Chapter 1: Introduction
Page 1
CHAPTER 1 INTRODUCTION 1.1 Definition of Sensors
Much effort is being applied to the development of intelligent, autonomous machines. It is standard practice in aircraft to use auto-pilots which independently control in-flight course corrections; in the automotive industry systems which automatically keep a safe distance between cars and maintain the vehicle in the middle of the road are in the experimental phase. Robots are increasingly autonomous and carry out tasks in environments that are hostile or extremely difficult to access for human beings such as space or radioactive areas. Factories that run 24 hours with only the supervision of very few humans are already reality. The sophisticated manufacturing process is entirely controlled and carried out by machines. Such complex artificial systems rely heavily on an interaction between the environment and the machine which reacts according to changing parameters in its environment. Consequently the sensing of these parameters is of utmost importance. It is a common feature of control system design that a system cannot be controlled more accurately than the method of measurement of the controlled variable. These sensing devices are commonly referred to as sensors or transducers.
A transducer normally measures the physical quantity of interest indirectly through its effect on one of the transducer’s parameters. Typically this change is then converted into an electrical signal, e.g. voltage or current which provide a measure for the quantity of interest. Fig. 1.1 illustrates this point.
Physical quantity
Voltage or current Transducer
Fig. 1.1: Principle of a transducer.
Chapter 1: Introduction
Page 2
1.2 Mechanical Sensors
One important class of transducers is mechanical sensors for measuring acceleration, velocity, position, pressure, weight, flow rate, force, sound, etc. The demand for these sensors has increased recently quite considerably because of the growing number of applications dependent on accurate measurements of physical quantities such as in plant and process control, monitoring systems, medical applications, etc. These sensors should provide high precision measurement data at both low cost and low power consumption. Consequently, much effort is being applied to research and development to improve existing designs for mechanical sensors and to find novel approaches.
1.3 Recent Innovations
Two technical advances can be identified to be of importance in the search for new concepts: first, the ever increasing use of digital signal processing (DSP) for which powerful tools are available and the enormous advantages of digital data transmission providing a robust signal even in an electrical noisy environment; second, the progress in micromechanics which, it is claimed, has similar potential for development as microelectronics.
Micromachined sensors often have a superior performance to their conventional counterparts in terms of robustness, reliability, accuracy, flexibility and sensitivity at reduced weight, dimensions and power consumption. Since the microfabrication process offers the possibility to manufacture in a batch process in which hundreds of sensors can be fabricated on a wafer, Sasayama et. al. [87], the cost of these devices is expected to fall to similar levels to that of components produced by the microelectronic industry.
A combination of micromachined sensors with the above mentioned advantages of digital signal processing yield “intelligent” or “smart” sensors, Brignell, J. H. [11]. The availability of such transducers will enable applications to become feasible which in the past were impossible or at least economically unjustified. Consequently, the economic potential of such sensors is
Chapter 1: Introduction
Page 3
considerable: according to Brignell the turnover of micromachined sensors alone in the USA was $750 million in 1989 and the predicted growth rate is between 10 % and 15 % a year. A more recent figure, Knutti, J. W. [41], quotes a U.S. market value of $252 million in 1996 and $870 million in the year 2000 just for one type of micromachined sensor: the accelerometer. Reuber, C. [82] quotes a figure for the European Market where in 1998 an estimated twenty million micromachined accelerometers are required just for one particular application: airbag release in cars.
In the near future rapid technical and economical development in the field of sensors and transducers is commonly expected, according to Bau, H. H. [6] ‘ ... sensor technology has developed tremendously in a relatively short period of time. But we have only scratched the surface.’
1.4 The Accelerometer
One important mechanical transducer, already mentioned above, is the accelerometer. It provides a measure of acceleration in form of an electrical output signal. One reason this sensor is of special significance is the fact that by integrating the output signal, an accelerometer can also provide a measure of velocity and position, fig. 1.2.
Electrical Signal ~
Acceleration Accelerometer
Acceleration
I
Electrical Signal ~
Velocity
I
Electrical Signal ~
Position
Fig. 1.2: Accelerometers also provide a measure for velocity and position.
Chapter 1: Introduction
Page 4
As a result these devices are widely used in all sorts of terrestrial, marine and aerospace applications. Recently the automotive industry has shown high demand for accelerometers where the release of the airbag is the most well known application, Goodenough [34], but also for active suspension control, brake control, fuel cut-off and engine knock, MacDonald [71]. High precision inertial navigation and guidance systems are based upon accelerometers; they are also used in vibration control, e.g. if containers with fragile goods are shipped, to monitor the vibration of the hard disk in portable computers, to measure the vibration of machines and in the investigation of earthquakes, Kraft et. al. [46]. In aerospace accelerometers are used to sense microgravity in space laboratories, Kulzer et. al. [51]. In medical applications accelerometers can help to monitor the motion of a patient, e.g. suffering from Parkinson disease. The variety of applications for accelerometers is enormous and this list is far from complete.
To date the main limitations have been the cost, size, weight and power consumption of these devices. Consequently accelerometers have an especially high potential for improvements through new developments in micromachining because this technique addresses the problems mentioned above. However, if the number of accelerometers and other sensors in a plant increases significantly, another problem arises: the transfer and processing of a huge amount of data. Obviously, this can be solved by using microprocessors for the data processing and digital data transmission. Consequently a micromechanically fabricated accelerometer with a direct digital output signal would result in an ideal transducer for most of the applications described above.
1.5 Scope of this Research Project
In 1993 the aim of the Nonlinear Systems Design Group under direction of Dr. C. P. Lewis was to develop such a digital accelerometer. Additionally to the direct digital output signal the accelerometer should employ some sort of feedback for closed loop operation which yields the well known advantages over open loop devices such as an increase in bandwidth, dynamic range and linearity. Quite often accelerometers and other sensors are closed loop devices having an analogue output signal which is then subjected to an analogue to digital converter (ADC). From the systems engineering point of view the entire transducer is a chain of the closed loop sensor
Chapter 1: Introduction
Page 5
Fig. 1.3: A chain of a closed loop sensor and an ADC results in an open loop system.
and an ADC, consequently many of the advantages of having a closed loop structure are lost. Fig. 1.3 illustrates this point. In this work a closed loop accelerometer is described which incorporates the analogue to digital conversion within the loop, to produce a true digital transducer. However due to the complexity of this design procedure the research programme was divided in three stages: ‚
design of an open loop accelerometer,
‚
design of a closed loop, analogue accelerometer,
‚
and finally the design of an inherently digital, closed loop accelerometer.
For the design of these devices the following key factors could be identified: ‚
choice of the sensing element,
‚
method of signal pick-off,
‚
reset mechanism (for the closed loop devices),
‚
suitable form of compensation (for the closed loop devices),
‚
system stability, static and dynamic performance.
Each stage comprises the development of a mathematical model, simulation of the accelerometer, implementation in hardware, measurement and testing.
The design approach used in this research work was attempted mainly from the control engineering’s point of view as these devices have a closed loop structure. The last stage of the
Chapter 1: Introduction
Page 6
project, the design of a digital accelerometer, was an especially challenging task because the device is a highly nonlinear, discrete system.
Coventry University had not the facilities to manufacture micromachined structures and therefore industrial support was sought. Druck Ltd. in Leicester showed interest in such a project because they were currently developing the micromachined sensing element of an accelerometer. They agreed to supply samples of the micromachined prototype sensing element and consequently the research project, which is subject of this thesis, was started in January 1994.
