300

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 2, NO. 3, JULY 2011

Enhanced Identication of Battery Models for Real-Time Battery Management

Mark Sitterly, Student Member, IEEE, Le Yi Wang, Senior Member, IEEE, G. George Yin, Fellow, IEEE, and Caisheng Wang, Senior Member, IEEE

AbstractRenewable energy generation, vehicle electrication, and smart grids rely critically on energy storage devices for enhancement of operations, reliability, and efciency. Battery systems consist of many battery cells, which have different characteristics even when they are new, and change with time and operating conditions due to a variety of factors such as aging, operational conditions, and chemical property variations. Their effective management requires high delity models. This paper aims to develop identication algorithms that capture individualized characteristics of each battery cell and produce updated models in real time. It is shown that typical battery models may not be identiable, unique battery model features require modied input/output expressions, and standard least-squares methods will encounter identication bias. This paper devises modied model structures and identication algorithms to resolve these issues. System identiability, algorithm convergence, identication bias, and bias correction mechanisms are rigorously established. A typical battery model structure is used to illustrate utilities of the methods. Index TermsBattery management system, battery model, bias correction, convergence, identiability, parameter estimation, system identication.

I. INTRODUCTION

ENEWABLE energy generation, vehicle electrication, and smart grids rely critically on energy storage devices for enhancement of operations, reliability, and efciency. A large-scale battery system consists of hundreds, even thousands, of battery cells, which have different characteristics even when they are new, and change with time and operating conditions due to a variety of factors such as aging, operational conditions, and chemical property variations. State of charge, battery health, remaining life, charge and discharge resistance

Manuscript received September 10, 2010; accepted February 06, 2011. Date of publication February 17, 2011; date of current version June 22, 2011. The work of L. Y. Wang was supported in part by the National Science Foundation under DMS-0624849, and in part by the Air Force Ofce of Scientic Research under FA9550-10-1-0210. The work of G. G. Yin was supported in part by the National Science Foundation under DMS-0907753, and in part by the Air Force Ofce of Scientic Research under FA9550-10-1-0210. The work of C. Wang was supported by the National Science Foundation under ECS-0823865. M. Sitterly, L. Y. Wang are with the Department of Electrical and Computer Engineering, Wayne State University, Detroit, MI 48202 USA (e-mail: ac3981@wayne.edu; lywang@wayne.edu). G. G. Yin is with the Department of Mathematics, Wayne State University, Detroit, MI 48202 USA (e-mail: gyin@math.wayne.edu). C. Wang is with the Division of Engineering Technology and the Department of Electrical and Computer Engineering, Wayne State University, Detroit, MI 48202 USA (e-mail: cwang@wayne.edu). Color versions of one or more of the gures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identier 10.1109/TSTE.2011.2116813

and capacitance demonstrate nonlinear and time-varying dynamics [4], [5], [8], [12], [18], [24]. Consequently, for enhanced battery management, reliable system diagnosis, and improved power efciency, it is necessary to capture battery cell models in real time [18]. This paper aims to develop a real-time automated battery characterizer that will capture individualized characteristics of each battery cell and produce updated models in real time. The core of such a system is advanced system identication techniques [15], [27], [28], [29] that provide fast tracking capability to update each battery cells individual model when it is plugged into the battery system and to track battery aging and health conditions during operation. The main ideas of such identication methods include: 1) using load changes to stimulate or excite the internal behavior of the cell; 2) using measured voltage and current data to estimate model parameters such that the model-produced voltage and current trajectories match those from the actual battery cell; 3) employing recursive algorithms to accomplish automated model generation and real-time operation so that new measurement data are immediately used to update the model. To facilitate discussions on the key issues and challenges in real-time battery model identication, we employ, as a typical model structure, the resistancecapacitance battery model, which is part of ADVISOR, developed at National Renewable Energy Laboratory (NREL) [12]. The parameters of the components are functions of the state of charge (SOC) and cell temperature. In addition, the resistance values differ when the battery is in charge mode or discharge mode. Typically the parameters of the model and their dependence on the SOC and temperatures are experimentally established. In this work, we will use system identication methods to identify them in real time, without laboratory testing facilities. When identication can capture the circuit model parameters faster than a batterys charge/discharge operations, one may employ the frozen time concept and view the model parameters as their linearized values at a given operating condition. As such, identication of the system can concentrate on time-varying but linearized models. The goal of system identication is to identify system parameters on the basis of observations on the measured inputs and outputs. Continuing our recent work on battery models [22], [23], in this paper, we will use adaptive ltering algorithms to derive convergent algorithms. System identiability, algorithm convergence, identication bias, and bias correction mechanisms are rigorously established. The paper is organized as follows. Section II introduces battery model structures.

1949-3029/$26.00 2011 IEEE

SITTERLY et al.: ENHANCED IDENTIFICATION OF BATTERY MODELS FOR REAL-TIME BATTERY MANAGEMENT

301

Fig. 1. Battery model. Fig. 2. Battery thermal model.

Although our methodology is generic, it is helpful to use a typical model structure to reveal key issues and discuss their resolutions. System identiability is rst discussed in Section III. It is shown that physically meaningful models may not be real-time identiable. Modications of model structures to facilitate system identication are studied, leading to a modied regression expression for system identication. The section further explores the issue of identication bias correction. It is revealed that due to noise corruption in observation data and inherent correlation in the regressors, the standard least-squares (LS) algorithms will encounter identication bias, implying that parameter estimates will have an inherent error even when data sizes become very large. Bias correction measures are introduced to overcome this drawback. Strong convergence of the modied algorithms is established. For computational efciency, recursive algorithms are derived. The methodologies of this paper are then evaluated and illustrated by using simulation studies in Section IV. Both standard LS algorithms and modied algorithms are illustrated for their utilities. Finally, our ndings and potential applications are discussed in Section V. II. BATTERY MODEL STRUCTURES The methodologies developed in this paper can be applied to different battery models. We choose a well-developed and validated model structure to facilitate discussions of essential issues, including identication algorithms, identiability, bias correction, noise correlation, and convergence analysis. We employ the resistancecapacitance battery model which is part of ADVISOR, developed at NREL [3], [12] (see Fig. 1). The model contains two capacitors ( and ) and three resistors ( , , and ). The capacitor models the main storage capacity of the battery. The capacitor captures the fast chargedischarge aspect of the battery and is much smaller than . The parameters of the components are functions of the SOC and cell temperature ( ). In addition, the resistance depends on also if the battery is in charge or discharge mode. These will be expressed as when needed. To describe the model details, we dene the following symbols: terminal voltage (V); terminal current (A); voltage of the capacitor (V); (A); current through the capacitor

voltage of the capacitor

(V); (A);

current through the capacitor air temperature (degree Celsius);

cell temperature (degree Celsius); conducting heat transfer rate (W); heat transfer rate generated by the battery cell (W); air conditioning forced heat transfer rate (W); SOC; SOC SOC ; ;

It was given in [12] that the overall state of charge is a weighted combination of the states of charge on and (1) where . on . In the NREL/Saft model, and and will be related to the voltages

and . The thermal model is a lumped rst-order linear dynamics, shown in Fig. 2

where is the equivalent thermal resistance ( ) and is the equivalent heat capacitance ( ). In [3] and [12], the parameters of the model and their dependence on the SOC and temperatures are experimentally established. In this work, we will use system identication methods to identify them in real time, without laboratory testing facilities. The states of charge of and are closely related to their voltages and . The relationship and can be experimentally established. In the normal operating ranges of and , they can be well approximated by linear functions [12]. From , model parameters become functions of the state variables and the mode

(2)

302

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 2, NO. 3, JULY 2011

The battery model is derived from the following basic equations:

For simulation studies, we will use the NREL/Saft model [10][12]. The model structure is given in Fig. 1. Nominal values of the parameters at temperature 20 C are

(5) Although dependence of these parameters on and can be derived from experimental data, in this paper we employ system identication methods to capture real-time parameter values. III. IDENTIFICATION OF BATTERY MODELS and Although system parameters in the battery model (3) can be established by using experimental data from laboratory testing, for real-time operation under a battery management system (BMS), the parameters must be derived using real-time operational data. This is due to several factors: 1) New batteries have different characteristics even for the same model, due to manufacturing variations. 2) Battery features depend on many factors that cannot be totally captured in model details. 3) Batteries experience signicant aging effects. Consequently, it is not only desirable, but in fact imperative that model parameters be obtained individually in real time. A. Identiability of Battery Models We rst derive the transfer function for the battery model shown in Fig. 1. Using for the capacitors and deriving the transfer function in the -domain, we obtain (3) (6) where example, under m Remark 1: By (2), the model parameters depend on the state. Consequently, the system is highly nonlinear. Within the inputs, is measured and fed back as an input to the model. is a disturbance but is measured with some measurement noise. Since is not modeled in detail, it will be considered as an unmeasured disturbance to the battery system. is controlled (by the cooling system). The cooling system dynamics are not considered in this model. As a result, we will view as a control input. The outputs and are measured with measurement noise. is not measured. Denote the state vector by , input vector , and output vector . The state space model (3) may be written in an abstract form, for methodology and algorithm development, as (4) This is a nonlinear system in an afne form. Since is a control variable taking only two possible values, the system (4) is a hybrid system. kF, m , we have kF, . For m

These lead to

With state variables ; inputs (current load), (air temperature), (battery cell generated heat ow rate), and (convective heat ow rate due to cooling air); operating mode ; and outputs (terminal voltage), (state of charge), and ; the state space model is

(7) The inputoutput model (6) contains four coefcients. However, the internal circuit model in Fig. 1 contains ve parameters. In other words, there will be one redundant parameter in the model. As a result, the circuit model is not inputoutput identiable when only the terminal voltage and current are measured. Due to its physical meanings, the pair represents a much faster branch than the branch. Consequently, it is a sensible choice to let , reducing the battery system into a four-parameter circuit model. We show now that this simplication results in a direct inverse mapping from the transfer function parameters to the circuit parameters. When , (6) is reduced to

(8)

SITTERLY et al.: ENHANCED IDENTIFICATION OF BATTERY MODELS FOR REAL-TIME BATTERY MANAGEMENT

303

where , , can be derived as

, , . The inverse mapping of the parameters

where the regressor is . It should be pointed out that in practical applications, the measurements on the current and voltage are subject to noises. As a result, is measured as and as . Consequently,

(9) This denes a mapping from to . As a result, simplied circuit model (8) is always identiable. B. Regression Structure for System Identication Although the battery models physical input is the load (current ), since the model contains an integrator, for system identication the input/output setting needs to be modied. Suppose that the signal sampling interval in a BMS is . Let the sampled values be , , etc. To derive a sampled expression to identify parameters in (9), we note that where . In other words, noises enter the regressor also. This is a more complicated problem of errors-in-variable identication. C. Identication Algorithms For algorithm implementation, only From the observation equation and can be used.

(10) where have . Using the forward approximation, we . Equation (10) is mapped to

(11)

and the mapping , we would like to derive identication algorithms to track time-varying parameters. Assumption 1: Suppose that the measurement on is subject to measurement noise and the voltage measurement is subject to measurement noise . The joint vector sequence is stationary and ergodic [in the sense of convergence with probability one (w.p.1)] such that , , and that both and are ergodic. That is,

where , from (7) and

, second, we have

, . For the data

(12) Dene rived as a mapping . from can be deNote that we do not need the sequences and to be independent. A sufcient condition to ensure the ergodicity in the above assumption is that the underlying sequence is a stationary -mixing sequence, which is a sequence whose remote past and distant future are asymptotically independent. The well-known results [13, p. 488] then yield that and are strongly ergodic. It is noted that for time-varying systems, identication of system parameters can only be performed on the basis of most recent data, such as a nite-time moving window or exponentially decaying window. Hence, suppose the data on and are collected in a nite-time window . Let , the largest integer below . When , , and convergence analysis in this case should be interpreted as asymptotic accuracy of identication when sampling intervals become small.

Consequently, we have a mapping to the circuit parameters (13) The discrete-time system (11) leads to the regression expression of the new input/output relationship (14)

304

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 2, NO. 3, JULY 2011

Let

Identication bias can be corrected if The algorithm (15) is modied to . . . . . .

and

are known.

(16) It follows from Theorem 1 that this modied has a desired convergence property. If and are unknown, we can use statistical methods to estimate them. Then in the bias correction algorithm (16), in place of the true and , we may use their estimates. Theorem 2: Under the assumptions of Theorem 1, the estimates in (16) satisfy

. . . Then,

. . .

. . .

. From , the observation equation becomes . When is nonsingular, w.p.1, the standard LS estimate is

Proof: By the strong law of large numbers, as (15) We should demonstrate that due to signal correlation, this algorithm will introduce identication bias, namely, parameter estimates will converge to values away from the true value. D. Identication Bias and Correction Under Assumption 1, since , with probability one, and are deterministic, as

E. Recursive Algorithms for Bias Corrected LS Algorithms We now introduce a recursive algorithm for (16). Theorem 3: The estimates in (16) can be updated recursively as

which imply

Proof: From (16),

for some matrices . , , and . As a result, w.p.1. On the other hand, the true parameter satises . Consequently, we have the following theorem. Theorem 1: Assume that Assumption 1 holds and that exists. Then the least-squares estimate (15) is asymptotically biased in that

Let , we have

. Since

Proof: This follows from

By the matrix inversion lemma

This completes the proof.

(17)

SITTERLY et al.: ENHANCED IDENTIFICATION OF BATTERY MODELS FOR REAL-TIME BATTERY MANAGEMENT

305

Moreover,

Dene

By (17)

Fig. 3. Comparison of identication algorithms with and without bias correction for a second-order FIR system with a delay.

It follows that

Finally,

Fig. 4. Comparison of identication algorithms with and without bias correction for a third-order ARX system.

IV. SIMULATION STUDIES Although we use a specic battery model structure in this paper, the main methodologies introduced in this paper are generic and can be applied to different battery models. In this section, we rst illustrate the utility of bias correction algorithms by using some models of various structures and complexity. Then the specic circuit model is used for more detailed evaluations. A. Illustrative Examples on Identication Bias Correction We now illustrate effectiveness of the bias correction mechanisms described in Section III. Example 1: This example involves a nite impulse response (FIR) system of order 2 and one step delay: . The true parameter is . The input and the corresponding output are shown in the top plot of Fig. 3. Suppose that the input is corrupted by an independent and identically distributed (i.i.d.) Gaussian noise of mean zero

and variance 0.01, and the output is subject to an i.i.d. Gaussian measurement noise of mean zero and variance 0.16. Input and output noises are independent. In this example, and , where is the 2 2 identity matrix. Without direct noise correction, identication errors demonstrate a persistent bias of norm 0.6373 (averaged Euclidean norm), with the exit estimate . With direct bias correction, such a bias is substantially reduced to an error norm 0.0156 and the exit estimate . These are shown in the bottom plot of Fig. 3. Example 2: This example involves an auto-regression with external input (ARX) system

The true parameter is . The input and the corresponding output are shown in the top plot of Fig. 4. The input is corrupted by an i.i.d. Gaussian noise of mean zero and variance 0.04, and the output is subject to an

306

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 2, NO. 3, JULY 2011

Fig. 5. Battery model in Simulink.

i.i.d. Gaussian measurement noise of mean zero and variance 0.01. Input and output noises are independent. In this example, and

Under a data window of size , without direct noise correction, identication errors demonstrate a persistent bias of norm 0.1778, with the exit estimate

With direct bias correction, such a bias is reduced to an error norm 0.0061 and the exit estimate

These are shown in the bottom plot of Fig. 4. B. Identication of Circuit Models for Battery Systems The circuit model of the nominal parameter values (5) is simulated in a Simulink model shown in Fig. 5. The Simulink model generated the current and voltage proles shown in Fig. 6, which will be used for parameter estimation. It should be emphasized here that to enhance identication quality, a very small random dither has been added to the load (current). This is indicated in the Simulink model by a random signal generator, and reected in the signal proles with noise-like signals. No measurement noises are added in the proles at this point. The true system parameter in the modied regression structure (11) is

Fig. 6. Current and voltage proles of the battery.

Due to added dithers, the estimates are highly accurate, even with a very small data window of size 50. Although the estimation is of high delity in this case, estimation quality deteriorates substantially when measurement noises are introduced. When a very small measurement noise, a Gaussian noise of mean 0 and variance 0.00001, is added to the voltage measurements, parameter estimation loses its accuracy, shown in Fig. 7. The identication algorithms are then modied to include bias correction. Fig. 8 demonstrates substantially improved identication accuracy after bias correction. V. CONCLUDING REMARKS This paper has introduced identication methods for realtime battery model updates. The methods capture time-varying model parameters which can then be used for real-time control, optimization, diagnosis, and management. Due to some unique features in battery model structures, battery model identication encounters some challenging issues, such as identiability,

The estimated system by using the LS (or equivalently RLS) algorithms produced the parameter estimation for

SITTERLY et al.: ENHANCED IDENTIFICATION OF BATTERY MODELS FOR REAL-TIME BATTERY MANAGEMENT

307

Fig. 7. Parameter estimation errors increase when voltage measurements are slightly noise corrupted.

Fig. 8. Parameter estimation becomes more accurate when bias correction algorithms are implemented.

bias, and convergence. Modied identication algorithms are introduced to resolve these issues. Applications of the methods are demonstrated in some typical models in this paper. Much broader utilities are currently under investigation in different aspects of battery management systems. Also, discussions of this paper are limited to a single battery cell. Extensions to a battery pack consisting of a network of battery cells remain an open and critically important scenario for investigation. REFERENCES
[1] A. Benveniste, M. Metivier, and P. Priouret, Adaptive Algorithms and Stochastic Approximations. Berlin: Springer-Verlag, 1990. [2] K. Astrom and B. Wittenmark, Adaptive Controls, 2nd ed. Mineola, NY: Dover, 2008, pp. 4252. [3] Advanced Vehicle Simulator (ADVISOR) ver. 3.2 [Online]. Available: http://www.ctts.nrel.gov/analysis

[4] C. Barbier, H. Meyer, B. Nogarede, and S. Bensaoud, A battery state of charge indicator for electric vehicle, in Proc. Int. Conf. Institution of Mechanical Engineers, Automotive Electronics, London, U.K., May 1719, 1994, pp. 2934. [5] E. Barsoukov, J. Kim, C. Yoon, and H. Lee, Universal battery parameterization to yield a non-linear equivalent circuit valid for battery simulation at arbitrary load, J. Power Sour., vol. 83, no. 1/2, pp. 6170, 1999. [6] H. Chan and D. Sutanto, A new battery model for use with battery energy storage systems and electric vehicle power systems, in Proc. 2000 IEEE Power Engineering Society Winter Meeting, Singapore, Jan. 2327, 2000, pp. 470475. [7] H.-F. Chen and G. Yin, Asymptotic properties of sign algorithms for adaptive ltering, IEEE Trans. Automatic Control, vol. 48, no. 9, pp. 15451556, Sep. 2003. [8] R. Giglioli, P. Pelacchi, M. Raugi, and G. Zini, A state of charge observer for lead-acid batteries, Energia Elettrica, vol. 65, no. 1, pp. 2733, 1988. [9] W. B. Gu and C. Y. Wang, Thermal-electrochemical modeling of battery systems, J. Electrochem. Soc., vol. 147, no. 8, pp. 29102922, 2000. [10] V. Johnson, A. Pesaran, and T. Sack, Temperature-dependent battery models for high-power lithium-ion batteries, in Proc. 17th Electric Vehicle Symp., Montreal, Canada, Oct. 2000. [11] V. Johnson, M. Zolot, and A. Pesaran, Development and validation of a temperature-dependent resistance/capacitance battery model for ADVISOR, in Proc. 18th Electric Vehicle Symp., Berlin, Germany, Oct. 2001. [12] V. H. Johnson, Battery performance models In ADVISOR, J. Power Sources, vol. 110, pp. 321329, 2001. [13] S. Karlin and H. M. Taylor, A First Course in Stochastic Processes, 2nd ed. New York: Academic, 1975. [14] H. J. Kushner and G. Yin, Stochastic Approximation and Recursive Algorithms and Applications, 2nd ed. New York: Springer-Verlag, 2003. [15] L. Ljung, System Identication: Theory for the User. Englewood Cliffs, NJ: Prentice-Hall, 1987. [16] S. Piller, M. Perrin, and A. Jossen, Methods for state-of-charge determination and their applications, J. Power Sour., vol. 96, pp. 113120, 2001. [17] G. Plett, Extended Kalman ltering for battery management systems of LiPB-based HEV battery packs. Part 1. Background, J. Power Sour., vol. 134, no. 2, pp. 252261, 2004. [18] G. Plett, Extended Kalman ltering for battery management systems of LiPB-based HEV battery packs. Part 2, Modeling and Identication, J. Power Sour., vol. 134, no. 2, pp. 262276, 2004. [19] G. Plett, Extended Kalman ltering for battery management systems of LiPB-based HEV battery packs. Part 3. State and parameter estimation, J. Power Sour., vol. 134, no. 2, pp. 277292, 2004. [20] S. Rodrigues, N. Munichandraiah, and A. Shukla, A review of state-of-charge indication of batteries by means of ac impedance measurements, J. Power Sour., vol. 87, no. 1/2, pp. 1220, 2000. [21] A. Salkind, C. Fennie, P. Singh, T. Atwater, and D. Reisner, Determination of state-of-charge and state-of-health of batteries by fuzzy logic methodology, J. Power Sour., vol. 80, no. 1/2, pp. 293300, 1999. [22] M. Sitterly and L. Y. Wang, Identication of battery models for enhanced battery management, in Proc. Clean Technology Conf. Expo 2010, Anaheim, CA, Jun. 2124, 2010. [23] M. Sitterly, L. Y. Wang, and G. Yin, Enhanced identication algorithms for battery models under noisy measurements, in Proc. SAE Power Systems Conf., Ft. Worth, TX, Nov. 24, 2010. [24] K. Takano, K. Nozaki, Y. Saito, A. Negishi, K. Kato, and Y. Yamaguchi, Simulation study of electrical dynamic characteristics of lithium-ion battery, J. Power Sour., vol. 90, no. 2, pp. 214223, 2000. [25] Battery Modeling for HEV Simulation by ThermoAnalytics Inc. [Online]. Available: http://www.thermoanalytics.com/support/publications/batterymodelsdoc.html [26] E. Wan and A. Nelson, , S. Haykin, Ed., Dual extended Kalman lter methods, in Kalman Filtering and Neural Networks. New York: Wiley/Interscience, 2001, pp. 123174.

308

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY, VOL. 2, NO. 3, JULY 2011

[27] L. Y. Wang, Persistent identication of time varying systems, IEEE Trans. Automatic Control, vol. 42, no. 1, pp. 6682, Jan. 1997. [28] L. Y. Wang, J. F. Zhang, and G. Yin, System identication using binary sensors, IEEE Trans. Automat. Control, vol. 48, no. 11, pp. 18921907, Nov. 2003. [29] L. Y. Wang, G. Yin, J.-F. Zhang, and Y. Zhao, System Identication With Quantized Observations. Boston, MA: Birkhuser, 2010, ISBN: 978-0-8176-4955-5. [30] G. Yin, Stochastic approximation: Theory and applications, in Handbook of Stochastic Analysis and Applications, D. Kannan and V. Lakshmikantham, Eds. New York: Marcel Dekker, 2002, pp. 577624.

and currently is an Associate Editor of the Journal of System Sciences and Complexity and Journal of Control Theory and Applications.

Mark Sitterly (S02) received the B.S. and M.S. degrees in electrical engineering from Wayne State University, Detroit, MI, in 2000 and 2005, respectively. He is currently a Graduate Teaching Assistant and Ph.D. student at Wayne State University, Detroit, MI. He also serves in the U.S. Air Force with the Michigan Air National Guard as a Bioenvironmental Engineer. He previously worked at Eaton Corporation in Southeld, MI, as a Senior Electrical Engineer.

G. George Yin (S87M87SM96F02) received the B.S. degree in mathematics from the University of Delaware in 1983, the M.S. degree in electrical engineering, and the Ph.D. degree in applied mathematics from Brown University in 1987. He joined the Department of Mathematics, Wayne State University in 1987, and became a Professor in 1996. His research interests include stochastic systems, applied stochastic processes and applications, stochastic recursive algorithms, identication, signal processing, and control and optimization. He severed on the IFAC Technical Committee on Modeling, Identication and Signal Processing, and many conference program committees; he was the editor of SIAM Activity Group on Control and Systems Theory Newsletters, Cochair of the 1996 AMS-SIAM Summer Seminar in Applied Mathematics and the 2003 AMS-IMS-SIAM Summer Research Conference: Mathematics of Finance, Co-organizer of the 2005 IMA Workshop on Wireless Communications and the 2006 IMA PI conference. He is Cochair of 2011 SIAM Control Conference, Program Director of SIAM Activity Group on Control and Systems Theory, is an associate editor of Automatica and SIAM Journal on Control and Optimization, and is on the editorial board of a number of other journals. He was an Associate Editor of IEEE TRANSACTIONS ON AUTOMATIC CONTROL 19941998.

Le Yi Wang (S85M89SM01) received the Ph.D. degree in electrical engineering from McGill University, Montreal, Canada, in 1990. Since 1990, he has been with Wayne State University, Detroit, MI, where he is currently a Professor in the Department of Electrical and Computer Engineering. His research interests are in the areas of complexity and information, system identication, optimization, time-varying robust control, systems, adaptive systems, hybrid and nonlinear systems, information processing and learning, as well as medical, automotive, communications, power systems, and computer applications of control methodologies. He was a keynote speaker in several international conferences. He serves on the IFAC Technical Committee on Modeling, Identication and Signal Processing. He was an Associate Editor of the IEEE TRANSACTIONS ON AUTOMATIC CONTROL and several other journals,

Caisheng Wang (S02M06SM08) received the B.S. and M.S. degrees from Chongqing University, China in 1994 and 1997, respectively, and the Ph.D. degree from Montana State University, Bozeman, in 2006, all in electrical engineering. From August 1997 to May 2002, he worked as an electrical engineer in Zhejiang Electric Power Test and Research Institute, Hangzhou, China. Since August 2006, he has been a Faulty Member at the Division of Engineering Technology and the Department of Electrical and Computer Engineering at Wayne State University. His current research interests include modeling and control of power systems and electrical machines, energy storage devices, alternative/hybrid energy power generation systems, and fault diagnosis and on-line monitoring of electric apparatus.