CDC2000 Nonlinear Filtering and Control

Nonlinear Filtering and Control
Home Full List of Titles 1: Proceedings of CDC2000 Discrete Event Systems Control in Communication Systems Optimal Control and Applications I Optimisation Approaches and Methods Model Predictive Control Advances in Linear Estimation Stochastic and Uncertain Systems Nonlinear Control and Applications Nonlinear Estimation and Filtering Formation Control and its Applications New Approaches to Fuzzy Control Manufacturing Systems Automotive Applications Stability Issues in Hybrid Control Recent Advances in Stochastic Networks Optimal Control and Applications II Robust Controller Design - mu, L1 and H2 Constrained and Receding Horizon Control Identification and Control around the World Markov Decision Processes Nonlinear Optimisation Observers for Nonlinear Systems Motion Planning Neural / Fuzzy Stability and Control Motor Control Control of Quantum Phenomena I Hybrid Systems Methods Control in Communication Networks Robustness and Optimisation Bumpless Transfer, Antiwindup and Saturation Adaptive Control: Linear Systems Estimation and Closed Loop Identification Control of Markov Processes Nonlinear Filtering and Control Modelling, Identification and Validation of Nonlinear Systems Differential Geometric Control Theory for Mechanical Systems Nonlinear Output Feedback Control Pneumatics and Compression Systems Control of Quantum Phenomena II Stability of Hybrid Systems Performance Analysis in Communication Networks Adaptive Control of Nonlinear Systems LMI Methods in Design Robust Control of Time Delay Systems Subspace Identification Methods Nonlinear Stochastic Filtering and Estimation Bifurcations, Chaos and Control I New Progress in Synthesis of Nonlinear Systems I Implementation Issues of Sliding Mode Control Theory Control of Mixing in Shear Flows Novel Neural Network Control Techniques for Industrial Motion Control Systems Physiological Control Systems Optimal Control of Hybrid Systems Stochastic Models for Communication Networks Control and Stabilisation of Nonlinear Systems New Directions in Robust Control Linear Systems Theory Advanced Topics in Systems Theory Estimation in Action Bifurcations, Chaos and Control II New Progress in Synthesis of Nonlinear Systems II Numerical Design and Analysis Techniques for Nonlinear Systems Analysis and Control of Underactuated Systems Sliding Mode Control I Challenges in the Application of Control to Computer Systems Estimation and Diagnosis of Discrete Event Systems Communications and Games Optimal Control Stochastic Systems Model Reduction Methodologies Identification and Subspace Methods Applications of Nonlinear Adaptive Control Advances in Nonlinear Output Feedback Design The Behavioural Approach to Systems and Control Vision Based Estimation and Control: Recent Advances and Open Problems Agile Control of Military Operations Sliding Mode Control II Model-based Fault Diagnosis of Industrial Processes Discrete Event Systems / Petri Nets System Identification and Confidence Estimation New Approaches to H-Infinity Control I Probabilistic Approaches to Robust Control Time Delay System Stabilisation Identification Methods Controlled Stochastic Processes Output Feedback of Nonlinear Systems Topics in Nonlinear Stabilisation Mobile Robots: Tracking Control Robust Control of Nonlinear Systems Power Systems Stabilisation and Control Disk Drive Control Hybrid Control Applications Discrete Time Systems New Approaches to H-Infinity Control II Linear Systems with Saturating Actuators New Theories in Distributed Parameter Systems Applications of Estimation and Identification Stochastic Control and Tuning Methodologies Control of Nonlinear Systems Iterative Learning and Control Coordinating Robot Systems Nonlinear Time Varying Systems Novel Applications of Neural Networks Aerospace Applications Switched Systems Implicit and Descriptor Systems LQG Periodic Systems and Disturbances New Horizons for Distributed Parameter Systems State Estimation Learning and Neuro-Control Nonlinear Control and Stabilisation I Tracking Vision Servoing Controllability of Nonlinear Systems Control of Flexible Systems Electro-Mechanical Systems Robust Control Methods and Applications Fault Detection and Diagnosis Optimisation and Applications Robust Stability Analysis Numerical Methods in Control Filtering in Continuous Time Stochastic Systems Interplay between Control and Signal Processing Fault Detection and Analysis Nonlinear Dynamical Systems Nonlinear Time Delay Systems Computational Issues in Nonlinear Control Disturbance Rejection Process Control Industry Applications Linear Parameter Varying Systems Linear Control Systems Dynamic and Nonlinear Programming Model Reduction Applications New Techniques for Control and Systems: Numerical Linear Algebra Estimation and Identification using Hidden Markov Models Applications of Stochastic Control Topics in Linear Design Nonlinear Control and Stabilisation II Ambulatory Robot Systems Chaotic and Oscillatory Systems Biomedical System Control Integrated Control and CPU Scheduling Linear Design Techniques Adaptive Disturbance / Noise Compensation Nonlinear Model Predictive Control Sensitivity Design, Analysis and Limitations Analysis of Linear Systems Linear Matrix Inequalities in Design Lyapunov's 2nd Method Robotics: Tracking Control Lagrangian and Hamiltonian Theory Variable Structure Control Machine Vision Signal Processing Methods in Control Applied Nonlinear Control Author Index A B C D E F G H I J K L M N O P Q R S T U V W X Y Z	Tracking Volatility Authors: Jaksa Cvitanić, Robert S. Liptser, Boris Rozovskii, Volume: 1, Page 1189 Paper number 3801 Abstract: This paper is concerned with nonlinear filtering of the volatility coefficient in a Black-Scholes type model that allows stochastic volatility. The stochastic volatility is modeled as a nonnegative function of a homogeneous Markov jump process. Following to Frey and Runggaldier, we assume that the asset price is measured only at random times. This assumption is designed to reflect the discrete nature of high frequency financial data (e.g. tick-by-tick stock prices). In the above setting the problem of volatility estimation is reduced naturally to a nonlinear filtering problem. We remark that while quite natural, the latter problem does not fit into the ''standard'' framework and requires new technical tools. In this paper, we derive a mean-square optimal recursive Bayesian filter for the volatility process. In particular, we derive Duncan-Mortensen-Zakai and Wonham-Kushner type equations for posterior distributions of the volatility process. CD003801.PDF (From Author) TOP Convergence and Error Analysis for a Max-Plus Algorithm Authors: William M. McEneaney, Volume: 1, Page 1194 Paper number 3803 Abstract: We consider max-plus based algorithms for the solution of nonlinear H-infinity problems. This class of algorithms has been described for several problem types such as nonlinear H-infinity filtering, nonlinear H-infinity control and nonlinear H-infinity control under partial information. Previous treatments have been oriented towards the general introduction of the algorithms. It has been noted that the corresponding convergence analysis was lacking in those papers. Here we demonstrate, in the case of nonlinear H-infinity control, that the errors introduced by the truncation to a finite number of max-plus basis functions go to zero as the number of basis functions increases. Some error bounds are also obtained. CD003803.PDF (From Author) TOP Efficient Particle Methods for Residual Generation in Partially Observed SDE's Authors: Frédéric Cérou, François LeGland, Volume: 1, Page 1200 Paper number 3804 Abstract: In this paper, the problem of detecting a change in the drift coefficient of a partially observed stochastic differential equation is addressed. The score function, evaluated at the nominal value, is used as the residual, and only the problem of residual generation is considered. In the special case where the drift coefficient depends on the parameter only in directions that are affected by nondegenerate noise, an efficient numerical approximation of the residual is proposed, using particle filters. The more complicated problem of residual evaluation will be considered elsewhere, under the small noise asymptotics. CD003804.PDF (From Author) TOP Risk-Sensitive Potfolio Optimization With Partial Information Authors: Hideo Nagai, Volume: 1, Page 1206 Paper number 3805 Abstract: We shall show some results applying risk-sensitive control with partial observation to mathematical finance, discussing portofolio optimization problems for factor models recently treated by Bielecki and Pliska(1999). They have formulated the factor models where securities prices are governed by the stochastic differential equations like geometric Brownian motion processes, whose drift coefficients (indicating expected growth rate in the case of geometric B.M.) are affine functions of the underlying economic factors such as price-earning ratios, short term interest rates, dividend yields, and macroeconomic measures and diffusion coefficients are constant indicating volatility of the securities. The set of such securities may include stocks, bonds, cash, and derivative securities. On the other hand the factors are assumed to be governed by the stochastic differential equations with linear drift coefficients. Then they consider the problem maximizing the risk-sensitized long run expected growth rate of capital value which the investor possesses by taking a portfolio strategy among the above defined securities. Representing the maximum as the function of the initial values of the factors and the value process, they introduce a Belleman equation of ergodic type. By using the solution of the Bellman equation they construct an optimal strategy. In their case the admissible investment strategies have been considered to be selected by using past informations of the securities to invest and also the factors. In the present paper we shall discuss similar problems to Bielecki and Pliska's, taking up such factor models and considering the optimization problems on finite time horizon. However, in our setting, our admissible strategies are considered to be chosen by using only past informations of securities to invest. We shall formulate our problems as risk-sensitive control problems of partially observable systems by regarding the factors as the system processes and the logarithm of securities prices as the observation processes and then obtain optimal strategies for the control problems having explicit representation by the solutions of ordinary differential equations and the ones of stochastic differential equations. Indeed we shall first express our criterion function by using the solution of a modified Zakai equation. Then, after giving the solution an explicit representation by means of the solution of a matrix Riccati differential equation and the one of a finite dimensional stochastic differential equation, we shall show the optimal strategy can be explicitly represented by the solutions and the ones of other three kinds of ordinary differential equations. We note that our results relate to Bensoussan and Van Schuppen's work (1985) having discussed the LEQG problem of a partially observable system. Difference from the work lies in that our noises are correlated and the performance index includes a stochastic integral. CD003805.PDF (From Author) TOP Discrete-Time Linear Filtering in Arbitrary Noise Authors: X. Rong Li, Chongzhao Han, Jie Wang, Volume: 1, Page 1212 Paper number 3806 Abstract: The Kalman filter is a recursive Best Linear Unbiased Estimator (BLUE) for a linear dynamic system with uncorrelated white process and measurement noises. It has been extended to the case where the noises are Markov and/or crosscorrelated for the same time instant. This paper presents optimal batch and semi-recursive filters and a suboptimal recursive filter for a linear discrete-time system with arbitrarily colored (not necessarily Markov) noises that are arbitrarily cross-correlated and correlated with the initial state of the system. They are generalizations of the Kalman filter for the case of arbitrary additive noise of known first two moments. Numerical examples are provided. They demonstrate the superiority in terms of performance and efficiency of the proposed recursive filter. CD003806.PDF (From Author) TOP Gaussian Filter for Nonlinear Filtering Problems Authors: Kazufumi Ito, Volume: 1, Page 1218 Paper number 3807 Abstract: In this paper we develop and analyze real-time and accurate filters for nonlinear filtering problems based on the Gaussian distributions. We present the systematic formulation of Gaussian filters and develop efficient and accurate numerical integration of the proposed filter. We also discuss the mixed Gaussian filters in which the conditional probability density is approximated by the sum of Gaussian distributions. Our numerical testings demonstrate that new filters significantly improve the extended Kalman filter with no additional cost and the new Gaussian sum filter has a nearly optimal performance. CD003807.PDF (From Author) TOP