On-line Estimation Algorithm Research Articles

Dynamic modeling of mean-reverting spreads for statistical arbitrage

Statistical arbitrage strategies, such as pairs trading and its generalizations rely on the construction of mean-reverting spreads enjoying a certain degree of predictability. Gaussian linear state-space processes have recently been proposed as a model for such spreads under the assumption that the observed process is a noisy realization of some hidden states. Real-time estimation of the unobserved spread process can reveal temporary market inefficiencies which can then be exploited to generate excess returns. We embrace the state-space framework for modeling spread processes and extend this methodology along three different directions. First, we introduce time-dependency in the model parameters, which allows for quick adaptation to changes in the data generating process. Second, we provide an on-line estimation algorithm that can be constantly run in real-time. Being computationally fast, the algorithm is particularly suitable for building aggressive trading strategies based on high-frequency data and may be used as a monitoring device for mean- reversion. Finally, our framework naturally provides informative uncertainty measures of all the estimated parameters. Experimental results based on Monte Carlo simulations and historical equity data are discussed, including a co-integration relationship involving two exchange-traded funds.

On-line topical importance estimation: an effective focused crawling algorithm combining link and content analysis

Focused crawling is an important technique for topical resource discovery on the Web. The key issue in focused crawling is to prioritize uncrawled uniform resource locators (URLs) in the frontier to focus the crawling on relevant pages. Traditional focused crawlers mainly rely on content analysis. Link-based techniques are not effectively exploited despite their usefulness. In this paper, we propose a new frontier prioritizing algorithm, namely the on-line topical importance estimation (OTIE) algorithm. OTIE combines link- and content-based analysis to evaluate the priority of an uncrawled URL in the frontier. We performed real crawling experiments over 30 topics selected from the Open Directory Project (ODP) and compared harvest rate and target recall of the four crawling algorithms: breadth-first, link-context-prediction, on-line page importance computation (OPIC) and our OTIE. Experimental results showed that OTIE significantly outperforms the other three algorithms on the average target recall while maintaining an acceptable harvest rate. Moreover, OTIE is much faster than the traditional focused crawling algorithm.

On-Line Policy Gradient Estimation with Multi-Step Sampling

In this note, we discuss the problem of the sample-path-based (on-line) performance gradient estimation for Markov systems. The existing on-line performance gradient estimation algorithms generally require a standard importance sampling assumption. When the assumption does not hold, these algorithms may lead to poor estimates for the gradients. We show that this assumption can be relaxed and propose algorithms with multi-step sampling for performance gradient estimates; these algorithms do not require the standard assumption. Simulation examples are given to illustrate the accuracy of the estimates.

An Extended Kalman Filtering Approach to Modeling Nonlinear Dynamic Gene Regulatory Networks via Short Gene Expression Time Series

In this paper, the extended Kalman filter (EKF) algorithm is applied to model the gene regulatory network from gene time series data. The gene regulatory network is considered as a nonlinear dynamic stochastic model that consists of the gene measurement equation and the gene regulation equation. After specifying the model structure, we apply the EKF algorithm for identifying both the model parameters and the actual value of gene expression levels. It is shown that the EKF algorithm is an online estimation algorithm that can identify a large number of parameters (including parameters of nonlinear functions) through iterative procedure by using a small number of observations. Four real-world gene expression data sets are employed to demonstrate the effectiveness of the EKF algorithm, and the obtained models are evaluated from the viewpoint of bioinformatics.

An estimation model of time-varying origin -destination flows in expressway corridors based on unscented Kalman filter

On the basis of measurable time series of mainline and ramp flows from traffic counts and the assumption of travel time distributions, this research presents a dynamic system model and its on-line estimation algorithm for recursive estimation of time-varying origin-destination (OD) matrices in expressway corridors. The proposed model employs a macro-traffic flow model to estimate travel times of OD flows and uses parameters of the traffic model as state variables, which are added to the constrained function of the system. To improve the model efficiency, we revise the travel time distribution based on the feature of normal distribution. The research employs a newly developed filtering technique, called unscented Kalman filter. The proposed model is evaluated with simulation experiments. Numerical analyses with respect to the sensitivity of the selection of initial parameters on the estimation results indicate that the proposed model is sufficiently reasonable and stable for real-world applications.

On-line Estimation of Departure Time-Based Link Travel Times from Spatial Detection System

The link travel time (LTT) estimation logic, using the spatial detection system, can be classified into two classes based on the aggregation point of view: departure time-based (DTB) and arrival time-based (ATB) aggregations. The former is conceptually advantageous over the latter because it provides the LTTs, which are experienced by vehicles departing at the same time period. However, the existing studies on link travel time (LTT) estimation have focused on ATB- LTT estimation, rather than on DTB-LTT estimation. Although some research in this area has developed DTB-LTT estimation algorithms, they are not applicable in a real-time mode. The objective of this study is to develop a DTB-LTT estimation algorithm, which is applicable in a real-time mode. This study compares the mean DTB- and ATB-LTT estimates from the two estimation perspectives: on-line and off-line. Then the tradeoff between the accuracy and timeliness of the on-line DTB-LTT estimates and their implications on the LTT estimation procedure are discussed. Lastly; this study develops an on-line DTB-LTT estimation algorithm which utilizes the Bayesian inference logic and demonstrates it using spatial travel time data from the Toll Collection System (TCS) of the Korea Highway Corporation. It was found that the proposed approach could estimate DTB-LTTs in a real-time context with an acceptable level of accuracy and timeliness.

Interaction Control of Robot Manipulators Using Force and Vision

An approach to force and visual control of robot manipulators in contact with a partially known environment is proposed in this article. The environment is modeled as a rigid object of known geometry but of unknown and time-varying position and orientation. An algorithm for online estimation of the object pose is adopted, based on visual data provided by a camera as well as on forces measured during the interaction. This information is used by a force/position control scheme, in charge of managing the interaction. Simulation and experimental results are presented for the case of an industrial robot manipulator.

Modelling and on-line estimation of zinc sulphide precipitation in a continuously stirred tank reactor

In this paper the ZnS precipitation in a continuously stirred tank reactor (CSTR) is modelled using mass balances. The dynamics analysis of the model reveals that the ZnS precipitation shows a two time-scales behaviour with inherent numerical stability problems, which therefore needs special attention during implementation. The mass balance model allows the design of an on-line estimator of the zinc concentration and precipitation rate, given only wet-chemical measurements of the sulphide concentration and of the pH in the reactor. The on-line estimation algorithm, also called observer or software sensor, is tested in a simulation experiment with a block disturbance showing numerically stable results. Furthermore, on the basis of real-world experimental data, the observer shows an appropriate performance. The estimation procedure offers good possibilities to be applied for other metal removal processes.

Online Parameter Estimation and Convergence Property of Dynamic Bayesian Networks

In this paper, we investigate a novel online estimation algorithm for dynamic Bayesian network(DBN) parameters, given as conditional probabilities. We sequentially update the parameter adjustment rule based on observation data. We apply our algorithm to two well known representations of DBNs: to a first-order Markov Chain(MC) model and to a Hidden Markov Model(HMM). A sliding window allows efficient adaptive computation in real time. We also examine the stochastic convergence and stability of the learning algorithm.

Autonomous Control Reconfiguration of Aerospace Vehicle Based on Control Effectiveness Estimation

Future aerospace vehicles (ASV) are designed to fly in both inner and extra atmospheric fields, which requires autonomous adaptability to the uncertainties emanated from abrupt faults and continuously time-varying environments. An autonomous control reconfiguration scheme is presented for ASV to deal with the uncertainties on the base of control effectiveness estimation. The on-line estimation methods for the time-varying control effectiveness of linear control system are investigated. Some sufficient conditions for the estimable system are given for different cases. There are proposed corresponding on-line estimation algorithms which are proved to be convergent and robust to noise using the least-square-based methods. On the ground of fuzzy logic and linear programming, the control allocation algorithms, which are able to implement the autonomous control reconfiguration through the redundant actuators, are put forward. Finally, an integrated system is developed to verify the scheme and algorithms by way of numerical simulation and analysis.

An On-Line Estimation Algorithm for Periodic Autoregressive Models

This paper develops an on-line estimation algorithm for periodic autoregressive models (PAR). Indeed, we provide an adaptation of the well known recursive least squares algorithm (RLS), which has been successfully applied to classical autoregressive models (AR), to deal with PAR models. The obtained estimators are shown to be asymptotically efficient under mild conditions. Moreover, the performance of the periodic least squares algorithm (PRLS) is assessed via an intensive simulation study.

Adaptive subsample delay estimation using windowed correlation

An on-line delay estimation algorithm based upon the windowed average magnitude difference function (AMDF) that is suitable for determining the subsample delay between two noisy signals is introduced. The new estimator uses the AMDF to directly update the delay estimate, which is in turn used to adapt the coefficients of a fractional delay filter (FDF). The new estimator consists predominately of addition and subtraction operations and outperforms conventional FDF based estimators in low signal-to-noise ratio situations

Design of neural networks using genetic algorithm for on-line property estimation of crude fractionator products

The products from the crude distillation unit (CDU) of a petroleum refinery need to conform to certain standards and hence it is important that the properties, which characterize these products, be measured on-line so that necessary control action can be taken through feedback mechanism. In view of nonavailability of on-line hardware sensors, software-based sensors have been developed for prediction of product properties on-line. Artificial neural network (ANN) based models have been used for this purpose. To overcome the problem of designing the networks by a cumbersome trial and error procedure, a methodology based on genetic algorithm (GA) has been developed. This involves optimally designing the network architecture including number of hidden layers, number of neurons in each layer, connectivity and activation functions. For on-line prediction of any property, the network design can be completed in a matter of a few hours. Additionally GA was also used for the selection of most relevant set of input features to the neural network. The above methodology was used for designing networks to predict properties of various side-draw products from a crude fractionator. The properties include ASTM temperatures, specific gravities and Flash Points of various products. All the predicted results were generally in good agreement with lab measurements with average deviation ranging from 0.5 to 3% for the properties investigated. The methodology is quite general and can be used to design similar other nets.

Parameter estimation for chaotic system based on particle swarm optimization

It's of vital importance to estimate the unknown parameters of chaos systems in chaos control and synchronization. We firstly improve the newly developed particle swarm optimization (PSO) in view of the population initialization and objective function treatment. Then we use the improved algorithms for parameter estimation and on-line estimation of chaotic system for its global searching ability. Experiments show that the improved method has better adaptability, reliability and high precision is robust to noise. It is proved to be a successful approach in parameter estimation for chaotic systems.

Contact impedance estimation for robotic systems

In this paper, the problem of online estimation of the mechanical impedance during the contact of a robotic system with an unknown environment is considered. This problem is of great interest when controlling a robot in an unstructured and unknown environment, such as in telemanipulation tasks, since it can be easily shown that the exploitation of the knowledge of the mechanical properties of the environment can greatly improve the performance of the robotic system. In particular, a single-point contact is considered, and the (nonlinear) Hunt-Crossley model is taken into account, instead of the classical (linear) Kelvin-Voigt model. Indeed, the former achieves a better physical consistency and also allows describing the behavior of soft materials. Finally, the online estimation algorithm is described and experimental results are presented and discussed.

Rapid On-Line Frequency Determination Using Non-Linear Signal Processing

A rapid on-line frequency estimation algorithm with the help of non-linearity is presented. A full-wave rectification of the raw signal is carried out such that the fundamental frequency is doubled. The rectified signal is filtered and the signal frequency is extracted based on the filtered fundamental component and second order harmonics. Simulation example and real measurement are included to demonstrate the effectiveness of the algorithm.

New sequential Monte Carlo methods for nonlinear dynamic systems

In this paper we present several new sequential Monte Carlo (SMC) algorithms for online estimation (filtering) of nonlinear dynamic systems. SMC has been shown to be a powerful tool for dealing with complex dynamic systems. It sequentially generates Monte Carlo samples from a proposal distribution, adjusted by a set of importance weight with respect to a target distribution, to facilitate statistical inferences on the characteristic (state) of the system. The key to a successful implementation of SMC in complex problems is the design of an efficient proposal distribution from which the Monte Carlo samples are generated. We propose several such proposal distributions that are efficient yet easy to generate samples from. They are efficient because they tend to utilize both the information in the state process and the observations. They are all Gaussian distributions hence are easy to sample from. The central ideas of the conventional nonlinear filters, such as extended Kalman filter, unscented Kalman filter and the Gaussian quadrature filter, are used to construct these proposal distributions. The effectiveness of the proposed algorithms are demonstrated through two applications--real time target tracking and the multiuser parameter tracking in CDMA communication systems.

ON-LINE IDENTIFICATION OF HYDRODYNAMICS IN UNDERWATER VEHICLES

In this work an algorithm for on-line estimation of the nonlinear hydrodynamics in underwater vehicles is presented. The algorithm is able to estimate physical parameters from natural operation signals. Identifiability and convergence of the parameter trajectories are analyzed. The application of the algorithm to a spherical vehicle is described with numerical simulations.

An equivalent soil mechanics formulation for rigid wheels in deformable terrain, with application to planetary exploration rovers

A simplified, closed-form version of the basic mechanics of a driven rigid wheel on low-cohesion deformable terrain is presented. This approach allows the formulation of an on-line terrain parameter estimation algorithm, which has important applications for planetary exploration rovers. Analytical comparisons of the original and simplified equations are presented, and are shown to closely agree. Experimental results from a single-wheel testbed operating in dry sand shows that the simplified equations can be used for mobility prediction with good accuracy. Methods for incorporating the simplified equations into an on-line terrain parameter algorithm are discussed.

A novel variable‐length sliding window blockwise least‐squares algorithm for on‐line estimation of time‐varying parameters

AbstractMotivated by the advances in computer technology and the fact that the batch/block least‐squares (LS) produces more accurate parameter estimates than its recursive counterparts, several important issues associated with the block LS have been re‐examined in the framework of on‐line identification of systems with abrupt/gradual change parameters in this paper. It is no surprise that the standard block LS performs unsatisfactorily in such a situation. To overcome this deficiency, a novel variable‐length sliding window‐based LS algorithm, known as variable‐length sliding window blockwise least squares, is developed. The algorithm consists of a change detection scheme and a data window with adjustable length. The window length adjustment is triggered by the change detection scheme. Whenever a change in system parameters is detected, the window is shortened to discount ‘old’ data and place more weight on the latest measurements. Several strategies for window length adjustment have been considered. The performance of the proposed algorithm has been evaluated through numerical studies. In comparison with the recursive least squares (RLS) with forgetting factors, superior results have been obtained consistently for the proposed algorithm. Robustness analysis of the algorithm to measurement noise have also been carried out. The significance of the work reported herein is that this algorithm offers a viable alternative to traditional RLS for on‐line parameter estimation by trading off the computational complexity of block LS for improved performance over RLS, because the computational complexity becomes less and less an issue with the rapid advance in computer technologies. Copyright © 2004 John Wiley & Sons, Ltd.