George Hassapis

Nonlinear model identification and adaptive model predictive control using neural networks

This paper presents two new adaptive model predictive control algorithms, both consisting of an on-line process identification part and a predictive control part. Both parts are executed at each sampling instant. The predictive control part of the first algorithm is the Nonlinear Model Predictive Control strategy and the control part of the second algorithm is the Generalized Predictive Control strategy. In the identification parts of both algorithms the process model is approximated by a series-parallel neural network structure which is trained by a recursive least squares (ARLS) method. The two control algorithms have been applied to: 1) the temperature control of a fluidized bed furnace reactor (FBFR) of a pilot plant and 2) the auto-pilot control of an F-16 aircraft. The training and validation data of the neural network are obtained from the open-loop simulation of the FBFR and the nonlinear F-16 aircraft models. The identification and control simulation results show that the first algorithm outperforms the second one at the expense of extra computation time.

A Modified DPWS Protocol Stack for 6LoWPAN-Based Wireless Sensor Networks

In this paper, a modification of the protocol stack of the device profile for web services (DPWS) is proposed which can be applied in wireless sensor networks (WSNs) that comply with the IPv6 over low-power wireless personal area networks (6LoWPAN) architecture. The modification is based on a new format for the DPWS message exchanges without prohibiting the usage of the web services (WS) and the extensible markup language (XML) set of rules. The modified DPWS was implemented on the SunSPOT wireless sensor mote (WSM) and it was observed that it processes XML documents with a mean computation time less by 53% than the respective computation time of the DPWS while it consumes less EEPROM and RAM by 84% and 85%, respectively. Furthermore, its network performance was assessed by testing it over a real 6LoWPAN-based WSN with its maximum number of WSMs being 12. In order to validate these results and extend them to larger-scale 6LoWPAN-based WSNs, the network simulator 2 (NS-2) was used by enhancing it with a developed 6LoWPAN object. The NS-2 was also utilized for comparing the modified DPWS, the DPWS and a binary-based DPWS in terms of packet delivery ratio and maximum transmission delay. Simulation results have shown that the modified DPWS presents better performance than the DPWS and offers inferior results only when it is compared with the binary-based DPWS which, however, does not retain the WSs interoperability feature as it does not use XML documents.

Automatic Recognition of Boredom in Video Games Using Novel Biosignal Moment-Based Features

This paper presents work conducted toward the biosignals-based automatic recognition of boredom, induced during video-game playing. For this purpose, common biosignal feature extraction methods were exploited and their capability to identify boredom was assessed. Moreover, for the first time, Legendre and Krawtchouk moments, as well as novel moment variations, were extracted as biosignal features and their potential toward automatic affect recognition was examined using the specific application scenario. The present analysis was conducted with ECG and GSR data collected from 19 different subjects, while boredom was naturally induced during the repetitive playing of a 3D video game. Conventional biosignal features as well as moment-based ones were found to be effective for the automatic recognition of boredom by achieving classification accuracies around 85 percent. Then, the joint use of moments and moment variations with conventional features was found to significantly improve classification accuracy by producing a maximum correct classification ratio of 94.17 percent.

Integrating Wireless Sensor Networks into Enterprise Information Systems by Using Web Services

This paper proposes an advanced middleware solution to the problem of integrating a Wireless Sensor Network into the information system of an enterprise at a high abstraction level. This is achieved by using the proposed middleware which provides to the wireless sensors a Service Oriented Architecture connection to the Internet. The proposed middleware is based on the Device Profile for Web Services which is a Service Oriented Architecture technology at the device level. Since this technology is based on exchanging eXtensible Markup Language documents, a technique is utilized which compresses and reduces the data volume of such documents at a level that can be handled by the use of the resource constrained environment of the wireless sensors. By utilizing the proposed middleware which implements only the basic functions of the Device Profile for Web Services, we demonstrate how such a Wireless Sensor Network can be connected to the Internet achieving in this way its integration into an enterprise information system in which all its components conform to a Service Oriented Architecture standard.

Using activity-related behavioural features towards more effective automatic stress detection

This paper introduces activity-related behavioural features that can be automatically extracted from a computer system, with the aim to increase the effectiveness of automatic stress detection. The proposed features are based on processing of appropriate video and accelerometer recordings taken from the monitored subjects. For the purposes of the present study, an experiment was conducted that utilized a stress-induction protocol based on the stroop colour word test. Video, accelerometer and biosignal (Electrocardiogram and Galvanic Skin Response) recordings were collected from nineteen participants. Then, an explorative study was conducted by following a methodology mainly based on spatiotemporal descriptors (Motion History Images) that are extracted from video sequences. A large set of activity-related behavioural features, potentially useful for automatic stress detection, were proposed and examined. Experimental evaluation showed that several of these behavioural features significantly correlate to self-reported stress. Moreover, it was found that the use of the proposed features can significantly enhance the performance of typical automatic stress detection systems, commonly based on biosignal processing.

Mining affective needs of automotive industry customers for building a mass-customization recommender system

Mass customization systems aim to receive customer preferences in order to facilitate personalization of products and services. Current online configuration systems are unable to efficiently identify real customer affective needs because they offer an excess variety of products that usually confuse customers. On the other hand, mining affective customer needs may result in recommender systems, which can enhance existing configuration systems by recommending initial configurations according to customer affective needs. This paper introduces a mass customization recommender system that exploits data mining techniques on automotive industry customer data aiming at revealing associations between user affective needs and the design parameters of automotive products. One key novelty of the presented approach is that it deploys the Citarasa engineering, a methodology that focuses on the provision of the appropriate characterizations on customer data in order to associate them with customer affective needs. Based on the application of classification techniques we build a recommendation engine, which is evaluated in terms of user satisfaction, tool’s effectiveness, usefulness and reliability among other parameters.

Implementation of model predictive control using real-time multiprocessing computing

Abstract This work addresses the issue of finding the programming procedure that results to the fastest implementation of the core calculations of the model predictive control (MPC) algorithms that are amenable to parallel processing on a real-time multiprocessing system. Three concurrent programming procedures were considered for the MPC implementation, each one consisting of a number of tasks configured to a linear array, mesh and tree parallel architectures. Both theoretical analysis and measurements, taken from running these procedures on a four-processor computer platform, indicate that the procedure of the linear array architecture presents the best speed-up ratio.

Soft-testing of industrial control systems programmed in IEC 1131-3 languages.

This paper presents a technique and the means for testing PLC-based control software outside the actual plant environment with the purpose of increasing the confidence level on the compliance of the software to functional and temporal requirements. The need to obtain a high confidence level on the correct software operation arises from the fact that in most of the cases it is quite dangerous and expensive to test unproved PLC operation by linking it with the actual facilities that it is going to control. The proposed technique relies on a combined simulation of the controlled plant and the PLC system and an analysis of the plant responses. The PLC system simulation imitates the way software is executed on a PLC that is programmed in the languages of the IEC1131-3 standard. It is based on the programming model of the IEC standard and analytical formulae for estimating the program runtime. The simulation of the plant is based on a discrete convolution model that is solved at the same rate with the rate determined by the control algorithms. A tool realizing these concepts has been developed and its use in testing the control software of three critical outputs of a distillation column is demonstrated.

An interactive electronic book approach for teaching computer implementation of industrial control systems

There is an overall consensus on the importance of laboratory work that exposes the students to broader and more practical issues of industrial control systems, such as their implementation by distributed computer systems (DCSs) and programmable logic controllers (PLCs). However, setting up appropriate laboratory facilities to serve this purpose is expensive. For this reason, an interactive learning environment has been developed around the concept of the electronic book. The architecture of the environment allows the integration of hypertext with simulators of DCS, PLC, and process operation. The simulators are specially designed to serve an application-oriented teaching approach, which involves the student in the simulation setup and the running of the application. They are able to simulate not only the execution of the software that realizes the regulatory control algorithms but also the start-up and emergency control strategies of an industrial process, the manual, automatic, and cascade modes of controller operation, and the man-machine interface of a DCS- or PLC-based control system. The applications on which the teaching of DCS and PLC-based control system implementation is based are the interactive advanced control of a distillation column and the pH control of a reactor solution.

Adaptive predictive control using recurrent neural network identification

This paper presents a new adaptive predictive control algorithm which consists of an on-line process identification part and a predictive control strategy which is updated every time a process model change is identified. The identification method is based on recurrent neural network (RNN) nonlinear AutoRegressive with eXternal input (NARX) model derived from dynamic feedforward neural network by adding feedback connection between output and input layers. Two model-based predictive control strategies have been studied: the generalized predictive controller (GPC) and nonlinear adaptive model predictive controller (NAMPC). The neural network training and validation data are obtained from the open-loop simulation of a validated first principles plant model. The identified neural network (NN) model is validated using the following three different validation algorithms: (1) one-step ahead cross-correlation; (2) Akaikes final prediction error (AFPE) estimate; and (3) 5-step ahead prediction simulations. The algorithm has been applied to the temperature control of a fluidized bed furnace reactor of the steam deactivation unit of a fluid catalytic cracking (FCC) pilot plant used to evaluate catalyst performance. The validation results show that the RNN models the reactor to a high degree of accuracy. Simulation results show that the proposed NAMPC control strategy outperforms the GPC at the expense of extra computation time.