- About this Journal ·
- Abstracting and Indexing ·
- Aims and Scope ·
- Annual Issues ·
- Article Processing Charges ·
- Author Guidelines ·
- Bibliographic Information ·
- Citations to this Journal ·
- Contact Information ·
- Editorial Board ·
- Editorial Workflow ·
- Free eTOC Alerts ·
- Publication Ethics ·
- Recently Accepted Articles ·
- Reviewers Acknowledgment ·
- Submit a Manuscript ·
- Subscription Information ·
- Table of Contents
Mathematical Problems in Engineering
Volume 2012 (2012), Article ID 631759, 19 pages
Least-Mean-Square Receding Horizon Estimation
1Department of Control and Instrumentation Engineering, Kangwon National University, Samcheock 245-711, Republic of Korea
2Department of Electrical Engineering, Konkuk University, Seoul 143-701, Republic of Korea
Received 31 October 2011; Accepted 20 December 2011
Academic Editor: M. D. S. Aliyu
Copyright © 2012 Bokyu Kwon and Soohee Han. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
We propose a least-mean-square (LMS) receding horizon (RH) estimator for state estimation. The proposed LMS RH estimator is obtained from the conditional expectation of the estimated state given a finite number of inputs and outputs over the recent finite horizon. Any a priori state information is not required, and existing artificial constraints for easy derivation are not imposed. For a general stochastic discrete-time state space model with both system and measurement noise, the LMS RH estimator is explicitly represented in a closed form. For numerical reliability, the iterative form is presented with forward and backward computations. It is shown through a numerical example that the proposed LMS RH estimator has better robust performance than conventional Kalman estimators when uncertainties exist.
Several criteria have been often employed for the design of optimal estimators. In particular, the mean-square-error criterion is the most popular and has many applications since it offers a simple closed-form solution as well as important geometric and physical interpretations. It is well known that the optimal estimators based on the mean-square-error criterion is obtained from the conditional expectation of the estimated variable given the known measurements.
For state estimation, many trials have been conducted to obtain a receding horizon (RH) estimator based on the mean-square-error criterion. At the current time , the least-mean-square (LMS) RH estimator is to estimate the state at the time from the inputs and the outputs over the recent finite horizon , which can be written as where and , that is, the lag size and the memory size are design parameters to be determined, respectively. Available inputs and outputs are regarded as given conditions and the corresponding conditional expectation of the state at time is obtained as its optimal estimate. Practically, inputs and outputs are known variables and hence can be considered as conditions. The structure of the LMS RH estimator (1.1) is depicted in Figure 1. As mentioned before, the LMS RH estimator (1.1) minimizes the mean-square-error criterion, . For , , and , the estimators (1.1) are often called the smoothers, the filters, and the predictors, respectively. “Receding horizon’’ is traced from the fact that the finite-time horizon, where inputs and outputs necessary for estimating unknown states are available, recedes with time. In designing the controls, receding horizon schemes have already been popular [1, 2]. If approaches , the estimator (1.1) reduces to the well-known stationary Kalman estimator with infinite-memory. So, the LMS RH estimator (1.1) can be also called a finite memory estimator for the finite and fixed memory size . It has been illustrated through numerical simulation and analysis that the LMS RH estimators with finite memory have better robust performance than conventional Kalman estimators with infinite memory [3, 4]. Also in input/output models arising in signal processing area, it is acknowledged that the finite-memory or finite-impulse-response (FIR) filters have been preferable for practical reasons .
In spite of the good performance and the usefulness of the LMS RH estimators, no one has proposed a general result on the conditional expectation (1.1). Since it was difficult to obtain the conditional expectation (1.1) for a general state space model, some assumptions have been made to simplify the problem and then obtain a solution easily. At first, system or measurement noise were set to zero, which offers a closed-form solution easily [6–10]. In [11, 12], a priori information on the initial state on the horizon was assumed to be known for obtaining a solution easily. Instead of directly obtaining a closed-form solution, the duality of a control and the complicated scattering theory from a physical phenomenon were used to show the feasibility of implementation for a general state space model [13, 14]. For easy derivation, the Kalman filter was also employed with infinite covariance . However, this approach is so heuristic that the optimality is not guaranteed in the sense of the mean-square-error criterion. Besides, the system matrix is required to be nonsingular. As in other conventional estimators, there were also some trials that unbiased and linear constraints are imposed to obtain the optimal estimator [3, 15–17]. However, external control inputs are not considered  and the system matrix is required to be nonsingular . Furthermore, it is not guaranteed that even though such constraints are removed, the optimality is still preserved. Though computing the conditional expectation (1.1) looks like a very simple problem, there is no result on a closed-form solution corresponding to (1.1) for a general state space model without any artificial assumptions and requirements.
In this paper, existing artificial assumptions for obtaining a solution easily are not made and any conditions are not required. Unlike the existing results, the system matrix is not required to be nonsingular. Both system and measurement noise are considered together with external control inputs. The LMS RH estimator will be directly obtained from the conditional expectation (1.1), which automatically guarantees its optimality. It turns out that the proposed LMS RH estimator has the deadbeat property and the linear structure with inputs and outputs over the recent finite horizon.
The rest of this paper is organized as follows. In Section 2, the LMS RH estimator is obtained from the conditional expectation and its iterative computation is introduced in Section 3. A numerical simulation is carried out in Section 4 to illustrate the performance of the proposed LMS RH estimator. Finally, conclusions are presented in Section 5.
2. LMS RH Estimator
Consider a linear discrete-time state space model with an external control input: where , , and are the state, the input, and the output, respectively, and the system noise and the measurement noise are assumed to be zero-mean white Gaussian with where nonzero often happens when I/O models are converted to state space models. It is also assumed that of the system is observable. Through this paper, the current time is denoted by . For mathematical tractability, the state space model (2.1) is equivalently changed to where and . It can be easily shown that and in (2.3) are not correlated. In other words, we have where . It is noted that off-diagonal blocks are filled with zeros.
For simple representations, several variables are defined as where denotes the direct sum of the matrices. Additionally, ,, and are defined by replacing in (2.5) with , , and , respectively. and are also defined by replacing in with and , respectively.
First, we consider the case of in (1.1). A prediction problem for will be discussed later on.
By using the defined variables, the state to be estimated is represented in terms of the initial state on the horizon, inputs, and system noise on the recent finite horizon , as where is given by and and are obtained by replacing in with and , respectively. From (2.9), the conditional expectation in (1.1) can be represented as Note that if and are known, can be obtained. In order to compute and , we first try to find the relation among , , , and . For this purpose, the system (2.3) is represented in a batch form on the recent finite horizon as follows: from which we obtain We can see from (2.13) that and are linearly, more correctly affinely, dependent on and . The joint probability density function of and , that is, , can be expressed as for an appropriate scaling factor . How to choose will be discussed later on. By using the probability density function of and given as can be computed from where a constant is chosen to satisfy the following normalization condition: In the same way, can be obtained. It is noted that is an exponential function with the following exponent: where denotes symmetric parts and , , and are given by By completing the square of the terms of the exponent and recalling the integration of Gaussian functions from to , we can compute and . To begin with, we introduce the following completion of squares: to obtain for some vectors and , and some matrices , , and of appropriate dimensions. The relation (2.20) can be easily obtained in a similar way to the mean of normal distribution. From the following correspondences: we have where , , and are given by (2.18). It is noted that if of the system (2.1) is observable, is positive definite and is also positive definite, which implies that the block matrix including , , and is guaranteed to be positive definite and hence nonsingular. Substituting (2.22) into (2.11), we can obtain . Through a long and tedious algebraic calculation, we have the covariance matrix of the mean-square-error as follows: where is defined by for .
What we have done until now can be summarized in the following theorem.
Theorem 2.1. Suppose that of the system (2.1) is observable. Then, the LMS RH estimator (1.1) is given by where denotes and , , and are defined in (2.18). The corresponding covariance matrix is given as (2.23).
It is noted that the proposed LMS RH estimator (2.25) is designed without requirements of the removal of some noise and the nonsingular system matrix . In previous work, those requirements are adopted to solve the problems more easily. If constraints or assumptions of existing results are applied to the proposed result, the latter reduces to the former. For example, and can be set to zero for comparison with the results on systems without system noise and inputs, respectively.
As seen in (2.25), the LMS RH estimator is linear with respect to inputs and outputs on the recent finite horizon . So, “finite memory’’ may be called “finite impulse response’’ that is often used in linear signal processing systems. In addition, the deadbeat property is guaranteed in the following theorem.
Theorem 2.2. If no noise is applied, the LMS RH estimator (2.25) is a deadbeat estimator, that is, at the fixed-lag () or the current () time.
Proof. The LMS RH estimator (2.25) can be rearranged in terms of and , that is, for some gain matrices and . Substituting in (2.12) into , using (2.9), and removing all noise, we have It can be easily seen that and terms associated with inputs and outputs become zero. It follows then that we have . This completes the proof.
According to the deadbeat property, the LMS RH estimator tracks down the real state exactly if no noise is applied.
The prediction problem for can be easily solved from the previous results for . Since is represented as we have which means that can be obtained from . Setting in (2.25) to zero, we can compute easily.
3. Iterative Computation
The LMS FM estimator (2.25) is of the compact and simple form. However, this form requires the inverse of big matrices, which may lead to long computation time and large numerical errors. To overcome these weak points, in this section, we provide an effective iterative form of (2.25).
First, we represent (2.25) in another form for getting recursive equations. Recalling the following fact: we obtain Note that nonsingularity of is guaranteed for the observability of . From (3.2), it can be easily seen that the LMS RH (2.25) can be represented as
Now, we consider two recursive equations for the batch form (3.3). One is for obtaining and . The other is for (3.3) given and . The first one is computed in a backward time and the second one in a forward time. Next subsections deal with each recursive equation.
3.1. Recursive Equation for Backward Computation
Here, how to obtain and in (3.3) will be discussed. These values will be computed in a backward time. and will be denoted by and , respectively, for consistency with the next section. The following theorem provides the main result.
Theorem 3.1. and can be computed recursively as follows: where and are obtained from for , , and and are zero matrices with appropriate dimensions.
Proof. Before going into a main proof, we introduce some variables , , , and as
for . In terms of in (3.10) and in (3.11), and can be represented as
where , , and are obtained by removing the first row blocks of , , and , respectively, and , , and are given by
for . In order to obtain and , we have only to know and . Now, in order to get and , we try to find recursive equations for and on . By using recursions in (3.8) and (3.9), we have the following equality:
where is given by
Pre- and postmultiplying (3.15) by
respectively, we have (3.6). Pre- and postmultiplying (3.15) by
respectively, we have (3.7). Note that (3.6) and (3.7) hold for . From (3.11) and (3.14), and can be written as
If and are set to zero matrices with appropriate dimensions, in (3.20) and in (3.21) can be calculated from (3.6) and (3.7). Thus, we can say that (3.6) and (3.7) hold for and are initiated with and . After obtaining from (3.6), we can calculate from (3.4). in (3.5) can be obtained from that comes from (3.7).
This completes the proof.
3.2. Recursive Equation for Forward Computation
Here, the recursive equation for (3.3) on the horizon is derived under the assumption that and are given. and in Section 3.1 are computed in a backward time while variables introduced in this section are computed in a forward time. Before proceeding to a main result, we introduce some variables and the necessary lemma.
, , and are defined as where and is the matrix obtained by removing the last zero column block from , that is, . In particular, , , and are defined as , and . The following lemma shows how variables , , and are related to one another.
Lemma 3.2 is useful for breaking up big matrices of (3.3) into small matrices. We now exploit the recursive equations for two following quantities: for for , where and are defined in a form (2.24), and , , , , and are given by Note that , , and . However, is recursively computed from to and from to with the initial value . Even though does not look useful on , it is still used to compute on that horizon. Now, we try to find out recursive equations for and in what follows.
3.2.1. Recursive Equation for on
Using Lemma 3.2, we will obtain a recursive equation for , which is introduced in the following theorem.
Proof. First, we will obtain a closed-form of in (3.34) and then, using the closed-form of , we show that in (3.30) can be computed recursively from (3.33).
Using the defined variables (3.22) and (3.24), we assume that in (3.34) is of the form:
By an induction method, we will prove (3.35). For the first step, we check (3.35) for . Given the initial value , we transform to the form (3.35) by (3.34) as follows: Equation (3.38) can be written in terms of and as . Now, we check under the assumption that . From (3.34), we have Thus, we can see that in (3.34) can be represented as the form (3.35) in terms of and . Using this result, we are in a position to show that in (3.30) can be computed recursively from (3.33). We can rewrite in (3.30) as where is given by As in a derivation of in a batch form, we show by an induction method that in (3.38) is equivalent to the one in (3.33). First, we show that can be obtained from and : where, in terms of , , and , can be written as . Next, we show that in the form (3.38) can be obtained from of the batch form, that is, . First, note that we have the following relations: Substituting (3.41) into (3.33) yields where and in the first and second matrix blocks on the right-hand side of the first equality have no effect on the equation. This completes the proof.
It is observed that the recursive equations with (3.33) and (3.34) are the same as the Kalman filter with initial conditions and . on provides initial values and inputs for a recursive equation of (3.31), which will be investigated in what follows.
3.2.2. Recursive Equation for on
Here, we discuss the recursive equation for on . As mentioned before, the recursive equation for starts from with the initial value . On , can be recursively computed with the help of . However, is what we want to find out finally. The recursive equation for is given in the following theorem.
Proof. Using Lemma 3.2, in (3.31) for can be represented as If we denote by , we have only to prove (3.44). Since , is equal to . Using Lemma 3.2, we can represent in a recursive form as follows: This completes the proof.
It is noted that the recursive equation (3.43) in Theorem 3.4 is a fixed-point smoother of the state , which runs with a recursive equation (3.33) in Theorem 3.3. Variables for recursive equations in Sections 3.1 and 3.2 are visualized in Figure 2. Starting from and , we compute and recursively in a backward time. From and , we compute and , from which we drive the forward recursive equation for to get .
In this section, a numerical example is given to demonstrate the performance of the proposed LMS RH estimator. Suppose that we have a state space model represented as where is an uncertain model parameter given as The system noise covariance and the measurement noise covariance are set to and , respectively. The memory size and the fixed-lag size are taken as and , respectively. A sinusoidal input is applied as an input.
We carry out a simulation for the system (4.1) with temporary modeling uncertainties (4.2). In Figure 3, we compare the estimation errors of the LMS RH estimator with the fixed-lag Kalman estimator . When uncertainties do not exist, the fixed-lag Kalman estimator has the smaller estimation error than the proposed LMS RH estimator. It can, however, be seen that the estimation error of the LMS RH estimator is considerably smaller than that of the fixed-lag Kalman estimator when modeling uncertainties exist. Actually, one of poles of the fixed-lag Kalman estimator is so close to a unit that even small uncertainties have a good chance of divergence. Additionally, we can see that the estimation error of the LMS RH estimator converges much more rapidly than that of the fixed-lag Kalman estimator after temporary modeling uncertainty disappears. The slow response is also related to the pole near a unit. To be summarized, we can say that the proposed LMS RH estimator is more robust than other estimators with infinite memory when applied to systems with modeling uncertainties.
In this paper, we proposed a receding horizon (RH) estimator based on the mean-square-error criterion for a discrete-time state space model, called a least-mean-square (LMS) RH estimator. An unknown state was estimated by making use of the finite number of inputs and outputs over the recent finite horizon without any arbitrary assumptions and any a priori state information. The proposed LMS RH estimator was obtained from the conditional expectations of the initial state and the system noise on the corresponding horizon. It was shown that the LMS RH estimator has a deadbeat property and has good robust performance through a numerical example.
To the best of authors’ knowledge, the proposed LMS RH estimator would be the most general version among existing RH or finite-memory estimators in the mean-square-error sense. Furthermore, the LMS RH estimator could be extended to other stochastic systems with imperfect communications, uncertainties, and so on [19, 20].
This work was supported by the Ministry of Knowledge Economy, Rebublic of Korea (ID: 10040132). The material in thid paper was presented in part at the International Symposium on Intelligent Informatics, Qingdao, China, May 2011.
- J. B. Rawlings and K. R. Muske, “The stability of constrained receding horizon control,” IEEE Transactions on Automatic Control, vol. 38, no. 10, pp. 1512–1516, 1993.
- G. De Nicolao, L. Magni, and R. Scattolini, “Stabilizing receding-horizon control of nonlinear time-varying systems,” IEEE Transactions on Automatic Control, vol. 37, no. 9, pp. 1405–1410, 1998.
- B. K. Kwon, S. Han, O. K. Kwon, and W. H. Kwon, “Minimum variance FIR smoothers for discrete-time state space models,” IEEE Signal Processing Letters, vol. 14, no. 8, pp. 557–560, 2007.
- W. H. Kwon, P. S. Kim, and P. Park, “A receding horizon Kalman FIR filter for discrete time-invariant systems,” IEEE Transactions on Automatic Control, vol. 44, no. 9, pp. 1787–1791, 1999.
- A. Oppenheim and R. Schafer, Discrete-Time Signal Processing, Prentice-Hall, Englewood Cliffs, NJ, USA, 1989.
- G. J. Bierman, “Fixed memory least squares filtering,” IEEE Transactions on Information Theory, vol. 21, no. 6, pp. 690–692, 1975.
- P. J. Buxbaum, “Fixed-memory recursive filters,” IEEE Transactions on Information Theory, vol. 20, no. 1, pp. 113–115, 1974.
- A. H. Jazwinski, “Limited memory optimal filtering,” IEEE Transactions on Automatic Control, vol. 13, pp. 558–563, 1968.
- F. Schweppe, Uncertain Dynamic Systems, Prentice-Hall, Englewood Cliffs, NJ, USA, 1973.
- K. V. Ling and K. W. Lim, “Receding horizon recursive state estimation,” IEEE Transactions on Automatic Control, vol. 44, no. 9, pp. 1750–1753, 1999.
- A. Alessandri, M. Baglietto, and G. Battistelli, “Receding-horizon estimation for discrete-time linear systems,” IEEE Transactions on Automatic Control, vol. 48, no. 3, pp. 473–478, 2003.
- A. Alessandri, M. Baglietto, and G. Battistelli, “Receding-horizon estimation for switching discrete-time linear systems,” IEEE Transactions on Automatic Control, vol. 50, no. 11, pp. 1736–1748, 2005.
- A. M. Bruckstein and T. Kailath, “Recursive limited memory filtering and scattering theory,” IEEE Transactions on Information Theory, vol. 31, no. 3, pp. 440–443, 1985.
- W. H. Kwon and A. E. Pearson, “A modified quadratic cost problem and feedback stabilization of a linear system,” IEEE Transactions on Automatic Control, vol. 22, no. 5, pp. 838–842, 1977.
- W. H. Kwon, P. S. Kim, and S. H. Han, “A receding horizon unbiased FIR filter for discrete-time state space models,” Automatica, vol. 38, no. 3, pp. 545–551, 2002.
- S. Han and W. H. Kwon, “- FIR smoothers for deterministic discrete-time state-space signal models,” IEEE Transactions on Automatic Control, vol. 52, no. 5, pp. 927–932, 2007.
- S. Han, B. K. Kwon, and W. H. Kwon, “- FIR smoothers for deterministic continuous-time state space signal models,” Automatica, vol. 45, no. 6, pp. 1561–1566, 2009.
- T. Kailath, A. H. Sayed, and B. Hassibi, Linear Estimation, Prentice Hall, 2000.
- B. Shen, Z. Wang, H. Shu, and G. Wei, “Robust finite-horizon filtering with randomly occurred nonlinearities and quantization effects,” Automatica, vol. 46, no. 11, pp. 1743–1751, 2010.
- B. Shen, Z. Wang, H. Shu, and G. Wei, “On nonlinear filtering for discrete-time stochastic systems with missing measurements,” IEEE Transactions on Automatic Control, vol. 53, no. 9, pp. 2170–2180, 2008.