Luís Brito Palma

SMCRVI - a Labview/Matlab based tool for remote monitoring and control

Modeling and dynamic simulation are now basic tools for understanding and verifying theoretical subjects. However, the experimentation with a plant cannot be replaced by simulation. Laboratory experiments play and will certainly play an important role in control-engineering education. The high number of students and the limitations of economic resources require an efficient use of laboratory experiments. This paper presents SMCRVI - a tool that combines the use of Matlab and Labview to allow students to gain remote access to laboratory experiments. By means of a client/server architecture students are able to share online experiment data and to perform practical tests, without leaving their Matlab environment

Gain Tuning of Fuzzy PID Controllers for MIMO Systems: A Performance-Driven Approach

A new methodology for tuning the scaling factors, or gains, of fuzzy proportional-integral-derivative controllers, by taking explicitly into account the closed-loop system performance is proposed in this study. The solution is obtained by solving a nonlinear constrained optimization problem, considering a set of constraints on the scaling factors of the Mamdani-type fuzzy system, and on the plants inputs and outputs. Two distinct approaches are presented, which are associated with the optimization being carried out offline or in real time. The offline tuning scheme assumes the system dynamics described by a nonlinear model, while for the real-time implementation, the plants dynamics is locally approximated by a linear model, with the underlying parameters recursively updated. In order to cope with rather stringent sampling time requirements, the constrained online optimization problem is implemented based on the grid computing paradigm. Given the adaptive nature of the real-time scheme, time-varying dynamics and unknown disturbances can be accommodated in such a way that the closed-loop performance is effectively maximized, while avoiding wind-up phenomena induced by the integrator term. The proposed tuning methodologies are assessed on a benchmark three-tank system and compared against a conventional-based tuning approach. Results from experiments illustrate the feasibility of the proposed approaches and also all the relevance in optimal control systems based on Mamdani-type fuzzy controllers.

Process control based on PCA models

In this paper an approach to design controllers based on principal components analysis (PCA) models is presented. Closed-loop control can be formulated and implemented within the reduced space defined by a PCA model. This PCA controller, results in an integral controller, which can be used as an inferential controller when a measurement of a primary variable is not available. The main contributions of the paper are: a) the incorporation of an adjustable gain on the classical PCA controller; b) the proposal of a set of tuning rules; c) the performance evaluation of this new controller, in nominal operation conditions and in faulty situations. Some experimental results, obtained with the three tank benchmark (European COSY project), are presented.

A remote laboratory environment for blended learning

In this paper, a remote laboratory environment for blended learning is described. The main contributions are: a) the architecture proposed that facilitates the student-centered learning; b) an environment that improves the students learning process; c) the set of experiments available that can be easily incorporated in an e-learning platform. The Labview® platform was used on the server, a data acquisition board (NI-USB-6009) interfaces with the real setups, and the clients access remotely using a web-browser. The practice showed that the proposed remote laboratory can be used, not only for local experiments in a laboratory, but also for e-teaching and e-learning activities using a remote laboratory.

A support vector machine based technique for online detection of outliers in transient time series

This paper deals with online detection and accommodation of outliers in transient time series by appealing to a machine learning technique. The methodology is based on a Least Squares Support Vector Machine technique together with a sliding window-based learning algorithm. A modification to this method is proposed so as to extend its application to transient raw data collected from transmitters attached to a Wireless Sensor Network. The performance of two approaches are compared on a particular controlled data set.

A distributed multi-agent approach for resilient supervision over a IPv6 WSAN infrastructure

Wireless Sensor and Actuator Networks has become an important area of research. They can provide flexibility, low operational and maintenance costs and they are inherently scalable. In the realm of Internet of Things the majority of devices is able to communicate with one another, and in some cases they can be deployed with an IP address. This feature is undoubtedly very beneficial in wireless sensor and actuator networks applications, such as monitoring and control systems. However, this kind of communication infrastructure is rather challenging as it can compromise the overall system performance due to several factors, namely outliers, intermittent communication breakdown or security issues. In order to improve the overall resilience of the system, this work proposes a distributed hierarchical multi-agent architecture implemented over a IPv6 communication infrastructure. The Contiki Operating System and RPL routing protocol were used together to provide a IPv6 based communication between nodes and an external network. Experimental results collected from a laboratory IPv6 based WSAN test-bed, show the relevance and benefits of the proposed methodology to cope with communication loss between nodes and the server.

Estimating the number of hidden neurons in recurrent neural networks for nonlinear system identification

The problem of complexity is here addressed by defining an upper bound for the number of the hidden layers neurons. This majorant is evaluated by applying a singular value decomposition to the contaminated oblique subspace projection of the row space of future outputs into the past inputs-outputs row space, along the future inputs row space. Full rank projections are dealt with by i) computing the number of dominant singular values, on the basis of a threshold related to the Euclidean norm of an artificial error matrix and ii) finding the argument of minimizing the singular value criterion. Results on a benchmark three-tank system demonstrate the effectiveness of the proposed methodology.

Fault diagnosis based on black-box models with application to a liquid-level system

This paper proposes an on-line robust approach to fault detection and isolation (FDI) of dynamic systems. This FDI approach is based on black-box models: artificial neural networks (ANNs) and the autoregressive with exogenous input (ARX) models. ANNs are used as observers and pattern classifiers, and adaptive ARX models are used as observers. The generalized likelihood ratio (GLR) algorithm is used for change detection. Process faults are considered, and the robust FDI problem is also addressed. The approach is applied to a laboratory set-up tank system under closed-loop control.

A machine learning technique in a multi-agent framework for online outliers detection in Wireless Sensor Networks

Wireless Sensor Networks enable flexibility, low operational and maintenance costs, as well as scalability in a variety of scenarios. However, in the context of industrial monitoring scenarios the use of Wireless Sensor Networks can compromise the systems performance due to several factors, being one of them the presence of outliers in raw data. In order to improve the overall systems resilience, this paper proposes a distributed hierarchical multi-agent architecture where each agent is responsible for a specific task. This paper deals with online detection and accommodation of outliers in non-stationary time-series by appealing to a machine learning technique. The methodology is based on a Least Squares Support Vector Machine along with a sliding window-based learning algorithm. A modification to this method is considered to improve its performance in transient raw data collected from transmitters over a Wireless Sensor Networks (WSNs). An empirical study based on laboratory test-bed show the feasibility and relevance of incorporating the proposed methodology in the context of monitoring systems over Wireless Sensor Networks.

Recursive subspace system identification for parametric fault detection in nonlinear systems

Graphical abstractDisplay Omitted HighlightsEigenstructure based fault detection for nonlinear systems.Recursive subspace based system identification techniques.Two linear models updated in parallel.Local eigenvalues residuals as symptoms. This work addresses the problem of detecting parametric faults in nonlinear dynamic systems by extending an eigenstructure based technique to a nonlinear context. Two local state-space models are updated online based on a recursive subspace system identification technique. One of the models relies on input-output real-time data collected from the plant, while the other is updated using data generated by a neural network predictor, describing the nonlinear plant behaviour in fault-free conditions. Parametric faults symptoms are generated based on eigenvalues residuals associated with two linear state-space model approximators. The feasibility and effectiveness of the proposed framework are demonstrated through two case studies.