Search

This paper considers the optimal control of a linear scalar model with a finite communication rate between the sensor and the controller. We analyze the optimization of the quantizer and the controller where the latter utilizes the overall history of the received symbols to determine the control input. Making use of a tree structure representation, it is shown that the resulting optimal control problem can be reduced to a combined quantization and constrained quadratic minimization problem. We characterize the necessary conditions for the optimal control and develop numerical algorithms. A localized computational method for long time horizon is also discussed.

Active sonar systems exhibit high false alarm rates that reduce their usefulness in target detection. An idea to reduce false alarm rates is to track the hypothetical targets from the sonar system and reject false detections that do not generate a reasonable track. The IPDA algorithm includes a procedure for estimating the probability a track is a true track. This estimate is investigated as a detection statistic. The use of IPDA ability is compared to a single pulse detector and to a multistatic active sonar system using IPDA to combine measurements from different sensors as well as different scans. © 2005 IEEE.

In a cluttered environment, measurements originate not only from the objects being tracked, but also from spurious sources. The number of existing objects (targets) to be tracked is also unknown. The Multi-Hypotheses Tracking (MHT) filter is generally considered to be the "optimal" n-scan target tracking filter in such a multi-target cluttered environment. This paper introduces a new MHT filter based on the target existence. The probability that a target exists is estimated, and the estimate of the target kinematic state is subsequently conditioned on target existence. False track discrimination uses the probability of target existence as the track quality measure. This approach simplifies the structure of the algorithm and reduces complexity. A simplified version of the algorithm, which decouples new track hypotheses from the measurement-to-existing-track allocations is also presented. A simulation study shows the effectiveness of this approach in an environment of heavy and non-uniform clutter, with multiple maneuvering targets with crossing trajectories.

The significance of parasitic impedances of the connecting structures in power electronic converters (PCB interconnects) increases with frequency and switching speed and influences transients and radiated EMI. Design aspects such as physical layout and interconnect routing requires accurate extraction of parasitic parameters. This paper presents a computational procedure for partly achieving this for typical power converter systems.

Advances in technology are making distributed networks of sensor platforms a reality for surveillance systems. Such networks provide many benefits including the ability to cover a broader area, to compute more accurate target estimates and provide a reduction in susceptibility to countermeasures. In spite of these advantages, many practical issues remain in the deployment of such networks. One such significant issue is bandwidth limitations on the communication channels between the sensor platforms. This paper examines the effect of bandwidth restrictions on a decentralised system architecture. In such a structure, each sensor tracks targets within its individual surveillance region and then transmits its track data to a global fusion centre. The minimum necessary data rates required to permit reconstruction of the global track estimates to a given level of accuracy are presented in the case when the track state can be represented by a Gauss-Markov system.

Sensor scheduling has been a topic of interest to the target tracking community for some years now. Recently, research into it has enjoyed fresh impetus with the current importance and popularity of applications in Sensor Networks and Robotics. The sensor scheduling problem can be formulated as a controlled Hidden Markov Model. In this paper, we address precisely this problem and consider the case in which the state, observation and action spaces are continuous valued vectors. This general case is important as it is the natural framework for many applications. We present a novel simulation-based method that uses a stochastic gradient algorithm to find optimal actions.1

Since the release of IEEE Std. 1159.3 recommend practice for the transfer of power quality data, vendors of power quality monitoring equipment have been interested in testing their implementations with third parties to determine if the implementation is compliant with the standard. This summary paper and the accompanying presentation describes the findings for a test conducted for the PMI PMR600 power meter

In this paper we measure and characterize 55-65 GHz wire-less channels for a typical desktop environment. An Angle of Arrival (AoA) Modified Saleh-Valenzuela (S-V) model is used for the desktop environment. Key AoA Modified S-V model parameters such as Cluster Decay Factor, Ray Decay Factor, Cluster Arrival Rate, Ray Arrival Rate and AoA Ray angular standard deviation are extracted from the measured data.

Objective—To investigate factors influencing screw loosening after triple pelvic osteotomy (TPO) and ischial wire stabilization of the acetabular segment. Design—Retrospective case series. Animals—227 dogs with congenital hip dysplasia or subluxated hip joints. Procedures—Medical records and radiographs of 227 dogs that underwent 332 TPO procedures were evaluated, and data pertaining to screw type, plate position, sacral screw engagement, use of ischial interfragmentary wires, and pelvic alignment were assessed for associations with screw loosening. Results—Complications developed in 96 of the 332 (29%) procedures. Cancellous screws without sacral engagement were associated with the lowest frequency (6%) of loosening, compared with cancellous and cortical screws engaging the sacrum and cortical screws that did not engage the sacrum. Frequency of screw loosening increased when cortical or cancellous screws engaged the sacrum and when cortical screws were used. In dogs that had surgery bilaterally, the first limb on which TPO was performed had a higher frequency of screw loosening than the second limb. Pelvic alignment loss was greatest (5.4°) when the 3 most cranial screws were loosened. Loss of pelvic alignment was significantly different between dogs that underwent surgery and had complications and those that underwent surgery and did not have complications in association with loosening of 1, 2, and 3 screws. Conclusions and Clinical Relevance—TPO screw loosening was multifactorial and related to stability of the affected ilium, screw type, and screw position. Placing cancellous screws that do not engage the sacrum in pelvic osteotomy plate positions 1 through 3 may decrease the number of screws that loosen.

A survey of the field related to control of chaotic systems is presented. Several major branches of research that are discussed are feed-forward (‘non-feedback’) control (based on periodic excitation of the system), the ‘Ott–Grebogi–Yorke method’ (based on the linearization of the Poincaré map), the ‘Pyragas method’ (based on a time-delayed feedback), traditional for control-engineering methods including linear, nonlinear and adaptive control. Other areas of research such as control of distributed (spatio-temporal and delayed) systems, chaotic mixing are outlined. Applications to control of chaotic mechanical systems are discussed.

In this paper target tracking using measurements whose probability density function can be described (approximated) by a Gaussian Sum Mixture is described. The approach is illustrated for two classes of these measurements. One class is the measurements obtained by acoustic amplitude / time difference of arrival measurements typically obtained using motes. The other is measurements obtained by radar with large angular measurement errors. Such measurements in a Cartesian system cannot be accurately modelled by a single Gaussian pdf. By following a Bayessian approach and the target existence paradigm, two target tracking filters have been derived. A variant of IPDA - the single scan target tracking filter, and a variant of the ITS - the multi scan target tracking filter, have been derived for situations when Gaussian sum measurements are available.

In this paper we extend the idea of adaptive waveform selection for radar target tracking to interacting multiple model (IMM) trackers to permit the modelling of maneuvering targets by allowing multiple possible dynamical models. We develop a one step ahead solution to the problem of waveform selection, which is designed to decrease dynamic model uncertainty for the target of interest. It is based on maximization of the expected information obtained about the dynamical model of the target from the next measurement. We also discuss the design of waveform libraries for target tracking applications.

In this paper, we explore the application of signal processing methods to the design of optical monitoring equipment for use in optical networks. In particular, we study inverse signal processing to enhance the performance of optical power spectrum estimation, based on low finesse (low Q) tunable optical filters. We consider both batch and recursive methods. Simulation results confirm that accurate power spectrum estimates can be achieved using the proposed techniques.

A new version of BPM electronics based on the AD8362 RMS detector, which is a direct RF to DC converter, is under development at the JLAB FEL. Each of these new BPM electronics utilizes an embedded ColdFire Microprocessor for data processing and communication with the EPICS control system via TCP/IP. The ColdFire runs RTEMS, which is an open source real‐time operating system. The JLAB FEL is a SRF Energy Recovery LINAC capable of running up to 10 mA CW beam with a 74.85 MHz micropulse frequency. For diagnostic reasons and for machine tune up, the micropulse frequency can be reduced to 1.17 MHz, which corresponds to about 160 μA of beam current. It is required that the BPM system would be functional for all micropulse frequencies. By taking into account the headroom for the beam steering and current variations the dynamic range of the RF front end is required to be about 60 dB. A BPM resolution of at least 100 μm is required, whereas better resolution is very desirable to make it possible for more accurate measurements of the electron beam optics. Some results of the RF front end development are presented as well as the first measurements made with an electron beam.

In this paper we consider two aspects of adaptive control. Firstly we assume that the unknown plant is linear and examine the effect of a strict inequality constraint on the control. Secondly we study the convergence properties of restricted complexity adaptive controllers when applied to nonlinear systems. Basically we are able to show that if the controller is capable of satisfactorily controlling the system then the adaptive version will find a satisfactory controller.

Presents an adaptive waveform selective probabilistic data association (WSPDA) algorithm for tracking a single target in clutter. The assumption of an optimal receiver allows the inclusion of transmitted waveform specification parameters in the tracking subsystem equations, leading to a waveform selection scheme where the next transmitted waveform parameters are selected to minimise the total mean-square tracking error at the next time step. A simulation example demonstrates the potential of WSPDA tracking filters over conventional probabilistic data association (PDA) tracking filters.

Optimal filtering, prediction and smoothing algorithms for hidden Markov models (HMMs) are presented. Employing a dynamic state space description of the probability evolution for an HMM, the close structural similarity with Kalman filtering, prediction and smoothing is shown. The HMM estimation algorithms are simpler than those of the Kalman filter. >

A version of the probabilistic data association (PDA) algorithm that is called integrated probabilistic data association (IPDA) is presented. Rather than deriving algorithms based on an assumption of track existence, IPDA starts from expressions which treat track existence as an event with an associated probability. This approach leads to a set of recursive formulae for both data association and probability of track existence or track quality. This is useful in the track initiation phase (to confirm a tentative track) and in track termination (to terminate a false track). The IPDA algorithm is compared with both the probabilistic data association algorithm (PDA) and the interactive multiple model (IMM) PDA algorithm. >

The Multiple Hypotheses Tracking (MHT) algorithm has been shown to have the best tracking performance among existing multi-target tracking algorithms using real world sensors with probability of detection less than unity and in the presence of false alarms. The improved performance of the Multiple Hypotheses Tracking comes at the cost of signicantly higher computational complexity. Most Multiple Hypotheses Tracking implementations only form the best global hypothesis. This paper compares the Linear Multitarget Integrated Track Splitting (LMITS) tracking algorithm with the Multiple Hypotheses Tracking algorithm. LMITS has a simpler structure than Multiple Hypotheses Tracking as it decouples local hypotheses and avoids the measurement to multi-track allocation entirely. The number of LMITS hypotheses equals the sum of the number of local hypotheses added to the number of initiation hypotheses. Thus LMITS can retain a deeper hypotheses subtree which can result in better performance. We compare tracking performances of LMITS and MHT algorithms using simulated data for multiple maneuvering targets in heavy and non-uniform clutter.

Multiple input and multiple output (MIMO) is a new technology that promises to significantly increase the capacity of wireless communication systems. MIMO technology increases capacity by exploiting the multipath channel. Unfortunately results presented in the literature (Gesbert et al., Smith et al.) neglect antenna effects. This paper investigates the effect that parameters such as number, spacing, coupling and far field antenna patterns of antennas have on the capacity of MIMO systems in Rayleigh channels.

In this paper, we investigate the performance of a set based tracker in clutter with no target present. If there is a positive probability of clutter-only tracks lasting an infinite time in such an environment, then a tracker cannot be expected to perform well under these clutter conditions when a target is present. In order to analyze the problem, a continuous space model is simplified and discretised. It is shown that this discretised model is a percolation process, which undergoes a phase transition. This means that there exists a critical false alarm probability for the system, below which clutter tracks die out infinite time, and above which clutter tracks have a positive probability of surviving indefinitely. A theoretical lower bound for this critical clutter probability is derived, which is then verified by Monte Carlo simulation.

We will present recent research activities in optical networking and photonic systems carried out at the two research centres at the University of Melbourne.

This paper considers the problem of data incest management in decentralized estimation networks. All nodes consist of both a measurement sensor and a local fusion centre. Global state information is exchanged across communications channels with bounded variable delays. Particular attention is devoted to the underlying delay structure that causes data incest. The investigation employs an efficient graph theoretic representation of the time varying nature of these delays. Applying this framework a data incest management algorithm is presented and is complimented with a tool for identifying the necessary and sufficient condition for its use.

A fully integrated Machine Protection System (MPS) is critical to efficient commissioning and safe operation of all high current accelerators. The Jefferson Lab PEL has multiple electron beam paths and many different types of diagnostic insertion devices. The MPS needs to monitor both the status of these devices and the magnet settings which define the beam path. The matrix of these devices and beam paths are programmed into gate arrays, the output of the matrix is an allowable maximum average power limit. This power limit is enforced by the drive laser for the photocathode gun. The Beam Loss Monitors (BLMs), RF status, and laser safety system status are also inputs to the control matrix. There are 8 Machine Modes (electron path) and 8 Beam Modes (average power limits) that define the safe operating limits for the FEL. Combinations outside of this matrix are unsafe and the beam is inhibited. The power limits range from no beam to 2 megawatts of electron beam power. The timing requirements for the BLMs and RF is /spl sim/1 microsecond and for the insertion devices (viewers) is /spl sim/1 millisecond. The entire system is VME based and EPICS compatible. This paper describes the system as it is installed on the JLab 10 kW FEL.

A maneuvering target is to be tracked based on noise corrupted measurements of the target's state that are received by a moving observer. Additionally, the quality of the target state observations can be improved by the appropriate positioning of the observer relative to the target during tracking. The bearings-only tracking problem is an example of this scenario. The question of optimal observer trajectory planning naturally arises, i.e. how should the observer maneuver relative to the target in order to optimise the tracking performance? In this paper, we formulate this problem as a discrete-time stochastic optimal control problem and present a novel stochastic approximation algorithm for designing the observer trajectory. Numerical examples are presented to demonstrate the utility of the proposed methodology.

This paper presents a new single-target tracking filter. Similar in structure to track-oriented MHT, the Integrated track splitting (ITS) filter is a multi-scan tracking algorithm that, like IPDA, integrates target state estimation with the estimation of target existence. The ITS filter models each track as a set of components, where each component is defined with a unique measurement history which consists of zero or one measurement received each scan. For each component the state estimate and the a-posteriori probability of component existence are computed recursively. After each scan, a new component is formed from each pair of "existing component, associated measurement". The probability of the new component existence is the probability that the parent component exists and that the measurement used to create the new component is the target measurement. The probability of target existence, mean and covariance of the state estimate for the track are then calculated and used for track maintenance and track output. ITS-MAP filter uses a-priori clutter density information provided by a clutter map, analytically or by some other means to better discriminate clutter from target measurements. Simulations are be used to verify the performance of the ITS-MAP algorithm and to compare its performance with that of other target tracking algorithms in a dense and non-homogeneous clutter environment.

Modern phased array radars are able to adaptively modify their performance to the environment. To make full use of this capability, scheduling algorithms need to be designed. This paper poses the problem of adaptive waveform scheduling for detecting new targets in the context of finite horizon stochastic dynamic programming. The result is a scheduling algorithm that minimises the time taken to detect new targets, detecting these targets in accordance with importance, while minimising the use of radar resources.

This paper considers the following question: given a deterministic linear plant with multiple sensors and controllers, how can one characterise the combination of channel data rates between all sensor-controller pairs which permits stability to be achieved? Under different stability objectives, two sets of data rate inequalities which must be satisfied by any stabilising decentralised coding and control scheme are presented. It is then shown that if another set of inequalities is feasible then a stabilising decentralised coding and control scheme can always be constructed. Finally, a tight data rate inequality for stabilisability is obtained for a special class of decentralised plant.

We report initial lasing results from the IR Upgrade FEL at Jefferson Lab (Proceedings: 2001 Particle Accelerator Conference, IEEE, Piscataway, NJ, 2001). The electron accelerator was operated with low average current beam at 80 MeV. The time structure of the beam was 120 pC bunches at 4.678MHz with up to 750 µs pulses at 2 Hz. Lasing was established over the entire wavelength range of the mirrors (5.5-6.6 µm). The detuning curve length, turn-on time, and power were in agreement with modeling results assuming a 1 ps FWHM micropulse. The same model predicts over 10kW of power output with 10mA of beam and 10% output coupling, which is the ultimate design goal of the IR Upgrade FEL. The behavior of the laser while the dispersion section strength was varied was found to qualitatively match predictions. Initial CW lasing results also will be presented.

This paper poses a simple question: what is the lowest rate, in bits per unit time, at which feedback information can be transmitted in order to stabilise a given dynamical system? Expressions for this fundamental quantity have recently been derived for linear systems, with and without noise. In this work, the case of deterministic, fully observed, continuously differentiable dynamical systems is investigated, under the additional assumptions of controllability to the desired set-point and bounded initial states. By the use of volume-partitioning arguments and local Jordan forms, the infimum feedback data rate is shown to be the base-2 logarithm of the magnitude of the determinant of the open-loop Jacobian on the local unstable subspace, evaluated at the set-point. Connections to the concept of local topological feedback entropy are briefly discussed.

Presents an approach for multi-target tracking. In multi-target situations, multiple tracks may share the same measurement(s). Rather than forming joint events by creating all possible combinations of track-measurement assignments, only a single track is processed at a time. Therefore, the number of operations is linear in the number of tracks. The a-priori probabilities of measurement origin are the conduit for inter-track information transfer. The LJIPDA algorithm follows three steps. In the first step, the a-priori probabilities of measurement origin are established for all tracks and all measurements. The next step is measurement oriented and corrects these probabilities. The final step is a modified IPDA algorithm, which uses visibility of measurements to the tracks to calculate the probability of target existence and the data-association probabilities. These probabilities allow track update in the classic PDA fashion, as well as automatic track initiation, maintenance and termination. Simulations are used to verify the performance of the LJIPDA algorithm and compare it with the performance of the IPDA and JIPDA algorithms in a dense and non-homogenous clutter environment, in crossing target situations.

A fundamental question in the field of communication-limited control is how low the closed loop data rate can be made before a given dynamical system is impossible to stabilise by any coding and control law. Analogous to the role of entropy in Shannon source coding, this number defines the smallest data rate sufficient to achieve 'reliable' closed loop performance. The objective here is to analyse this quantity for a general, finite-dimensional, discrete-time stochastic linear plant. By inductive arguments employing the entropy power inequality of information theory, an explicit expression for the infimum mean-square-stabilising data rate is derived, under very mild conditions on the initial state and noise probability densities.

Target tracking algorithms are often thought of as a conglomeration of a number of unrelated techniques and tricks. However, they all have firm foundations in recursive Bayesian logic. In this paper, we review the Bayesian underpinnings of several target tracking algorithms and highlight the assumptions under which appropriate Bayesian recursions are derived and used.

This paper presents a new algorithm for multi-target tracking. In multi-target situations, multiple tracks may share the same measurement(s). Joint events are formed by creating all possible combinations of track-measurement assignments. The probabilities for these joint events are calculated The expressions for the joint events incorporate the probabilities of track existence of individual tracks, as well as an efficient approximation for the cluster volume and an a-priori probability of the number of clutter measurements in each cluster. From these probabilities the data association and track existence probabilities of individual tracks are obtained These probabilities will allow track update in the classic PDA fashion, as well as automatic track initiation, maintenance and termination. The JIPDA algorithm is recursive and integrates seamlessly with the IPDA algorithm. Simulations are used to verify the performance of the algorithm and compare it with the per performance of the IPDA, IPDA-DLL and IJPDA algorithms in a dense and non-homogenous clutter environment, in crossing target situations.

In this paper we present two new results on Bode type integrals. Firstly we obtain, for a given scalar or multivariable continuous-time plant, the infimum of the Bode sensitivity integral which can be obtained with any stabilizing controller. The result involves the unstable plant poles and, perhaps surprisingly, a subset of the plant nonminimum phase zeros. Secondly, we obtain an apparently new expression for the Bode integral for the complementary sensitivity for a stable discrete-time scalar system.

We derive optimal and sub-optimal beam scheduling algorithms for electronically scanned array tracking systems. We formulate the scheduling problem as a multi-arm bandit problem involving hidden Markov models. A finite dimensional optimal solution to this multi-arm bandit problem is presented. The key to solving any multi-arm bandit problem is to compute the Gittins index. We present, a finite dimensional algorithm which computes the Gittins index. Suboptimal algorithms for computing the Gittins index are also presented.

The authors provide some unification of the results on various types of periodic and multirate sampling controllers by showing that properties of these controllers can be achieved by a single type of controller, i.e. the discrete periodic controller. The wideband and high-gain-like characteristic of this controller is addressed, and ways in which it may deteriorate system performance are described. >

The authors introduce the concept of iterative adaptive control and formulate an iterative adaptive control algorithm. The controller iterations must satisfy certain convergence and continuity assumptions for the closed-loop system stability, but only one step controller iteration is required for each new parameter update. This results in a considerable reduction in real-time computational load. The algorithm is presented in a general form and then applied to an adaptive LQ (linear quadratic) regulator. >

This paper describes a noncoherent ultrasonic array used to form three-dimensional images of defects in metal. The problem developed in terms of deconvolution in multiple dimensions to improve resolution of images blurred by the measuring system and degraded by noise is reduced to solution of a linear equation of the form y=Hx, where H is called the imaging operator H may be separated into the Kronecker product of smaller banded-Toeplitz matrices V(X)S(X)P. This structure is used to develop an algorithm to solve for X using least squares and singular value decomposition. >

Design of a sampled-data state feedback controller for stabilizing a nonlinear continuous-time plant is studied. The sampled-data state feedback controller is designed based on an approximate discrete-time model of the plant and it is able to exponentially stabilize the approximate discrete-time model. When the sampled-data controller applies to the continuous-time plant, the closed loop system is a hybrid system. This paper gives a sufficient condition under which the sampled-data controller exponentially stabilizes the hybrid system consisting of the continuous-time plant and the sampled-data controller.

A fully integrated Machine Protection System (MPS) is critical to efficient commissioning and safe operation of all high-current accelerators. The MPS needs to monitor the status of all devices that could enter the beam path, the beam loss monitors (BLMs), magnet settings, beam dump status, etc. This information is then presented to the electron source controller, which must limit the beam power or shut down the beam completely. The MPS for the energy recovery linac (ERL) at the Jefferson Lab Free Electron Laser [1] generates eight different power limits, or beam modes, which are passed to the drive laser pulse controller (DLPC) (photocathode source controller). These range from no beam to nearly 2 megawatts of electron beam power. Automatic masking is used for the BLMs during low-power modes when one might be using beam viewers. The system also reviews the setup for the two different beamlines, the IR path or the UV path, and will allow or disallow operations based on magnet settings and valve positions. This paper will describe the approach taken for the JLab 10-kW FEL. Additional details can be found on our website http://laser.jlab.org [2].

The increasing integration of power electronic circuit modules together with the continuing growth in power density, switching speed and frequency have resulted in a close coupling between electromagnetic, thermal and mechanical components and interconnections and a consequent increase in undesirable side effects. The switching capability of modern semiconductor devices (MOSFET, IGBT, MCT etc) results in voltage and current variations (dV/dt, dl/dt) reaching values in the range of V/ns and an A/ns. These large transient phenomena induce parasitic currents and radiated emissions. Computational and numerical approaches hold the promise of accurately predicting these undesirable phenomena.

An Out of Sequence Measurements (OOSM) based approach to decentralized fusion is pre- sented with specific application to multi-sensor target tracking. A full information network architecture is considered where a fixed delay exists between nodes. The communication between nodes is a fused posterior density whose optimal conditioning set contains com- plete information on individual measurements from all networked sensors.

This paper considers the problem of forming and maintaining tracks when measurements have both uncertain origin and are limited by finite resolution. The integrated probabilistic data association-finite resolution (IPDA-FR) algorithm is presented for the no a-priori clutter information situation. The algorithm consists of a set of recursive formulae for data association and probability of track existence, thus enabling automatic track initiation and track maintenance. Simulation results comparing the performance of IPDA-FR with the IPDA algorithm are presented. >

Uncertain nonlinear systems affine in the input are considered. It is shown that under some matching conditions these systems can be made practically stable by state feedback. We present an adaptive hybrid algorithm which renders the origin practically stable to any desired accuracy. The presented algorithms are simple and computationally cheap.

This paper contains results on the design of integral feedback controllers for discrete-time SISO plants, with maximal stability robustness against real time-invariant coefficient uncertainty in the plant numerator. The main result is the characterisation of nominal plants into two classes, according to whether or not maximal robustness is limited by the need to avoid a cancellation between the perturbed plant numerator and the integrator pole. Optimal stability margins and remarks about robust design are given for both kinds of plant.

The paper considers the output feedback robust stabilizability problem for hybrid dynamical systems. The hybrid system under consideration is a composite of a continuous plant and a discrete event controller. A necessary and sufficient condition of absolute stabilizability is given in terms of the existence of suitable solutions to a dynamic programming equation and a Riccati algebraic equation of the H/sup /spl infin// filtering type. A real time implementable method for absolute stabilization is also presented.

In this paper minimum norm duality results are applied to the problem of finding the largest l/sub p/ ball containing the coefficients of only stable polynomials and centred at a nominal stable point in coefficient space for both discrete-time and continuous-time systems. The results of this work are then applied to the computation of the largest stable hypercube (l/sub /spl infin// ball) in coefficient space. >

Considers the problem of estimating the state of a discrete-time, linear stochastic system given a number of distinct, noisy measurements at each time instant. The observation process consists of a finite set of known, linear measurement models with additive white noise. The number of measurements may vary with time and the correspondence of the measurements with the models is unknown. The authors derive a recursive, suboptimal filter that provides an effective solution to this multi-measurement association and filtering problem. >