay-dependent robust
stability criterion for a class of uncertain singular systems with multiple-state delays is presented, which ensures the
class of uncertain singular systems is asymptotically stable. Finally, two numerical examples are given to illustrate the
effectiveness of the obtained results.
Shield is a typical mechanical, electrical, hydraulic integration of equipment. Its faults species are complex
and diverse. To prevent because machine failure causes economic losses and casualties by shield, this article will
introduces rough set theory to the subway shield machine fault diagnosis, propose a method which is based on rough
set theory combined with neural network of Metro shield machine fault diagnosis. Use the strong advantage of rough
sets theory in attribute reduction, and remove the data redundancy of information which is not effective for decisionmaking.
Application of neural network algorithm to reduce date for diagnosis, the method can effectively improve
the speed and accuracy of the diagnosis. Then use BP neural network combined with least square method to forecast
fault, the least-square can reflect the trend of linear sequence, Neural network can seize the variation of nonlinear time
series, therefore the combination of two methods could well predict the future of unit operating conditions.
Planetary exploration is the embodiment of a nation’s synthesis science and technology power. The first
step is to land a rover to explore the unknown world. The science task always begins from rock and soil. Rock
sample and return is one of the important tasks of planetary exploration, which should be accomplished by the rover’s
hardware and software. A series of algorithms related to main science exploring object-rock are investigated in this
paper, such as a single rock’s segmentation, shadow elimination, stereo matching and 3D reconstruction. C-means
clustering and 3D surface evaluation realized by surrounding binocular vision are described. On the basis of the above
algorithms, a virtual mechanism is developed by VC++ and OpenGL. The evaluated information for localization is
transformed to the arm and leads the arm to locate the relative plane on the rock and shows the grinding process. The experiments and simulation results based on real image show the validity of these algorithms.
The necessity recurrently comes to align a strapdown inertial navigation system (SINS) in a moving
vehicle to guarantee the accuracy and efficiency in the long run-off of the inertial system after a take-off or launch
command is issued. This in-flight alignment is therefore achieved by integrating SINS data with some external aiding
source inlcluding airborne navigation equipments and networking sensors. In this paper, a localization architecture
and alignment scheme is presented for aircraft in a three-demensional fleet network, which is based on wireless
sensor network. Firstly, a 3D node localization scheme is designed based on weighed-multidimensional scaling,
which adopt spherical locating in the initial stage, and adaptively choose source nodes with high relative reliability to
achieve position update. Then a robust filter algorithm is applied to compensate time-varying delay error and large
azimuth uncertainty in alignment. Extensive simulation shows that the DMDG-3D localization scheme can provide
highly accurate and relatively reliable navigation information in real time, and filter algorithm can
accelerate convergence and give better estimation of the navigation parameters.
Intelligent video monitoring system has been used widely in daily life. In order to avoid the casualties, as
well as to predict potentially dangerous situations, real-time monitoring of crowd activities indoor has become an
urgent requirement. In recent years, the video-based passenger flow counting systems have been improved a lot with
constant equipment update. However, these systems are mostly aimed at a particular scene. In crowded
circumstances, the statistical precision is not very satisfactory. The analysis of abnormal situations is thus imperfect.
This paper improves the main algorithms, such as: the extraction of human, the segmentation of crowd and the
judgment of human moving direction. Finally, the improved algorithms integrate a system which achieves two-way
passenger flow counting. Consequently it first makes the early warning of abnormal conditions possible. The
system’s statistical accuracy is remarkably improved, and testing result under different scenes is shown.
Network scheduling plays an important role in Network Control System (NCS). Focused on problems of
the present scheduling of CAN bus, an improved Mixed Traffic Scheduling algorithm (MTS) was proposed on the
base of the communication principle of network scheduling. The cut-off time was denoted after introducing the
definitions about the scheduling. The NCS model for CAN Network was created then. The simulation platform was
created based on MATLAB/True Time toolbox, construed the new schedule affected the control system. The
simulation test was conducted and the result proved the validity of the method.
Traditional network control system offers many advantages but requires cables to inter-connect devices,
which leads to high installation and maintenance costs due to low scalability and high failure rate of connectors. As a
result, wireless technologies have gained an enormous success in the consumer goods industry in the last few years.
Wireless Networked Measurement Control System (WNMCS) is a distributed control system based on the wireless
measurement network, suitable for use at the device level of an automation system. For that purpose, Wireless Local
Area Networks (WLAN) and Wireless Personal Area Networks (WPAN) can be employed. In this paper, a
simulation model of WNMCS is designed on the basis of Matlab/Simulink, and the suitability of wireless sensor
networks for networked control loop is shown. Real time character is the common request of control systems and
multitasks is the mainstream of the distributed control system nowadays. Aiming at selecting communication
protocol in the wireless networks, the TrueTime toolbox is applied to establishing the simulation model of the
Wireless Network Measurement Control System (WNMCS). We compare the performance of WPAN-based
WNMCS to that of WLAN-based WNMCS. Experimental results show that WPAN-based WNMCS is a very
promising alternative for wireless industrial networking with short superframe, and WLAN-based WNMCS is more
suitable for high throughput application.
The integration of uncertain information from different time sources is a crucial issue in various
applications. In this paper, we propose an integration method of multiple Temporal Qualitative Probabilistic
Networks (TQPNs) in time series environments. First, we present the method for learning TQPN from time series
data. The TQPN’s structure is constructed using Dynamic Bayesian Networks learning based on Markov Chain
Monte Carlo. Furthermore, the corresponding qualitative influences are obtained by the conditional probabilities.
Secondly, based on rough set theory, we integrate multiple TQPNs into a single QPN that preserves as much
information as possible. Specifically, we take the rough-set-based dependency degree as the strength of qualitative
influence, and then make the rules to solve the ambiguities reduction and cycles deletion problems which arise from
the integration of different TQPNs. Finally, we verify the feasibility of the integration method by the simulation
Multi-Robot Task Allocation is a crucial issue before performing a certain task. This paper deals with a
distributed task allocation method based on some special relation defined according to the performance of history
cooperation between two robots. The algorithm we propose here is named TARARC—a Task Allocation algorithm based
on Robot Ability and Relevance with group Collaboration, where robot ability is weighed by reliability, relevance
represents a fresh concept of “history relevance” between every two robots to establish reasonable groups for better
collaboration, and the group collaboration includes inter and inner group help strategy that are adopted when different
nodes failures happen in unknown environment. TARARC emphasizes the role of “agent node” in each group that is
responsible for task competition, group leadership, formation maintenance as well as task execution with changing agents.
Simulation on Player/Stage shows that our mechanism is feasible and valid.