Electrical and Electronic Engineering - Research Publications

Permanent URI for this collection

Search Results

Now showing 1 - 10 of 38
  • Item
    Thumbnail Image
    A robust algorithm for foreground extraction in crowded scenes
    Rao, AS ; Gubbi, J ; Marusic, S ; Palaniswami, M (IEEE, 2012-12-01)
    The widespread availability of surveillance cameras and digital technology has improved video based security measures in public places. Surveillance systems have been assisting officials both in civil and military applications. It is helping to identify unlawful activities by means of uninterrupted transmission of surveillance videos. By this, the system is adding extraneous onus on to the already existing workload of security officers. Instead, if the surveillance system is intelligent and efficient enough to identify the events of interest and alert the officers, it alleviates the burden of continuous monitoring. In other words, our existing surveillance systems are lacking to identify the objects that are dissimilar in shape, size, and color especially in identifying human beings (nonrigid motions). Global illumination changes, frequent occurrences of shadows, insufficient lighting conditions, unique properties of slow and fast moving objects, unforeseen appearance of objects and its behavior, availability of system memory, etc., may be ascribed to the limitations of existing systems. In this paper, we present a filtering technique to extract foreground information, which uses RGB component and chrominance channels to neutralize the effects of nonuniform illumination, remove shadows, and detect both slow-moving and distant objects.
  • Item
    Thumbnail Image
    A UNIFYING FRAMEWORK FOR ANALYSIS AND DESIGN OF EXTREMUM SEEKING CONTROLLERS
    Nesic, D ; Tan, Y ; Manzie, C ; Mohammadi, A ; Moase, W (IEEE, 2012-01-01)
    We summarize a unifying design approach to continuous-time extremum seeking that was recently reported by the authors. This approach is based on a feedback control paradigm that was to the best of our knowledge explicitly summarized for the first time in this form in our recent work. This paradigm covers some existing extremum seeking schemes, provides a direct link to off-line optimization and can be used as a unifying framework for design of novel extremum seeking schemes. Moreover, we show that other extremum seeking problem formulations can be interpreted using this unifying viewpoint. We believe that this unifying view will be invaluable to systematically design and analyze extremum seeking controllers in various settings.
  • Item
    Thumbnail Image
    On a Shubert Algorithm-Based Global Extremum Seeking Scheme
    Nesic, D ; Nguyen, T ; Tan, Y ; Manzie, C (IEEE, 2012)
    This paper adapts the so-called Shubert algorithm for Extremum Seeking Control (ESC) to seek the global extremum (in presence of local extrema) of general dynamic plants. Different from derivative based methods that are widely used in ESC, the Shubert algorithm is a good representative of sampling optimization methods. With knowledge of the Lipschitz constant of an unknown static mapping, this deterministic algorithm seeks the global extremum. By introducing “waiting time” the proposed Shubert algorithm-based global extremum seeking guarantees the semi-global practical convergence (in the initial states) to the global extremum if compact sets of inputs are considered. Several numerical examples demonstrate how proposed method may be successfully deployed.
  • Item
    Thumbnail Image
    Small-gain theorems of LaSalle type for hybrid systems
    Liberzon, D ; Nesic, D ; Teel, AR (IEEE, 2012)
    We study stability of hybrid systems described as feedback interconnections of smaller subsystems, within a Lyapunov-based ISS small-gain analysis framework. We focus on constructing a weak (nonstrictly decreasing) Lyapunov function for the overall hybrid system from weak ISS-Lyapunov functions for the subsystems in the interconnection. Asymptotic stability of the hybrid system is then concluded by applying results of LaSalle type. The utility of this approach is illustrated on feedback systems arising in event-triggered control and quantized control.
  • Item
    Thumbnail Image
    Averaging in singularly perturbed hybrid systems with hybrid boundary layer systems
    Wang, W ; Teel, AR ; Nesic, D (IEEE, 2012-01-01)
    We analyze a class of singularly perturbed hybrid systems based on two auxiliary hybrid systems: the averaged system, which approximates the slow dynamics, and the boundary layer system, which approximates the fast dynamics. The average system is generated by averaging the solutions of the boundary layer system. The novelty of this work is that the boundary layer system is a hybrid system rather than a continuous-time system. This extends available results to cover new classes of hybrid systems. We illustrate how to apply our results through an example that is a power converter system under hybrid feedbacks implemented by pulse-width modulation (PWM).
  • Item
    Thumbnail Image
    Emulation-based tracking solutions for nonlinear networked control systems
    Postoyan, R ; van de Wouw, N ; Nesic, D ; Heemels, WPMH (IEEE, 2012-01-01)
    We investigate emulation-based tracking control for nonlinear networked control systems (NCS) affected by disturbances. We consider a general scenario in which the network is used to ensure the communication between the controller, the plant and the reference system generating the desired trajectory to be tracked. The communication constraints induce nonvanishing errors (in general) on the feedforward term and the output of the reference system. These network-induced errors affect the convergence of the tracking error. As a consequence, available results on the stabilization of equilibrium points for NCS are not applicable. Therefore, we develop an appropriate hybrid model and we give sufficient conditions on the closed-loop system, the communication protocol and an explicit bound on the maximum allowable transmission interval (MATI) guaranteeing that the tracking error converges to the origin up to some errors due to both the external disturbances and the aforementioned non-vanishing network-induced errors. Our results cover a large class of the so-called uniformly globally asymptotically stable protocols which include the well-known round-robin and try-once-discard protocols. We also introduce a new dynamic protocol suitable for tracking control.
  • Item
    Thumbnail Image
    Extremum-seeking control for periodic steady-state response optimization
    van de Wouw, N ; Haring, M ; Nesic, D (IEEE, 2012-01-01)
    Extremum-seeking control is a powerful adaptive technique to optimize steady-state system performance. To this date, extremum-seeking control has mainly been used to optimize plants with constant steady-state outputs, whereas the case in which the steady-state outputs are time varying, has received less attention. We propose an extremum-seeking scheme for the optimization of nonlinear plants with periodic steady-state outputs. Extremum-seeking control in this setting is relevant in e.g. the scope of tracking and disturbance rejection problems. We show that under certain assumptions the proposed extremum-seeking controller design guarantees that for an arbitrarily large set of initial conditions the steady-state performance of the plant converges arbitrarily close to its optimal value.
  • Item
    Thumbnail Image
    Parameter and state estimation for a class of neural mass models
    Postoyan, R ; Chong, M ; Nesic, D ; Kuhlmann, L (IEEE, 2012-01-01)
    We present an adaptive observer which asymptotically reconstructs the parameters and states of a model of interconnected cortical columns. Our study is motivated by the fact that the considered model is able to realistically reproduce patterns seen on (intracranial) electroencephalograms (EEG) by varying its parameters. Therefore, by estimating its parameters and states, we could gain a better understanding of the mechanisms underlying neurological phenomena such as seizures, which might lead to the prediction of the onsets of epileptic seizures. Simulations are performed to illustrate our results.
  • Item
    Thumbnail Image
    Semiglobal Practical Stability of a Class of Parameterized Networked Control Systems
    Wang, B ; Nesic, D (IEEE, 2012-01-01)
    This paper studies a class of parameterized networked control systems that are designed via an emulation procedure. In the first step, a controller is designed ignoring network so that semiglobal practical stability is achieved for the closed-loop. In the second step, it is shown that if the same controller is emulated and implemented over a large class of networks, then the networked control system is also semiglobally practically asymptotically stable; in this case, the controller parameter needs to be sufficiently small and communication bandwidth sufficiently high.
  • Item
    Thumbnail Image
    Distributed multi static sonar tracking in clutter
    Mušicki, D ; Nešić, D ; Song, TL (IEEE, 2012-10-24)