In this paper, we propose to use an inverse-based method that is implemented by a Genetic algorithm (GA) for
obtaining the strain and temperature profiles in a 10-mm fiber Bragg grating (FBG) sensor and a series set of ten 10-mm sensors. The changes of strain and temperatures are analyzed by utilizing the sensitivity of the refractive index
and grating period of the fiber Bragg grating sensor. This can be accomplished by reconstructing the FBG structural
shape by using a Genetic algorithm that is compared with the measured output data. Our ultimate objective for
utilizing these results are intended for real-time sensing of strain and temperature of these sensors which are ideally
suited for smart structures health monitoring and diagnostics applications.
Future space missions, such as those involving formation flying of multiple satellites require high operational
autonomy mainly with the aim of reducing the operation costs and improving reactivity to sensed data. In particular,
stringent performance requirements envisaged precision formation flying cannot be achieved by currently
available technologies. One of the main challenges in achieving autonomy is the capability of fault management
without extensive involvement of ground station operators. This paper uses a second order nonlinear sliding
mode observer to detect actuator faults in the attitude control system of a satellite with four reaction wheels in
a tetrahedron configuration. A post-processing of residuals is required to isolate and reconstruct the faults in all
four reaction wheels. Furthermore, the control strategy needs to be reconfigured to recover faults. Simulation
results show that the proposed strategy can detect, isolate and reconstruct reaction wheel faults in the attitude
control system of a satellite.
This paper investigates the problem of fault tolerant cooperative control for UAV rendezvous problem in which
multiple UAVs are required to arrive at their designated target despite presence of a fault in the thruster of any
UAV. An integrated hierarchical scheme is proposed and developed that consists of a cooperative rendezvous
planning algorithm at the team level and a nonlinear fault detection and isolation (FDI) subsystem at individual
UAV's actuator/sensor level. Furthermore, a rendezvous re-planning strategy is developed that interfaces the
rendezvous planning algorithm with the low-level FDI. A nonlinear geometric approach is used for the FDI
subsystem that can detect and isolate faults in various UAV actuators including thrusters and control surfaces.
The developed scheme is implemented for a rendezvous scenario with three Aerosonde UAVs, a single target,
and presence of a priori known threats. Simulation results reveal the effectiveness of our proposed scheme
in fulfilling the rendezvous mission objective that is specified as a successful intercept of Aerosondes at their
designated target, despite the presence of severe loss of effectiveness in Aerosondes engine thrusters.
It is widely understood that communication is a critical technological factor in designing autonomous unmanned
networks consisting of a large number of heterogeneous nodes that may be configured in ad-hoc fashions and
incorporating intricate architectures. In fact, one of the challenges in this field is to recognize the entire network as
a heterogenous collection of physical and information systems with complicated interconnections and interactions.
Using high data rates that are essential for real-time interactive command and control systems, these networks
require utilization of optimal integration of local feedback loops into a scheduling and resource allocation systems.
This integration becomes particularly problematic in presence of latencies and delays.
Given that dynamics of a network of unmanned systems could easily become unstable depending on interconnections
among nodes, in this paper stability of the resulting time-delayed controlled network based on
configuration changes is studied. We also formally investigate sufficient conditions for our proposed robust resource
allocation strategies to be able to cope with these interconnections and time-delays in an optimal fashion.
Our time-delayed dependent network consists of three nodes that can be configured into different architectures.
To model our traffic and network we use a fluid flow model that is of low order and simpler than a detailed
Markovian queueing probabilistic model. Using sliding mode-based variable structure control (SM-VSC) techniques
that enjoy robustness capabilities, we design on the basis of an inaccurate/uncertain model our proposed
robust nonlinear feedback-based control approaches. The results presented are analyzed analytically to guarantee
stability of known/unknown time-delayed dependent network of unmanned systems for different configurations.
Large scale unmanned networks consisting of a number of heterogeneous
nodes that may be configured in ad-hoc fashions and incorporating complicated
architectures result in challenging problems
for design of appropriate control and resource allocation
optimization techniques. The problem is further compounded by
the fact that designing appropriate network control methodologies
subject to bandwidth, latencies and computational resources for these
network-centric systems are highly non-trivial.
In this paper, we only investigate one of a number
of critical issues that are of interest in this domain, namely
the problem of congestion control of a network
that consists of three nodes
that can be configured into different architectures.
This study shows that depending on the
interconnections between the network nodes the dynamics of
the resulting closed-loop system can change
considerably so that the unmanned system could become even unstable
Therefore, a robust control strategy is required to be able to cope with any
configuration changes and to be able to address
the resource allocation problem subject to the propagation delays and latencies.
For sake of comparative evaluation, we first implement a standard PID
control scheme which is shown to lack sufficient capability
for achieving the desired performance requirements. Subsequently, a nonlinear
control scheme is proposed to resolve the limitation of
sensitivity of the closed-loop system
to propagation delays.
The proposed strategy is based on a well-known input-output
feedback linearization approach that is shown to achieve an appreciable
improvement in the performance of the closed-loop unmanned network and
which is also less sensitive to the network propagation delays.
The problem of Attitude Recovery of rigid and flexible spacecraft/satellite is investigated using the feedback linearization control approach. The attitude and flexible dynamics equations for a class of spacecraft/satellite are presented. Since the flexible spacecraft is under-actuated, the input-output linearization technique was specifically used to break up the system into two distinct parts, namely (1) an external linearizable system for which a linear controller can be easily implemented and (2) an internal nonlinear unobservable system for which the associated zero dynamics is shown to be asymptotically stable for a representative case. The overall closed-loop stability of the flexible satellite is analyzed rigorously and shown to be asymptotically stable using Lyapunov's method. In order to design and analyze attitude control system for satellites, it is important to be able to simulate the dynamics of the spacecraft. Hardware-in-the-loop simulations of a spacecraft, such as air-bearing spacecraft simulators, are not only expensive to build, but they cannot provide the full experience of micro-gravity. An alternative is to have a high fidelity software simulator. Consequently, the resulting nonlinear and coupled equations of the satellite are implemented into a high-fidelity, user-friendly simulation environment, named the Flexible Spacecraft Simulator (FS2). The development and utilization of the FS2 for our research will also be presented.
This paper presents a new sequential decision feedback approach for pattern classification in a changing environment. An adaptive classification system is developed that uses the decisions of multiple aspects that may not be separated uniformly. A tap delay mechanism is used to impact the final decision at the current aspect of the object. This system minimizes the error of the classifier while it maps the new feature vector to a familiar feature space for the classifier. The test results on an acoustic backscattered data set collected from six different objects: two mine-like and four non-mine-like at 72 aspect angles with 5 degrees of separation and with varying signal-to-reverberation ratio (SRR) from 4 to 16 dB are presented.
Compression of digital images has been a very important subject of research for several decades, and a vast number of techniques have been proposed. In particular, the possibility of image compression using Neural Networks (Nns) has been considered by many researchers in recent years, and several Feed-forward Neural Networks (FNNs) have been proposed with reported promising experimental results. Constructive One-Hidden-Layer Feedforward Neural Network (OHL-FNN) is one such architecture. At previous SPIE conferences, we have proposed a new constructive OHL-FNN using Hermite polynomials for regression and recognition problems, and good experimental results were demonstrated. In this paper, we first modify and then apply our proposed OHL-FNN to compress still and moving images and investigated its performance in terms of both training and generalization capabilities. Extensive experimental results for still images (Lena, Lake, and Girl) and moving images (football game) are presented. It is revealed that the performance of the constructive OHL-FNN using Hermite polynomials is quite good for both still and moving image compression.
Computer-based recognition of human facial expressions has been an active area of research since the 1970s. The ultimate goal is to realize intelligent man-machine interface. Recently, constructive One-Hidden-Layer Feedforward Neural Networks (OHL-FNNs) have been found promising for facial expression recognition. The hidden units in a FNN usually have the same activation functions typically selected as sigmoidal functions. However, it has not been proven that the use of the same activation functions for all the hidden units is the best or optimal choice in terms of network performance. In this paper, a new constructive polynomial OHL-FNN is proposed for pattern recognition. The well-known Hermite polynomials will be used as activation functions for the hidden units. Each time a new hidden unit is to be added to the network, a Hermite polynomial whose order is increased by one will be used as the activation function of the hidden unit. The proposed technique is applied to the facial expression recognition problem where the 2D DCT is performed over the entire face image before the resulting lower 2D DCT coefficients are fed to the constructive network training. The advantages and limitations of the constructive polynomial OHL-FNN for pattern recognition are also discussed.
The computer-based recognition of facial expressions has been an active area of research for quite a long time. The ultimate goal is to realize intelligent and transparent communications between human beings and machines. The neural network (NN) based recognition methods have been found to be particularly promising, since NN is capable of implementing mapping from the feature space of face images to the facial expression space. However, finding a proper network size has always been a frustrating and time consuming experience for NN developers. In this paper, we propose to use the constructive one-hidden-layer feed forward neural networks (OHL-FNNs) to overcome this problem. The constructive OHL-FNN will obtain in a systematic way a proper network size which is required by the complexity of the problem being considered. Furthermore, the computational cost involved in network training can be considerably reduced when compared to standard back- propagation (BP) based FNNs. In our proposed technique, the 2-dimensional discrete cosine transform (2-D DCT) is applied over the entire difference face image for extracting relevant features for recognition purpose. The lower- frequency 2-D DCT coefficients obtained are then used to train a constructive OHL-FNN. An input-side pruning technique previously proposed by the authors is also incorporated into the constructive OHL-FNN. An input-side pruning technique previously proposed by the authors is also incorporated into the constructive learning process to reduce the network size without sacrificing the performance of the resulting network. The proposed technique is applied to a database consisting of images of 60 men, each having the resulting network. The proposed technique is applied to a database consisting of images of 60 men, each having 5 facial expression images (neutral, smile, anger, sadness, and surprise). Images of 40 men are used for network training, and the remaining images are used for generalization and testing. Confusion matrices calculated in both network training and testing for 4 facial expressions (smile, anger, sadness, and surprise) are used to evaluate the performance of the trained network. By extensive simulations it is shown that when compared with the BP-based method, the proposed technique constructs OHL- FNN with significantly smaller number of hidden units and weights, and simultaneously yielding improved recognition performance.
To date numerous techniques have been proposed to compress digital images to ease their storage and transmission over communication channels. Recently, a number of image compression algorithms using Neural Networks NNs have been developed. Particularly, several constructive feed-forward neural networks FNNs have been proposed by researchers for image compression, and promising results have been reported. At the previous SPIE AeroSense conference 2000, we proposed to use a constructive One-Hidden-Layer Feedforward Neural Network OHL-FNN for compressing digital images. In this paper, we first investigate the generalization capability of the proposed OHL-FNN in the presence of additive noise for network training and/ or generalization. Extensive experimental results for different scenarios are presented. It is revealed that the constructive OHL-FNN is not as robust to additive noise in input image as expected. Next, the constructive OHL-FNN is applied to moving images, video sequences. The first, or other specified frame in a moving image sequence is used to train the network. The remaining moving images that follow are then generalized/compressed by this trained network. Three types of correlation-like criteria measuring the similarity of any two images are introduced. The relationship between the generalization capability of the constructed net and the similarity of images is investigated in some detail. It is shown that the constructive OHL-FNN is promising even for changing images such as those extracted from a football game.
Image compression is an important research domain in image processing. Recently, several neural netowkr (NN) based schemes developed in this are. In particular, constructive feed-forward neural networks have been attempted by many researchers to this problem. The constructive NN-based schemes are promising given their lower training cost, satisfactory performance and automatic determination of proper network size. In this paper, we first consider a NN- based technique that uses a constructive one-hidden-layer FNN for image compression. In standard NN-based schemes when a new hidden unit is added to the net the whole net is retrained while in this scheme the input-side weights are first trained and then all the network output-side weights are adjusted, resulting in a considerably less computational efforts. Next, two pruning techniques are proposed to remove the unnecessary input-side weights during the network construction, without sacrificing the performance of the network, to yield a smaller and a more economical network. To confirm the effectiveness of the prosed techniques, we have applied them to both regression problems and image compression. It has been found that a significant number of weights can be pruned without degenerating the network performance.
This paper presents results on appearance based 3D object recognition accomplished using Independent Component Analysis (ICA). A database of images captured by a ccd camera was used. The workspace was then sampled in a certain manner. Features were extracted from the sampled image using ICA employing information maximization approach reported recently. The features of all the objects thus obtained were saved in a database which formed the workspace manifold. The test images was also represented in a similar manner. Recognition was then performed by locating the closest point in the manifold using radial basis function network, which gave the identity and view (or pose) of the object. The use of ICA, in place of principle component analysis is expected to give a `natural' manifold with maximum significant information with least redundancy.
In cascade-correlation (CC) and constructive one-hidden- layer networks, structural level adaptation is achieved by incorporating new hidden units with identical activation functions one at a time into the active evolutionary net. Functional level adaptation has not received considerable attention, since selecting the activation functions will increase the search space considerably, and a systematic and a rigorous algorithm for accomplishing the search will be required as well. In this paper, we present a new strategy that is applicable to both the fixed structure as well as the constructive network trainings by using different activation functions having hierarchical degrees of nonlinearities, as the constructive learning of a one- hidden-layer feed-forward neural network (FNN) is progressing. Specifically, the orthonormal Hermite polynomials are used as the activation functions of the hidden units, which have certain interesting properties that are beneficial in network training. Simulation results for several noisy regression problems have revealed that our scheme can produce FNNs that generalize much better than one-hidden-layer constructive FNNs with identical sigmoidal activation functions, in particular as applied to rather complicated problems.
It is quite well-known that one-hidden-layer feed-forward neural networks (FNNs) can approximate any continuous function to any desired accuracy as long as enough hidden units are included. Due to this fact many developments in constructive neural networks have been concentrated on only constructive or adaptive one-hidden-layer FNNs. However, this fact does not necessarily imply that one-hidden-layer networks are the most efficient and the best network structure feasible, as one has no explicit guideline to properly select the network structure. Consequently, in practice it has been observed that networks with more than one hidden layer may perform better than the one-hidden- layer networks in some applications. In this paper, we propose a novel strategy for constructing a multi-hidden- layer FNN with regular connections. The new algorithm incorporates in part the policy for adding hidden units from a one-hidden-layer constructive algorithm, and has in part its own new policy for additional layer creation. Extensive simulations are performed for nonlinear noisy regression problems, and it is found that the proposed algorithm converges quite fast and produces networks with one or as many hidden layers/units as required, which are dictated by the complexity of the underlying problem.
This paper deals with adaptive force feedback control of haptic devices that will enable users to interact more realistically with a virtual environment. A system based approach to designing an indirect adaptive output feedback control for a class of single-input single-output nonlinear systems in the explicit input-output differential representation is presented. It is assumed that the zero dynamics associated with the coupled haptic interface/virtual environment is a nonlinear system that is exponentially stable, that is, the system enjoys a minimum- phase property. The proposed nonlinear adaptive controller is implemented using output feedback that can be obtained from the sensors available in the system.
This paper deals with the control design of haptic interfaces for virtual reality-based surgery simulations and training. It is well-known that realistic force reflection and realistic visual representations are among the most stringent requirements that are imposed by surgeons in this application domain. Our goal here is to develop adaptive control strategies that would facilitate the implementation of more realistic man-machine interface for the purpose of virtual reality based surgical simulations.