As the buildings account for about 40% of global final energy use, the efficient building climate control can significantly contribute to the saving effort. emph{Predictive control} can be used to operate buildings in energy and cost effective manner instead of conventional room automation such as PID, weather-compensated controllers or Rule-Based Controllers (RBC). However, the predictive controller has (besides many advantages as the possibility to incorporate the restriction directly into the controller design or handling of MIMO systems in a simple natural way) a drawback - it is the necessity of a proper mathematical model of the controlled system. Therefore, adequate attention should be paid to the procedure leading to its acquirement. In this paper a multi-step ahead error minimization approach to a building modeling is presented and influence of the solar radiation on the quality of the constructed model is examined. Moreover, the results are demonstrated on a real control of six-floor building of the Czech Technical University in Prague.

 Open canals are large distributed systems characterized by non-linear, time-varying and dependent with the operating point behavior. Then, these systems can be suitably represented for control by linear parameter varying (LPV) models. In this paper, two ways of obtaining an LPV model for a single reach open canal are proposed and compared: a LPV IDZ model based on hydraulic laws and a LPV model based on identification techniques. The first approach is a white-box modeling approach while the last is a grey-box methodology using experimental data. Finally, they will be applied and compared in a test-bench canal.

 The Stewart Platform, one of the most successful and popular parallel robots, has attracted the attention of many researchers in recent decades. The solution of the forward kinematics problem in realtime is one of the key aspects that continues to garner interest. In this paper we propose a new approach for solving this particular case using Support Vector Machines, a popular Machine Learning method for classification and regression. The algorithm involves a data generation and preprocessing offline phase, and a fast online evaluation. The experiments show that this method is very accurate and suitable for use in realtime.

 Polymer electrolyte membrane (PEM) fuel cell systems are highly efficient energy converters. Besides electrical power, low oxygen concentration cathode exhaust gas, water and heat are the byproducts if fed with pure hydrogen gas. So, this technology has become very attractive for the use on aircraft where it is investigated as replacement for the auxiliary power unit that is currently used for electrical power generation. Hence, controlling the fuel cell system for electrical power is a central topic. The electrical power output, however, is nonlinearly dependent on stack temperature, gas pressure, membrane humidity and stack current that is being drawn. This study deals with the controls of electrical power of an air and hydrogen fed self-powered fuel cell system. For controller design the nonlinear polarization curve is approximated by a linear current-voltage-characteristic. Based on a feedback-linearization approach a nonlinear control law for fuel cell system power is presented and implemented in a nonlinear fuel cell system simulation model. Even though the polarization curve is a linear approximation, the nonlinear control law leads to a fast response and zero steady state error.

 A large, segmented space telescope requires high precision and accuracy in its mirror shape to obtain clear images. The Structures, Propulsion, and Control Engineering (SPACE) telescope testbed at the NASA sponsored University Research Center of excellence must maintain a pointing control accuracy of 2 arc seconds. A Peripheral Pointing Architecture (PPA) has been designed to demonstrate the Testbeds pointing accuracy. A Finite Element Analysis (FEA) model of the PPA is developed. Normal mode analysis is performed to establish the PPAs natural frequencies, mode shapes, and the mass and stiffness matrices. Utilizing the Hinfinity controllers developed to achieve figure maintenance, the pointing control of the Testbed structure is achieved.

 In this paper, a linear dynamic control is developed for a faulty flexible structure subject to external ground perturbation. This active controller is based on H infinity theory and is designed using linear matrix inequality (LMI) theory. Lyapunov theory is invoked to validate the control design. According to experiments, where a two levels flexible building with active mass damper and external perturbation is employed, show that this strategy improves controller performance when it is compared with a given controller. This design faces the nonlinear structural system too.