Abstract
In previous works, a learning law with a dead zone function was developed for multilayer differential neural networks. This scheme requires strictly a priori knowledge of an upper bound for the unmodeled dynamics. In this paper, the learning law is modified in such a way that this condition is relaxed. By this modification, the tuning process is simpler and the dead-zone function is not required anymore. On the basis of this modification and by using a Lyapunov-like analysis, a stronger result is here demonstrated: the exponential convergence of the identification error to a bounded zone. Besides, a value for upper bound of such zone is provided. The workability of this approach is tested by a simulation example.
1. Introduction
During the last four decades system identification has emerged as a powerful and effective alternative to the first principles modeling [1–4]. By using the first approach, a satisfactory mathematical model of a system can be obtained directly from an input and output experimental data set [5]. Ideally no a priori knowledge of the system is necessary since this is considered as a black box. Thus, the time employed to develop such model is reduced significantly with respect to a first principles approach. For the linear case, system identification is a problem well understood and enjoys well-established solutions [6]. However, the nonlinear case is much more challenging. Although some proposals have been presented [7], the class of considered nonlinear systems can result very limited. Due to their capability of handling a more general class of systems and due to advantages such as the fact of not requiring linear in parameters and persistence of excitation assumptions [8], artificial neural networks (ANNs) have been extensively used in identification of nonlinear systems [9–12]. Their success is based on their capability of providing arbitrarily good approximations to any smooth function [13–15] as well as their massive parallelism and very fast adaptability [16, 17].
An artificial neural network can be simply considered as a nonlinear generic mathematical formula whose parameters are adjusted in order to represent the behavior of a static or dynamic system [18]. These parameters are called weights. Generally speaking, ANN can be classified as feedforward (static) ones, based on the back propagation technique [19] or as recurrent (dynamic) ones [17]. In the first network type, system dynamics is approximated by a static mapping. These networks have two major disadvantages: a slow learning rate and a high sensitivity to training data. The second approach (recurrent ANN) incorporates feedback into its structure. Due to this feature, recurrent neural networks can overcome many problems associated with static ANN, such as global extrema search, and consequently have better approximation properties. Depending on their structure, recurrent neural networks can be classified as discrete-time ones or differential ones.
The first deep insight about the identification of dynamic systems based on neural networks was provided by Narendra and Parthasarathy [20]. However, none-stability analyses of their neuroidentifier were presented. Hunt et al. [21] called attention to determine the convergence, stability and robustness of the algorithms based on neural networks for identification and control. This issue was addressed by Polycarpou and Ioannou [16], Rovithakis and Christodoulou [17], Kosmatopoulos et al. [22], and Yu and Poznyak [23]. Given different structures of continuous-time neural networks, the stability of their algorithms could be proven by using Lyapunov-like analysis. All aforementioned works considered only the case of single-layer networks. However, as it is known, this kind of networks does not necessarily satisfy the property of universal function approximation [24]. And although the activation functions of single-layer neural networks are selected as a basis set in such a way that this property can be guaranteed, the approximation error can never be made smaller than a lower bound [24]. This drawback can be overcome by using multilayer neural networks. Due to this better capability of function approximation, the case multilayer was considered in [25] for feedforward networks and for continuous time recurrent neural networks for first time in [26] and subsequently in [27]. By using Lyapunov-like analysis and a dead-zone function, boundedness for the identification error could be guaranteed in [26]. The following upper bound for the “average” identification error was reported, where is the identification error, is a positive definite matrix, is a upper bound for the modeling error, is an upper bound for a deterministic disturbance, and is a dead-zone function defined as Although, in [28], open-loop analysis based on the passivity method for a multilayer neural network was carried out and certain simplifications were accomplished, the main result about the aforementioned identification error could not be modified. In [29], the application of the multilayer scheme for control was explored. Since previous works [26–29] are based on this “average” identification error, one could wonder about the real utility of this result. Certainly, boundedness for this kind of error does not guarantee that belongs to or . Besides, none value for upper bound of identification error norm is provided. Likewise, none information about the speed of the convergence process is presented. Another disadvantage of this approach is that the upper bound for the modeling error must be strictly known a priori in order to implement the learning laws for the weight matrices. In order to avoid these drawbacks, in this paper, we propose to modify the learning laws employed in [26] in such a way that their implementation does not require anymore the knowledge of an upper bound for the modeling error. Besides, on the basis of these new learning laws, a stronger result is here guaranteed: the exponential convergence of the identification error norm to a bounded zone. The workability of the scheme developed in this paper is tested by simulation.
2. Multilayer Neural Identifier
Consider that the nonlinear system to be identified can be represented by where is the measurable state vector for is the control input, is an unknown nonlinear vector function which represents the nominal dynamics of the system, and represents a deterministic disturbance. represents a very ample class of systems including affine and nonaffine-in-control nonlinear systems. However, when the control input appears in a nonlinear fashion in the system state equation (2.1), throughout this paper, such nonlinearity with respect to the input is assumed known and represented by .
Consider the following parallel structure of multilayer neural network where is the state of the neural network, is the control input, is a Hurwitz matrix which can be specified by the designer, the matrices and are the weights of output layers, the matrices and are the weights of hidden layers, is the activation vector-function with sigmoidal components, that is, , where , , and are positive constants which can be specified by the designer, is also a sigmoidal function, that is, where , , and are positive constants which can be specified by the designer, represents the nonlinearity with respect to the input—if it exists—which is assumed a priori known for the system (2.1). It is important to mention that and , that is, the number of neurons for and the number of rows for , respectively, can be selected by the designer.
The problem of identifying system (2.1) based on the multilayer differential neural network (2.2) consists of, given the measurable state and the input , adjusting on line the weights , and by proper learning laws such that the identification error can be reduced.
Hereafter, it is considered that the following assumptions are valid;(A1)System (2.1) satisfies the (uniform on ) Lipschitz condition, that is,
(A2)The differences of functions and fulfil the generalized Lipschitz conditions where , , , are known positive definite matrices, and are constant matrices which can be selected by the designer.As and fulfil the Lipschitz conditions and from Lemma A.1 proven in [26] the following is true: where and are unknown vectors but bounded by , , respectively; , , and are positive constants which can be defined as , , where and are global Lipschitz constants for and , respectively.(A3)The nonlinear function is such that where is a known positive constant.(A4)Unmodeled dynamics is bounded by where and are known positive constants and is a known positive definite matrix and can be defined as and are constant matrices which can be selected by the designer.(A5)The deterministic disturbance is bounded, that is, , is a known positive definite matrix.(A6)The following matrix Riccati equation has a unique, positive definite solution : where is a positive definite matrix which can be selected by the designer.Remark 2.1. Based on [30, 31], it can be established that the matrix Riccati equation (2.12) has a unique positive definite solution if the following conditions are satisfied;(a)The pair is controllable, and the pair is observable.(b)The following matrix inequality is fulfilled: Both conditions can relatively easily be fulfilled if is selected as a stable diagonal matrix.
(A7)It exists a bounded control , such that the closed-loop system is quadratic stable, that is, it exists a Lyapunov function and a positive constant such that Additionally, the inequality must be satisfied.Now, consider the learning law: where is the number of columns corresponding to , and are positive definite matrices which are selected by the designer. is a dead-zone function which is defined as Based on this learning law, the following result was demonstrated in [26].
Theorem 2.2. If the assumptions (A1)–(A7) are satisfied and the weight matrices , , and of the neural network (2.2) are adjusted by the learning law (2.16), then(a)the identification error and the weights are bounded: (b)the identification error satisfies the following tracking performance:
In order to prove this result, the following nonnegative function was utilized: where ; .
3. Exponential Convergence of the Identification Process
Consider that the assumptions (A1)–(A3) and (A5)-(A6) are still valid but the assumption (A4) is slightly modified as follows.(B4) In a compact set , unmodeled dynamics is bounded by where is a constant not necessarily a priori known.
Remark 3.1. (B4) is a common assumption in the neural network literature [17, 22]. As mentioned in Section 2, is given by . Note that and are bounded functions because and are sigmoidal functions. As belongs to , clearly is also bounded. Therefore, assumption B4 implies implicitly that is a bounded function in a compact set .
Although certainly assumption (B4) is more restrictive than assumption (A4), from now on, assumption (A7) is not needed anymore.
In this paper, the following modification to the learning law (2.16) is proposed: where , and are positive constants which are selected by the designer; is the solution of the Riccati equation given by (2.12); ; is the number of columns corresponding to . By using the constants , and in (3.1) instead of the matrices , and in (2.16), the tuning process of the neural network (2.2) is simplified. Besides, none dead-zone function is now required. Based on the learning law (3.1), the following result is here established.
Theorem 3.2. If the assumptions (A1)–(A3), (B4), (A5)-(A6) are satisfied and the weight matrices , , , and of the neural network (2.2) are adjusted by the learning law (3.1), then(a)the identification error and the weights are bounded: (b)the norm of identification error converges exponentially to a region bounded given by
Proof of Theorem 3.2. Before beginning analysis, the dynamics of the identification error must be determined. The first derivative of is Note that an alternative representation for (2.1) could be calculated as follows: Substituting (2.2) and (3.5) into (3.4) yields Subtracting and adding the terms , , , and and considering that , , , , , , (3.6) can be expressed as In order to begin analysis, the following nonnegative function is selected: where is a positive solution for the Riccati matrix equation given by (2.12). The first derivative of is Each term of (3.9) will be calculated separately. For , substituting (3.7) into (3.10) yields The terms , , , and in (3.11) can be bounded using the following matrix inequality proven in [26]: which is valid for any and for any positive definite matrix . Thus, for and considering assumption (A2), For and considering assumptions (A2) and (A3) By using (3.12) and given assumptions (B4) and (A5), and can be bounded, respectively, by Considering (2.8), can be developed as By simultaneously adding and subtracting the term into the right-hand side of (3.16), By using (3.12) and considering assumption (A2), the term can be bounded as And consequently, is bounded by For and considering (2.9), Adding and subtracting the term into the right-hand side of (3.20), By using (3.12), can be bounded by but considering that and from assumptions (A2) and (A3), the following can be concluded: Thus, is bounded by Consequently, given (3.13), (3.14), (3.15), (3.19), and (3.25), can be bounded as With respect to , using several properties of the trace of a matrix, As , the derivative of is clearly . However, is given by the learning law (3.1). Therefore, by substituting (3.1) into and the corresponding expression into the right-hand side of (3.27), it is possible to obtain Proceeding in a similar way for it is possible to obtain By substituting (3.26), (3.28), and (3.30) into (3.9), the following bound for can be determined: Simplifying like terms Adding and subtracting into the right-hand side of the last inequality yields the expression . That is However, the expression is, in accordance with the assumption (A6), equal to zero. Therefore, Now, considering that and using Rayleigh’s inequality, the following can be obtained: or alternatively In view of (3.36), it is possible to establish that As , finally the following bound for can be concluded: Equation (3.38) can be rewritten in the following form Multiplying both sides of the last inequality by , it is possible to obtain The left-hand side of (3.40) can be rewritten as or equivalently as Integrating both sides of the last inequality yields Adding to both sides of the inequality, Multiplying both sides of the inequality (3.44) by , the following can be obtained: and, consequently, As and are positive definite matrices, then is always a positive scalar and therefore is an upperly bounded function. However, in reference to (3.8), is also a nonnegative function. Consequently, , and, thus, the first part of the Theorem 3.2 has been proven. With respect to the final part of this theorem, from (3.8), it is evident that . Besides, from Rayleigh’s inequality, . Consequently, . Nonetheless, in accordance with (3.46), is bounded by . This means that Finally, taking the limit as of the last inequality, the last part of Theorem 3.2 has been proven.
Remark 3.3. Based on the results presented in [32, 33], and, from the inequality (3.38), uniform stability for the identification error can be guaranteed.
Remark 3.4. Although, in [34], the asymptotic convergence of the identification error to zero is proven for multilayer neural networks, the considered class of nonlinear systems is much more restrictive than in this work.
Remark 3.5. In [35], the main idea behind Theorem 3.2 was utilized but only for the single layer case. In this paper, the generalization for the multilayer case is presented for first time.
4. Tuning of the Multilayer Identifier
In this section, some details about the selection of the parameters for the neural identifier are presented. In first place, it is important to mention that the positive definite matrices , , , , , and presented throughout assumptions (A2)–(A5) are known a priori. In fact, their selection can be very free. Although, in many cases, identity matrices can be enough, the corresponding freedom of selection can be used to satisfy the conditions specified in Remark 2.1.
Other important design decision is related to the proper number of elements or neurons for . A good point of departure is to select where is the dimension of the state vector . Normally, this selection is enough in order to produce adequate results. In other case, should be selected such as . With respect to , for simplicity, a first attempt could be to set the elements of this matrix as zeroes except for the main diagonal.
Another very important question which must be taken into account is the following: how should the weights be selected? Ideally, these weights should be chosen in such a way that the modelling error or unmodeled dynamics can be minimized. Likewise, the design process must consider the solution of the Riccati equation (2.12). In order to guarantee the existence of a unique positive definite solution for (2.12), the conditions specified in Remark 2.1 must be satisfied. However, these conditions could not be fulfilled for the optimal weights. Consequently, different values for and could be tested until a solution for (2.12) can be found. At the same time, the designer should be aware of that as take values increasingly different from the optimal ones, the upper bound for unmodeled dynamics in assumption B4 becomes greater. With respect to the initial values for , and , some authors, for example [26], simply select , and .
Finally, a last recommendation, in order to achieve a proper performance of the neural network, the variables in the identification process should be normalized. In this context, normalization means to divide each variable by its corresponding maximum value.
5. Numerical Example
In this section, a very simple but illustrative example is presented in order to clarify the tuning process of the neural identifier and compare the advantages of the scheme developed in this paper with respect to the results of previous works [26–29].
Consider the following first-order nonlinear system: with the initial condition and the input given by .
For simplicity, is assumed equal to zero. It is very important to note that (5.1) is only used as a data generator since apart from the assumptions (A1)–(A3), (B4), (A5)-(A6), none previous knowledge about the unknown system (5.1) is required to satisfactorily carry out the identification process.
The parameters of the neural network (2.2) and the learning laws (3.1) are selected as Note that Riccati equation (2.12) becomes a simple second-order algebraic equation for this case: As , and given the previous values for these parameters, (5.4) has the following solution: . The rest of the parameters for the neural identifier are selected as . The initial condition for the neural identifier is selected as .
The results of the identification process are displayed in Figures 1 and 2. In Figure 1, the state of the nonlinear system (5.1) is represented by solid line whereas the state is represented by dashed line. Both states were obtained by using Simulink with the numerical method ode23s. In order to appreciate better the quality of the identification process, the absolute value of the identification error is showed in Figure 2. Clearly, the new learning laws proposed in this paper exhibit a satisfactory behavior.
Now, which is the practical advantage of this method with respect to previous works [26–29]? The determination of and more still (parameters associated with assumption A.4) can result difficult. Besides, assuming that is equal to zero, can result excessively large inclusive for simple systems. For example, for system (5.1) and the values selected for the parameters of the identifier, can approximately be estimated as 140. This implies that the learning laws (2.16) are activated only when due to the dead-zone function . Thus, although the results presented in works [26–29] are technically right, on these conditions, that is, , the performance of the identifier results completely unsatisfactory from a practical point of view since the corresponding identification error is very high. To avoid this situation, it is necessary to be very careful with the selection of weights , and in order to minimize the unmodeled dynamics . However, with these optimal weights, the matrix Riccati equation could have no solution. This dilemma is overcome by means of the learning laws (3.1) developed in this paper. In fact, as can be appreciated, a priory knowledge of is not required anymore for the proper implementation of (3.1).
6. Conclusions
In this paper, a modification of a learning law for multilayer differential neural networks is proposed. By this modification, the dead-zone function is not required anymore and a stronger result is here guaranteed: the exponential convergence of the identification error norm to a bounded zone. This result is thoroughly proven. First, the dynamics of the identification error is determined. Next, a proper nonnegative function is proposed. A bound for the first derivative of such function is established. This bound is formed by the negative of the original nonnegative function multiplied by a constant parameter plus a constant term. Thus, the convergence of the identification error to a bounded zone can be guaranteed. Apart from the theoretical importance of this result, from a practical point of view, the learning law here proposed is easier to implement and tune. A numerical example confirms the efficiency of this approach.
Acknowledgments
The authors would like to thank the anonymous reviewers for their helpful comments and advice which contributed to improve this paper. First author would like to thank the financial support through a postdoctoral fellowship from Mexican National Council for Science and Technology (CONACYT).