Relay communications is a promising technique to extend the range of wireless communications by forwarding the message from the sender to the intended destination. While fixed or variable-power relays have been previously investigated, this paper addresses the collaborative use of variable-phase variable-power amplify-and-forward (AF) relays for robust beamforming, with the aid of imperfect channel state information (CSI) at the sender. In particular, the maximization of the worst-case signal-to-noise ratio (SNR) at the destination terminal is studied under a bounded spherical region for the norm of the CSI error vector from the relays to the destination. Our main contribution is that we prove, under a condition on the quality of the estimated CSI, the robust-optimal collaborative-relay beamforming (CRBF) can be obtained by S-Procedure and rank relaxation techniques. In addition, a distributed algorithm is developed by examining the structure of the optimal CRBF solution. Results demonstrate a significant gain of CRBF over non-robust approaches.
In this paper, we present a novel algorithm for representing facial expressions. The algorithm is based on the non-negative matrix factorization (NMF) algorithm, which decomposes the original facial image matrix into two non-negative matrices, namely the coefficient matrix and the basis image matrix. We call the novel algorithm graph-preserving sparse non-negative matrix factorization (GSNMF). GSNMF utilizes both sparse and graph-preserving constraints to achieve a non-negative factorization. The graph-preserving criterion preserves the structure of the original facial images in the embedded sub-space while considering the class information of the facial images. Therefore, GSNMF has more discriminant power than NMF. GSNMF is applied to facial images for the recognition of six basic facial expressions. Our experiments show that GSNMF achieves on average a recognition rate of 93.5% compared to that of discriminant NMF with 91.6%.
In this paper, a novel graph-preserving sparse nonnegative matrix factorization (GSNMF) algorithm is proposed for facial expression recognition. The GSNMF algorithm is derived from the original NMF algorithm by exploiting both sparse and graph-preserving properties. The latter may contain the class information of the samples. Therefore, GSNMF can be conducted as an unsupervised or a supervised dimension reduction method. A sparse representation of the facial images is obtained by minimizing the l(1)-norm of the basis images. Furthermore, according to the graph embedding theory, the neighborhood of the samples is preserved by retaining the graph structure in the mapped space. The GSNMF decomposition transforms the high-dimensional facial expression images into a locality-preserving subspace with sparse representation. To guarantee convergence, we use the projected gradient method to calculate the nonnegative solution of GSNMF. Experiments are conducted on the JAFFE database and the Cohn-Kanade database with unoccluded and partially occluded facial images. The results show that the GSNMF algorithm provides better facial representations and achieves higher recognition rates than nonnegative matrix factorization. Moreover, GSNMF is also more robust to partial occlusions than other tested methods.
Multiview video captured by a multi-camera system has been widely used in many applications such as environmental surveillance, industrial inspection and 3D television. A multiview video contains a vast amount of data required to be transmitted and/or stored, and therefore compression is essential. However, distortion or artifacts are usually witnessed in the reconstructed multiview video after a raw multiview video is compressed at a limited bit-rate. Previous results on video compression mainly focus on the refinements of compression algorithms to improve the quality of reconstructed multiview videos. This paper uses a camera control for video capture to improve the quality of reconstructed multiview videos at a limited bit-rate. Individual cameras are controlled to adjust their pan angles and focal lengths at the video acquisition stage to compensate for the difference in feature-point locations. The optimal pan angle and focal length are designed for each camera. Experimental results validate that the proposed method capturing with camera control outperforms the conventional method capturing without camera control and the heuristic method of 'keeping the object of interest in the image center'.
In this paper stabilization of nonlinear systems with quadratic multi-input is considered. With the help of control Lyapunov function (CLF), a constructive parameterization of controls that globally asymptotically stabilize the system is proposed. Two different cases are considered. Firstly, under certain regularity assumptions. the feasible control set is parameterized, and Continuous feedback stabilizing controls are designed. Then for the general case. piecewise Continuous stabilizing controls are proposed. The design procedure can also be used to verify whether a candidate CLF is indeed a CLF. Several illustrative examples are presented as well.
In this paper, the stabilization of quadratic-input nonlinear systems with bounded controls is considered. According to the type of quadratic-input forms, two cases, namely, positive definite and positive semi-definite, are considered. For the case of positive definiteness, a universal formula for bounded stabilizers is given via a known Lyapunov control function. For the case of positive semidefiniteness, a constructive parametrization of bounded stabilizers is proposed under the assumption that there exists a known Lyapunov control function with respect to a smaller control set than the admissible control set.
In this paper camera pose control for optimizing multiview distributed video coding is considered. The scenario considered is that multiple agents with monocular cameras observe a common scene in a three dimensional world. To get a good video reconstruction under a transmission rate constraint, the camera closest to the center of the camera array is chosen as the reference camera. The poses of all other cameras are controlled and designed such that their images are maximally similar under a constraint on their separation. Based on the rigid motion allowed for the cameras, two cases are considered. For the case where the rigid motion only involves translation, translation control is designed. For the case where it involves both translation and rotation, both controls are constructed. Some simulated results are given to show the efficiency of the designed controllers.
Multiview video captured by multi-camera systems has been widely used in many applications such as surveillance, 3D television, and free viewpoint television. After capturing, the multi-camera system generally requires to compress a large amount of multiview data due to limitations on the transmission capacity. To improve video quality, previous work on video compression focused on algorithm refinements when a raw multiview video is compressed at a limited bit-rate, whereas existing work on instrumentation generally introduced higher resolution capture and additional hardware. In contrast, this paper uses camera control to enhance video quality. Individual cameras are controlled to adjust their pan angles and focal lengths to compensate for both object location difference and color inconsistency among camera views at the video acquisition stage. Such a compensation makes the camera images more similar and therefore improves the video quality when a raw multiview video is compressed at a limited bit-rate. The optimal pan angle and focal length are designed for each camera. An example application is to monitor a potted plant in real time for ornamental horticulture study. Experimental results validate the camera control method.
In this paper, a distributed method for fault detection using sensor networks is proposed. Each sensor communicates only with neighboring nodes to compute locally an estimate of the state of the system to monitor. A residual is defined and suitable stochastic thresholds are designed, allowing to set the parameters so to guarantee a maximum false alarms probability. The main novelty and challenge of the proposed approach consists in addressing the individual correlations between the state, the measurements, and the noise components, thus significantly generalising the estimation methodology compared to previous results. No assumptions on the probability distribution family are needed for the noise variables. Simulation results show the effectiveness of the proposed method, including an extensive sensitivity analysis with respect to fault magnitude and measurement noise.
An algorithm for identification of single-input single-output Box-Jenkins models is presented. It consists of four steps: firstly a high order ARX model is estimated; secondly, the input-output data is filtered with the inverse of the estimated disturbance model; thirdly, the filtered data is used in the Steiglitz-McBride method to recover the system dynamics; in the final step, the noise model is recovered by estimating an ARMA model from the residuals of the third step. The relationship to other identification methods, in particular the refined instrumental-variable method, are elaborated upon. A Monte Carlo simulation study with an oscillatory system is presented and these results are complemented with an industrial case study. The algorithm can easily be generalized to multi-input single-output models with common denominator.
Strict quality of service requirements of industrial applications, challenged by harsh environments and huge interference especially in multi-vendor sites, demand incorporation of cognition in industrial wireless sensor networks (IWSNs). In this paper, a distributed protocol of light complexity for congestion regulation in cognitive IWSNs is proposed to improve the channel utilization while ensuring predetermined performance for specific devices, called primary devices. By sensing the congestion level of a channel with local measurements, a novel congestion control protocol is proposed by which every device decides whether it should continue operating on the channel, or vacate it in case of saturation. Such a protocol dynamically changes the congestion level based on variations of non-stationary wireless environment as well as traffic demands of the devices. The proposed protocol is implemented on STM32W108 chips that offer IEEE 802.15.4 standard communications. Experimental results confirm substantial performance enhancement compared to the original standard, while imposing almost no signaling/computational overhead. In particular, channel utilization is increased by 56% with fairness and delay guarantees. The presented results provide useful insights on low-complexity adaptive congestion control mechanism in IWSNs.
The problem of enhancing the quality of system state estimates is considered for a special class of dynamical systems. Specifically, a system characterized by a discrete-time, finite-state Markov chain state and observed via conditionally Gaussian measurements is assumed. The associated mean vectors and covariance matrices are tightly intertwined with the system state and a control input selected by a controller. Exploiting an innovations approach, finite-dimensional, non-linear approximate MMSE smoothing estimators are derived for the Markov chain system state. The resulting smoothers are driven by a control policy determined by a stochastic dynamic programming algorithm, which minimizes the MSE filtering error, and was proposed in our earlier work. An application of the smoothers derived in this paper is presented for the problem of physical activity detection in wireless body sensing networks, which illustrates the performance enhancement due to smoothing.
One of the main challenges in wireless sensor networks is to prolong the network lifetime by efficiently handling the limited battery life of the nodes. This problem becomes harder in applications where the nodes are randomly dropped in the field. In this paper we deal with the problem of the sink placement and of the network longevity, assuming a number of points in the field with known positions which must be covered by the sensors. Unlike other approaches, we consider the more realistic scenario where the coordinates of the sensors are not assumed to be known in advance and, thus, they cannot be used for the computation of the positions of the sinks. We present two solutions for the above problem; one based on the distance between the points and the second on the probability that a sensor may cover many points. We evaluate our approaches and compare them to algorithms that use the knowledge of the positions of the sensors in order to compute likely sink locations. It is shown that both proposed approaches present similar or better performance concerning network lifetime, while at the same time they significantly decrease the algorithm complexity.
This paper studies minimum-energy packet forwarding policies for communicating sensor measurements from plant to controller over an unreliable multi-hop wireless network so as to guarantee that the optimal controller achieves a prespecified closed-loop performance. For fixed sampling interval, we demonstrate that the minimal linear-quadratic control loss is monotonically decreasing in the reliability of the sensor-to-controller communication. This allows us to decompose the overall design problem into two separate tasks: finding the minimum end-to-end reliability that allows to achieve a prespecified linear-quadratic loss, and developing minimum-energy packet forwarding policies under a deadline-constrained reliability requirement. We develop optimal solutions for both subproblems and show how the co-designed system with minimum forwarding energy cost and guaranteed LQG control performance can be found by a one-dimensional search over admissible sampling periods. The paper ends with a numerical example which demonstrates the effectiveness of the proposed framework.
This paper studies the problem of optimal forwarding for reliable and energy-efficient real-time communication over multi-hop wireless lossy networks. We impose a strict per-packet latency bound and develop forwarding policies that maximize the probability that the packet is delivered within the specified deadline minus a transmission energy cost. A solution to this problem allows to characterize the set of achievable latency-reliability pairs and to trace out the Pareto frontier between achievable deadline-constrained reliability and transmission energy cost. We develop dynamic programming-based solutions under a finite-state Markov channel model. Particular instances with Bernoulli and Gilbert-Elliot loss models that admit numerically efficient solutions are discussed and our results are demonstrated on several examples.
In this paper, we investigate strategies for radio power control for wireless sensor networks that guarantee a desired packet error probability. Efcient power control algorithms are of major concern for these networks, not only because the power consumption can be signicantly decreased but also because the interference can be reduced, allowing for higher throughput. An analytical model of the Received Signal Strength Indicator (RSSI), which is link quality metric, is proposed. The model relates the RSSI to the Signal to Interference plus Noise Ratio (SINR), and thus provides a connection between the powers and the packet error probability. Two power control mechanisms are studied: a Multiplicative-Increase Additive-Decrease (MIAD) power control described by a Markov chain, and a power control based on the average packet error rate. A component-based software implementation using the Contiki operating system is provided for both the power control mechanisms. Experimental results are reported for a test-bed with Telos motes.
As a promising technology, wireless sensor networks have a wide range of applications. However, the development of wireless sensor networks is stillin face of multiple challenges. Among these challenges, the privacy issue is acritical parameter involved in providing secure and reliable services and attracts much attention from researchers and engineers. In the last decade, alarge number of privacy solutions for wireless sensor networks have been proposed. However, most of them are taken as additional secure functionality blocks rather than being integrated in the original sensor network designs.
In this thesis project, we will focus on the privacy assessment of a parallel distributed detection network, which represents a simplied physical-layer of wireless sensor networks. The security threat is assumed to come from a passive eavesdropper. Four privacy leakage criteria are proposed to evaluate the privacy issue of the distributed detection network in dierent scenarios. As references, the privacy leakages are evaluated by dierent criteria when the distributed detection system is optimized in the perspectives of Bayesian detection theory and information theory without considering the presence of the eavesdropper. Then, we propose the corresponding privacy-concerned distributed detection systems. Comparisons to the optimal detection systems are performed and they reveal the trade-off between privacy leakage suppression and detection performance degradation.
Component-based techniques revolve around composable, reusable software objects that shield the application level software from the details of the hardware and low-level software implementation and vice versa. Components provide many benefits that have led to their wide adoption it software and middleware developed for embedded systems: They are well-defined entities that can be replaced without affecting the rest of the systems, they can be developed and tested separately and integrated later, and they are reusable. Clearly such features are important for the design of large-scale complex systems more generally, beyond software architectures. We propose the use of a component approach to address embedded control problems. We outline a general coponent-based framework to embedded control aid show how it can be instantiated inspecific problems that arise in the control over/of sensor networks. Building on the middleware component framework developed under the European project RUNES, we develop a number of control-oriented components necessary for the implementation of control applications and design their integration. The paper provides the overview of the approach, discusses a real life application where the approach has been tested and outlines a number of specific control problems that arise in this application.
Wireless networked embedded systems are becoming increasingly important in a wide area of technical fields. In this tutorial paper we present recent results on the design of these systems and their use in control applications, that have been developed within the project Reconfigurable Ubiquitous Networked Embedded Systems (RUNES). RUNES is a European Integrated Project with the aim to control complexity in networked embedded systems by developing robust and scalable middleware systems. New components for control under varying network conditions are discussed for the RUNES architecture. The paper highlights how the complexity of the closed-loop system is increased, due to additional disturbances introduced by the communication system: additional delays, jitter, data rate limitations, packet losses, etc. Experimental work on integration test beds that demonstrates these results is presented, together with motivating links to the RUNES disaster relief tunnel scenario.
We propose a control law for stabilization of a quadrotor-load system, and provide conditions on the control law's gains that guarantee exponential stability of the equilibrium. The system is composed of a load and an unmanned aerial vehicle (UAV) attached to each other by a cable of fixed length, which behaves as a rigid link under tensile forces; and the control input is composed of a three dimensional force requested to the UAV, which the UAV provides with or without delay. Given the proposed control law, we analyze the stability of the equilibrium in two separate parts. In the first, the system is modeled assuming that the UAV provides the requested control input without delay, and we verify that the equilibrium is exponentially stable. In the second part, the UAV is modeled as possessing an attitude inner loop, and we provide a lower bound on the attitude gain for which exponential stability of the equilibrium is preserved. An integral action term is also included in the control law, which compensates for battery drainage or model mismatches, such as an unknown load mass. We present experiments for different scenarios that demonstrate and validate the robustness of the proposed control law.
This paper presents a formulation to identify the input-output maps of discrete- time bilinear state-space models. The new method uses input-output data from one experiment, and the input data need not be of special types such as extended pulses. The initial conditions can be non-zero and unknown. Recent methods, on the other hand, require data from multiple experiments, specialized input signals, and zero initial conditions. In the present method, a Nonlinear Auto-Regressive model with eXogenous input (NARX) is identified and used to predict the bilinear system response directly without returning to the bilinear state-space format. Numerical examples are provided to illustrate the identification method.