Recursive parameter estimator

The architecture of the self-tuning regulator is shown in Fig. 7.99. It is similar to that of the Model Reference Adaptive Controller in that it also consists basically of two loops. The inner loop contains the process and a normal linear feedback controller. The outer loop is used to adjust the parameters of the feedback controller and comprises a recursive parameter estimator and an adjustment mechanism. 

Procedures on how to make inferences on the parameters and the response variables are introduced in Chapter 11. The design of experiments has a direct impact on the quality of the estimated parameters and is presented in Chapter 12. The emphasis is on sequential experimental design for parameter estimation and for model discrimination. Recursive least squares estimation, used for on-line data analysis, is briefly covered in Chapter 13. 

Given a set of experimental data, we look for the time profile of A (t) and b(t) parameters in (C.l). To perform this key operation in the procedure, it is necessary to estimate the model on-line at the same time as the input-output data are received [600]. Identification techniques that comply with this context are called recursive identification methods, since the measured input-output data are processed recursively (sequentially) as they become available. Other commonly used terms for such techniques are on-line or real-time identification, or sequential parameter estimation [352]. Using these techniques, it may be possible to investigate time variations in the process in a real-time context. However, tools for recursive estimation are available for discrete-time models. If the input r (t) is piecewise constant over time intervals (this condition is fulfilled in our context), then the conversion of (C.l) to a discrete-time model is possible without any approximation or additional hypothesis. Most common discrete-time models are difference equation descriptions, such as the Auto-.Regression with eXtra inputs (ARX) model. The basic relationship is the linear difference equation ... 

In this chapter we present very briefly the basic algorithm for recursive least squares estimation and some of its variations for single input - single output systems. These techniques are routinely used for on-line parameter estimation in data acquisition systems. They are presented in this chapter without any proof for the sake of completeness and with the aim to provide the reader with a quick overview. For a thorough presentation of the material the reader may look at any of the following references Soderstrom et al. (1978), Ljung and Soderstrom (1983) Shanmugan and Breipohl, (1988), Wellstead and Zarrop (1991). The notation that will be used in this chapter is different from the one we have used up to now. Instead we shall follow the notation typically encountered in the analysis and control of sampled data systems. 

Chapter 11 presents the use of sequential least squares techniques for the recursive estimation of uncertain model parameters. There is a statistical advantage in taking this approach to model parameter identification over that of incorporating model parameter estimation directly into Kalman filtering. 

In this chapter we discuss the principles of the Kalman filter with reference to a few examples from analytical chemistry. The discussion is divided into three parts. First, recursive regression is applied to estimate the parameters of a measurement equation without considering a systems equation. In the second part a systems equation is introduced making it necessary to extend the recursive regression to a Kalman filter, and finally the adaptive Kalman filter is discussed. In the concluding section, the features of the Kalman filter are demonstrated on a few applications. 

After each new observation, the estimates of the model parameters are updated (= new estimate of the parameters). In all equations below we treat the general case of a measurement model with p parameters. For the straight line model p = 2. An estimate of the parameters b based ony - 1 measurements is indicated by b(/ - 1). Let us assume that the parameters are recursively estimated and that an estimate h(j - 1) of the model parameters is available from y - 1 measurements. The next measurement y(j) is then performed at x(j), followed by the updating of the model parameters to b(/). 

By way of illustration, the regression parameters of a straight line with slope = 1 and intercept = 0 are recursively estimated. The results are presented in Table 41.1. For each step of the estimation cycle, we included the values of the innovation, variance-covariance matrix, gain vector and estimated parameters. The variance of the experimental error of all observations y is 25 10 absorbance units, which corresponds to r = 25 10 au for all j. The recursive estimation is started with a high value (10 ) on the diagonal elements of P and a low value (1) on its off-diagonal elements. 

In Sections 41.2 and 41.3 we applied a recursive procedure to estimate the model parameters of time-invariant systems. After each new measurement, the model parameters were updated. The updating procedure for time-variant systems consists of two steps. In the first step the system state j - 1) at time /), is extrapolated to the state x(y) at time by applying the system equation (eq. (41.15)) in Table 41.10). At time tj a new measurement is carried out and the result is used to... 

Recursive estimation methods are routinely used in many applications where process measurements become available continuously and we wish to re-estimate or better update on-line the various process or controller parameters as the data become available. Let us consider the linear discrete-time model having the general structure ... 

We shall present three recursive estimation methods for the estimation of the process parameters (ai,...,ap, b0, b,..., bq) that should be employed according to the statistical characteristics of the error term sequence e s (the stochastic disturbance). 

Historically, treatment of measurement noise has been addressed through two distinct avenues. For steady-state data and processes, Kuehn and Davidson (1961) presented the seminal paper describing the data reconciliation problem based on least squares optimization. For dynamic data and processes, Kalman filtering (Gelb, 1974) has been successfully used to recursively smooth measurement data and estimate parameters. Both techniques were developed for linear systems and weighted least squares objective functions. 

The individual estimates for the parameter can be obtained from the recursion... 

Based on current knowledge of the process and its disturbance characteristics, one may know or choose a reasonable difference equation structure for the controller algorithm. Starting with some assumed initial parameter values in the controller equation, the controller can be implemented on the process as shown. The control algorithm is coupled with an on-line recursive estimation algorithm which receives information on the inputs and outputs at each sampling interval and uses this to recursively estimate the optimal controller parameters on-line and to update the controller accordingly. The idea is to use the data collected from the on-line control manipulations to tune the controller directly. 

For processes which operate over a wide range of conditions or production rates it is to be expected that the process and disturbance model parameters will be changing with time. With a very minor change in the recursive estimation algorithm one can use this scheme to track these slowly moving processes and thereby keep the controller well tuned at all times. 

The linear model structures discussed in this section can handle mild nonlinearities. They can also result from linearization around an operating point. Simple alternatives can be considered for developing linear models with better predictive capabilities than a traditional ARMAX model for nonlinear processes. If the nature of nonlinearity is known, a transformation of the variable can be utilized to improve the linear model. A typical example is the knowledge of the exponential relationship of temperature in reaction rate expressions. Hence, the log of temperature with the rate constant can be utilized instead of the actual temperature as a regressor. The second method is to build a recursive linear model. By updating model parameters frequently, mild nonlinearities can be accounted for. The rate of change of the process and the severity of the nonlinearities are critical factors for the success of this approach. Another approach is based on the estimation of nonlinear systems by using multiple linear models [11, 82, 83]. 

T Sastri. A recursive estimation algorithm for adaptive estimation and parameter change detection of time series models. J. Op. Res. Soc., 37 987-999, 1986. 

---