view). Nevertheless, in terms of industrial applications, other characteristics be- sides the transduction principle become more important, namely technology, ma- terials, accuracy, and costs. The insides of sensor systems become less important from an application perspective. All sensor systems are facing a noticeable upward trend in performance require- ments for maintenance, down-time, reliability, fault tolerance, fault recovery, and adaptability. Industrial applications in the area of consumer goods are also subject to cost pressure, where the value of electronic products has risen while costs have continuously fallen in past decades. Current trends in sensor systems can be summarized as follows (see Figure 1.1-10): 1. Higher performance of sensors. Sensor accuracy and linearity, both classical prob- lems of sensor systems, have been improved in all areas. Today’s sensor systems also work under better conditions and with higher reliability. Greater demands for environmental protection have led to the development of highly reliable sensors. Development efforts are still focused on maintenance-free sensors with long life expectancy and low electric power dissipation. New sensors have been developed, such as torque sensors, accelerometers, acoustic sensors, pattern information-sensing devices, and smell sensors. Olfac- tory sensors and voice recognition combined with synthetic speech generation technology will make remote-controlled home appliances a reality. 1.1 Sensors in Intelligent Buildings: Overview and Trends 19 Fig. 1.1-10 Trends in sensor systems 2. Smart sensor systems. Smart sensor systems are usually complex systems with in- tegrated intelligence. Sensor technology has developed towards a stronger integra- tion of signal transformation and signal processing. Microprocessors and sensors build an integrated unit for measuring and controlling technology. Sensors and signal processing have to be brought together at the place of measurement. These ‘smart sensors’ are characterized by an integral electronic design, optimized for defined applications. As a result, higher speeds will permit sensor signal proces- sing. In addition, the integration of sensors and related electronic circuits will lead to lower costs. Intelligent sensor systems rely on a set of algorithms and rules (knowledge base) which connect the signal enhancement from the sensors to the sensor out- put. Two kinds of knowledge are important: · empirical knowledge, which includes model-based calculations, hard rules about the process, and soft (fuzzy) rules based on experience; · analytical knowledge, which includes extrapolation from the past, involvement of the neighborhood, and learning from success (analytical of neuronal). 3. Multi-sensor systems. R&D in sensor technology is tending towards self-identify- ing, self-diagnosing, and self-calibrating sensors. Today there are already several sensor systems which integrate different transduction principles in a single sen- sor system, ie, mechanical, thermal, and optical. The output signals of several simple sensors are computed in parallel by a secondary processing unit. These multi-sensor systems create powerful and precise systems with high selectivity. New algorithms for pattern recognition support built-in self-supervision and in- crease the reliability of the sensor system. 4. Miniaturization of sensor systems. Driven by high-tech sectors, such as the aero- space, medical, and entertainment industries, sensor systems have been pushed towards smaller scale devices. Examples are the integration of optical sensors in fi- ber technology and intelligent noses with biochemical sensors. In the future the importance of nanotechnology for sensor systems will increase further and open up access to new and smaller dimensions. 5. Integration of sensor-actor systems. Sensor systems are becoming smaller in size and lighter in weight, which allows further integration of sensor and actor sys- tems. This trend became especially important in the area of mechatronics, a fu- sion of mechanical and electronic areas at a miniaturized level. 6. Standardization of sensor interfaces. Across industries and applications, the impor- tance of standardized interfaces has been on the increase. Economies of scale due to mass production of sensors have led to lower costs. Even complex sensor sys- tems have become commodity items because of standardized interfaces and inte- grated designs for sensor systems and control systems. Standardization allows eas- ier handling of sensor systems. Some industries support elaborate and advanced standards, such as the CAMAC system in the nuclear industry. The standardiza- 1 Introduction20 tion constraints, however, mainly apply to the communication aspects of sensor systems (external interfaces); the internal operation of industrial sensor systems remains open to free development [12]. 7. Cost reduction in sensor systems. A direct impact of standardization is high-volume production and thus cost reduction. In general, sensor technology has become more smart, integrated, and standardized. Industrialization and the trend towards more functionality at lower cost are the main drivers. The automotive industry has already demonstrated these trends. When the topic of intelligent sensors came up in the field of instrumentation in the late 1970s, it was greeted with some skepticism in industrial communities because of the cost and the current stage of development of computer technology [12]. Today, low-cost systems pro- duced in high volumes provide both intelligence and reliability. The enablers of these trends are new developments such as material technol- ogy, sensor and actuator technology, microelectronics/optoelectronics/information storage, electro-optical and electromechanical transducers, generation/distribu- tion/storage of electrical energy, information engineering, intelligence functions, and software. Major drivers include large efforts in research and development of high-tech industries in the area of microelectronics in the 1970s, in new materials and information technology in the 1980s, and in miniaturization and integration in the 1990s [13]. Huge improvements have been made in the manufacturability of sensor systems. This includes the integration of sensorics, actuators, ASICs, and mechatronics as well as communication elements into one system. At the same time, the complexity of the sensor-actor systems has increased drama- tically. The management of the increased complexity in R&D laboratories will be a critical success factor in the future, according to Bill Raduchel, Chief Strategy Offi- cer of Sun Microsystems: ‘The challenge over the next 20 years will not be speed or cost or performance; it will be a question of complexity.’ Future sensor systems will be independent, teachable, or adaptable systems, which will extend their geometries into the nano range, include non-electrical signal processing components, and will be based partly on materials from high-temperature electronics [13]. 1.1.4 Sensor Systems in Intelligent Buildings In the following sections of this book we focus on those areas of an intelligent building where we expect the biggest impact of sensor and control technologies. These are · energy and HVAC; · information and transportation; · safety and security; · maintenance and facility management; · system technologies. 1.1 Sensors in Intelligent Buildings: Overview and Trends 21 1.1.4.1 Energy and HVAC Albert So and Brian Tse describe intelligent air-conditioning control systems. After a review of conventional physical sensors, three new concepts and their associated de- velopments in sensing technology are discussed. Computer vision traces people movement and people flow in order to make the most intelligent decision for opti- mal HVAC control. Internet-based control permits fully remote monitoring and op- eration of building systems. Comfort-based control is highlighted as the future stan- dard; the authors suggest the use of the predicted mean vote (PMV) as comfort index. A self-commissioned heating control system, called NEUROBAT, is introduced by Jens Krauss, Manuel Bauer, Jürg Bichsel, and Nicolas Morel. The heating control- ler is predictive and adaptive and thus able to achieve energy savings and ensure optimal thermal comfort. Artificial neural networks permit the adaptation of the control model to real conditions, such as climate, building characteristics and user behavior. First empirical results at a real site have shown huge energy savings. Hanns-Erik Endres focuses on the measurement and management of air quality. Gas sensors play an important role in this. An optimal sensor-based management of energy and thermal comfort in build- ings is discussed by Thomas Bernard and Helge-Björn Kuntze. The automatic and remote reading of consumption data recorded by meters in- stalled in residential buildings is an important element of energy management. Dieter Mrozinski gives an overview of the opportunities of wireless and M-bus-en- abled metering devices. Since user behavior regarding energy consumption can be positively influenced by metering individual energy consumption, sensors for metering are recom- mended by the European Union. Günter Mügge describes such sensors and the means of energy cost allocation. Pressure sensors for the HVAC industry are described by Yves Lüthi, Rolf Meisin- ger, Marc Wenzler, and Kais Mnif. Special requirements from the HVAC industry are low pressure ranges, simple installation, independence of mounting orienta- tion, low sensitivity to dirt, robustness, a long lifetime, and low costs. Several solu- tions of pressure sensors which meet these requirements are discussed. 1.1.4.2 Information and Transportation Fieldbus systems enable decentralized control structures and therefore often lead to lower system costs and higher system reliability. Dietmar Dietrich, Peter Fischer, Dietmar Loy, and Thilo Sauter give an overview of fieldbusses. This includes the basics of the communication techniques, used in fieldbus systems, as well as a short history and an overview of EIB and LonWorks. Challenges which still re- main include interoperability and demands for agent-based systems. Wireless in-building networks are discussed by Mike Barnard, who focuses on low-power/low-bandwidth networks suitable for sensors, telemetry, etc. He also 1 Introduction22 gives an overview of existing and emerging standards and products, related to wired and wireless links from and to sensors. Sensor systems in modern high-rise elevators are discussed by Enrico Marchesi, Ayman Hamdy, and René Kunz. In order to travel higher and faster than ever be- fore, elevators need excellent shaft information systems. Modern elevators travel at 10 m/s up to 500 m in height with horizontal vibrations of less than 10 mg. The authors explain an active ride control system, which relies heavily on fast and accurate sensors. The perceptual user interface of a future intelligent office building includes in- formation on all elements in a room. Opportunities and first empirical results of sensing chair and floor interaction by using distributed sensors are presented by Hong Z. Tan, Alex Pentland, and Lynne A. Slivovsky. The key problem is the auto- matic processing and interpretation of touch sensor information and the model- ing of user behavior leading to such sensory data. 1.1.4.3 Safety and Security Marc Thuillard, Peter Ryser, and Gustav Pfister give an overview of life safety and security systems in intelligent buildings. These systems provide early warning of dangerous situations in buildings, eg, fire incidents, toxic or explosive gas concen- trations, and unwanted intrusion into premises. They discuss technologies and physical principles of those sensors, including the complete alarm systems and emergency systems involved. Biometric sensing systems represent innovative technologies for access control. Christoph Busch describes opportunities of biometric authentication for access control. Pattern recognition is one of the key areas. Smart cameras and their applications in intelligent buildings are outlined by Bedrich J. Hosticka. Smart cameras possess built-in intelligence and allow en- hanced imaging under critical light conditions, as well as low-cost image process- ing. New camera technology, called CMOS-based imaging, is introduced by the author. Peter L. Fuhr and Dryver R. Huston outline new sensing systems and techniques for improved construction site safety. Sensor systems are designed for monitoring construction site shoring and scaffolding. Such a sensor network can provide in- formation about the load distribution on shoring systems and thus increase per- sonal safety. The solutions presented have already been tested on field sites. 1.1.4.4 Maintenance and Facility Management New forms of maintenance management in industrial installations are described by Jerry Kahn. Predictive maintenance and periodical condition monitoring are the key elements. Sensor systems are required to measure and analyze equipment conditions and predict future equipment performance. 1.1 Sensors in Intelligent Buildings: Overview and Trends 23 Rolf Reinema from the GMD Institute for Secure Telecooperation discusses the opportunities presented by world-wide facility management. Office environments and their interiors are monitored, controlled, and managed from a globally stan- dardized platform, abstracted from particular local technology. A smart, embedded system, called roomServer, bridges the gap between physical and virtual work en- vironments. 1.1.4.5 System Technologies General trends in sensor systems in intelligent buildings are summarized by Hans-Rolf Tränkler and Olfa Kanoun . Friedrich Schneider et al. outline system tech- nologies for private homes. Future buildings will become more intelligent. The main trends will be reduced resource consumption, optimized convenience, and more comfort as well as better safety and security. Following other industries such as the automotive industry, the building industry will experience an increased impact of microsystems technologies and new communication systems. Within these trends, new sensor systems with higher performance characteristics will be key technologies and enablers. The per- formance of these sensor systems will increase regarding accuracy and linearity, maintenance, down-time, reliability, fault tolerance, fault recovery, and adaptabil- ity. Sensor systems themselves will become smarter and more integrated. Multi-sen- sor systems will be increasingly widespread, used with self-identifying, self-diagnos- ing, and self-calibrating sensors. Within the next 10 years megatrends will be min- iaturization, standardization of interfaces, and the close integration of smart sensors with actuators. Higher volumes will lead via economies of scale to low-cost sensor systems. Therefore, most visionary sensor system applications which are described in this book will become tomorrow’s commodity products. 1 Introduction24 1.1.5 References 1 Kaye, J., Working paper; Boston: MIT Media Laboratories, 1998. 2 Streitz, N. A., Geissler, J., Holmer, T., Konomi, Sh., Müller-Tomfelde, Ch., Reischl, W., Rexroth, P., Seitz, P., Steinmetz, R., in: Proceedings of the ACM Conference on Human Factors in Computing Systems; Pittsburgh, PA: May 15–20, 1999. 3 OECD, 21st Century Technologies: Promises and Perils of a Dynamic Future; Paris: OECD, 1998. 4 Festo, presented at IASS International Association for Shell and Spatial Struc- tures Symposium, University of Stutt- gart, 7–11 October 1996. 5 Festo, Portable Architecture; London: Roy- al Institute of British Architects, 1997. 6 Festo, in: Metropolis; New York, Decem- ber 1998, pp. 45–49. 7 Festo, Archit. Rec. (1999) 218. 8 Festo, Design Net 30 (2000) 145. 9 Kobayashi, T., in: Sensors. A Comprehen- sive Survey, Göpel, W., Hesse, J., Zemel, J. N. (eds.); Weinheim: Wiley VCH, 1989, Vol. 1, pp. 425–443. 1.1 Sensors in Intelligent Buildings: Overview and Trends 25 10 Lion, K. S., IEEE Trans. IECI-16 (1969) 2–5. 11 Grandke, T., Hesse, J., in: Sensors. A Comprehensive Survey, Göpel, W., Hesse, J., Zemel, J. N. (eds.); Weinheim: Wiley- VCH, 1989, Vol. 1, pp. 1–16. 12 Brignell, J.E., Smart Sensors, in: Sen- sors. A Comprehensive Survey, Göpel, W., Hesse, J., Zemel, J. N. (eds.); Weinheim: Wiley-VCH, 1989, Vol. 1, pp. 331–353. 13 Meixner, H., in: Sensors. A Comprehen- sive Survey, Göpel, W., Hesse, J., Zemel, J. N. (eds.); Weinheim: Wiley-VCH, 1995, Vol. 8, pp. 2–22. 2 Energy and HVAC Sensors in Intelligent Buildings. Edited by O. Gassmann, H. Meixner Copyright © 2001 Wiley-VCH Verlag GmbH ISBNs: 3-527-29557-7 (Hardcover); 3-527-60030-2 (Electronic) 2.1.1 Introduction A precise definition of a sensor (often called a detector in heating, ventilating, and air-conditioning (HVAC) control) has always been elusive. For practical purposes, a sensor in an HVAC system can be thought of as a device which converts a phys- ical property (eg, temperature or humidity) or quantity (eg, flow rate) into a conve- niently measurable effect or signal (eg, current, voltage, or number). For HVAC, sensors can be grouped into the following types: · temperature and humidity (enthalpy); · pressure; · flow rate; · comfort; · indoor air quality. Most sensors consist of two ‘components’, ie, a transducer which converts the raw, measured signal into a ‘convenient’ signal (usually electrical), and an asso- ciated signal conditioner which ensures that the raw signal is converted to a scal- able electrical signal which can be calibrated with the raw measured signal. Typi- cally, a linear relationship between the convenient signal and the quantity of the raw, measured signal is preferred. Recent and future generations of sensors have an additional feature, ‘intelligence’, where a built-in microprocessor enables data to be reported and analyzed besides pure measurement. Comfort and enthalpy ‘sensors’ are examples of this category. 2.1.2 General Specifications of a Sensor [1] A sensor can be fully specified with reference to at least 12 performance, practi- cal, and economic factors which can be divided into two classes. 29 2.1 Intelligent Air-conditioning Control Albert T. P. So and Brian W.L. Tse , City University of Hong Kong, Hong Kong Sensors in Intelligent Buildings. Edited by O. Gassmann, H. Meixner Copyright © 2001 Wiley-VCH Verlag GmbH ISBNs: 3-527-29557-7 (Hardcover); 3-527-60030-2 (Electronic) Performance-based Factors 1. Range. The range of the measured variable for which the characteristics are maintained at stated values. 2. Accuracy. The degree of equivalence to which the measured output matches with some known benchmark. 3. Repeatability. The ability of the sensor to reproduce consistently the same out- put from the same measured value. 4. Sensitivity. The smallest detectable change in measured value that results in outputs change by the sensor. 5. Drift. The degree to which the sensor fails to give a consistent performance throughout its stated life. 6. Linearity. The closeness to linear proportionality between the output and the measured value across the range. 7. Response time. The rate of response with respect to time of the output follow- ing an input change (often expressed as a time constant). Practical and Economic Factors 8. Cost. The cost should include power supply, transducer, the signal condi- tioner, and the connecting cables. Very often, the cost of installing the sensor consumes a very significant portion within the overall cost. 9. Maintenance. Any special maintenance and re-calibration requirements in- volving additional labor and expenses. 10. Compatibility. Interoperability and interchangeability with other components and standards. 11. Environment. The ability to withstand harsh or hazardous environments. 12. Interference. Susceptibility to ambient ‘noise’ such as electromagnetic waves or quasi-stationary electric or magnetic fields. 2.1.3 A Quick Review on HVAC Sensors A brief description of various sensors installed in HVAC systems is as follows [1– 12]. 2.1.3.1 Temperature Sensors Temperature is an important controlled parameter inside an air-conditioned envi- ronment because it is closely related to human comfort level as suggested by Fan- ger [3]. Obviously, temperature sensors are the most common sensors in an air- conditioning system. Usually, three types of temperature sensors are popular, namely thermocouples, resistance temperature detectors (RTDs), and thermistors. 2 Energy and HVAC30 . [13]. 1.1.4 Sensor Systems in Intelligent Buildings In the following sections of this book we focus on those areas of an intelligent building where we expect. and security systems in intelligent buildings. These systems provide early warning of dangerous situations in buildings, eg, fire incidents, toxic or explosive