Smart.Materials.and.New.Technologies Part 8 pdf

At the risk of excessive oversimplification, a typical system of sensors and actuators that are intended to accomplish specific tasks consists of elements providing a number of different

5.2 Control systems

BASIC FEATURES

We have so far briefly explored how different components of a complex system work Here we look at how more complex assemblies function This is a huge topic that is ultimately the domain of the engineering profession and beyond the scope

of this book Here we only look at basic characteristics and issues critical to how designers use these systems

At the risk of excessive oversimplification, a typical system

of sensors and actuators that are intended to accomplish specific tasks consists of elements providing a number of different functions:

* Sensors and transducers

* Signal conditioners

* Transmitters/converters/receivers

* Logic controllers

* Displays/recorders/actuators

In the traditional mechatronic (mechanical-electronic) model, several of these functions are provided by individual components that are interconnected and provide an overall desired response The roles and operations of sensors, transducers and actuators have already been briefly explored The output of a sensor that is responding to some energy stimulus may or may not be in a readily usable or interpretable state Transducers may or may not be needed to change the sensor output signal to some other energy state The output signal is also usually in need of further conditioning to boost its amplitude, filter out unwanted ‘noise’ or other reasons The conditioned signal invariably needs to be transmitted else-where so that it can be used directly as an input to something else, which in turn means that it has to be received elsewhere Transmitting and receiving devices are thus obviously needed These processes may again require signal conversion Actuation devices could then be directly activated

In the enhanced mechatronic model that is more sophisti-cated, transmitted signals that have been conditioned are first manipulated according to a logical intent and then trans-mitted to final actuators Of special interest here are the logic controllers It is here that the system is ultimately given its directions A designer might want a motion detector to cause

an alarm to go off when movement is detected, or a door to open when motion is sensed Operations of this type can be done using actual hard-wired circuits that use common

s Figure 5-14 Differerent input/output control models (from the simple sensor/actuator system to the biological system with its intricacies)

electrical devices to perform a series of operations, such as an inexpensive timer circuit where a charge builds up in a capacitor and is then periodically released, which in turn causes some action to occur Circuits with surprisingly intricate logic capabilities can be built in this way

An advantage of the use of both property-changing and energy-exchanging smart materials within the context of sensor/actuator systems is that many of the actions described occur internally within the materials themselves In some cases the sensor/actuation cycle is completely internalized In other cases, additional elements may be required for certain kinds of responses, but the complexity of the system is invariably reduced

For Type 1 property-changing materials, there are intrinsic and reversible property change responses The constitutive model I, shown in Figure 5–14, can be used to describe basic input and output relations for Type 1 smart materials When the full advantages of a complete control system are desired, including programmable logic capabilities or closed loop behaviors, it is clear that energy-exchanging smart materials that can generate electrical signals would be a seamless improvement While the behaviors of energy-exchanging materials are not normally programmable in the accepted sense of the word, they can easily become part of a complex system and still serve to reduce its overall complex-ity The constitutive model II, shown in Figure 5–14, can be used to describe basic input and output relations when these kinds of Type 2 energy-exchanging materials are used

It is interesting to note that more and more research is directed towards making as many overall system behaviors as internalized as possible Ultimately, smart materials offer the possibility of making the overall system seamless The biological model diagrammed in Figure 5–14 suggests the fundamental nature of this kind of completely internalized environment

OPEN AND CLOSED LOOPS

For complex systems or where there is a need to frequently change logical parameters, logic controllers inherently involve either direct connection to an external computational envir-onment or do so via an internal microprocessor For example,

a designer might want an activated motion detector to cause

a series of different lights to blink at different rates A simple program could be written that takes the conditioned signal as

an input and then uses it as a trigger to control different output signals to the final lights The sequence and rate of

light activation could be directly programmed Many of these programming environments use common interpreted pro-gramming languages, e.g., Basic or C, while others use more direct programming techniques Once a programming envir-onment is introduced, the control possibilities are exciting since all manner of operations can be envisioned (see Microprocessors below)

A control system can have many features Of particular importance herein is the distinction between open loop and closed loop systems In a simple open loop system, the input

is processed to produce a desired response Thus, a sensor might cause a slider on a mechanical device to move 25 mm

or for a shaft to twist 15o In another case, a sensor might cause an actuator to open a door In an open loop system the input system may be processed and an output signal sent to

an actuator to cause the desired action, but there is nothing inherent in the system that checks to see if the desired action actually took place Did the slider move 25 mm or did the door actually open? If something prevented the final action from occurring, there is no way that the system would know this

A closed loop system has additional features that allow the system to check to see if the intended output action did indeed occur in response to the input as desired and anticipated Thus, in the door opener or slider example, there might be some type of position sensor or limit switch that detected whether the door was in the open position or that the slider moved 25 mm as desired This information would be fed back into the system and the original state compared to the final state The comparison would then indicate whether the desired action took place If not, the system could be programmed to ‘try again,’ issue an error message, or do something else Feedback systems of this type normally involve use of microprocessors because of the logic involved, but a review of the history of automation suggests that there are many other ways, including mechanical means,

by which feedback control can be obtained

MICROCONTROLLERS The primary purpose of a microcontroller is to communicate

Microcontrollers are based on microprocessor technologies Microcontrollers are typically unseen to users, but they are buried in innumerable devices found in diverse settings such

as automobiles, home appliances or video equipment A microcontroller can be programmed to execute instructions in

a desired way, and thus to follow a series of sequenced operations that control the actions of a linked device Microcontrollers can be designed to be stand-alone devices that execute pre-programmed actions from memory, or they can be designed to interface directly between a primary user-controlled computer and an external device

A microcontroller converts inputs into outputs A micro-controller typically receives input signals and produces output signals that control an electrically activated mechanical device Microcontrollers invariably involve both hardware components and software that interprets inputs and controls outputs Microcontrollers come in a variety of types, sizes and capabilities A typical microcontroller might consist of a small microprocessor that has computing capabilities, a built-in capability for storing in memory a programming language interpreter (e.g., for Basic or other programming languages),

a series of input/output (I/O) pins to link to both input and output devices, and a built-in power supply or connection to

an external source A typical microcontroller can be repro-grammed at will via connection to a primary computer that houses the primary programming language Typically, pro-grammed instructions relate to how information from the input pins (say an electrical signal from a sensor) is manipu-lated and what signals are sent to each of the output pins, which in turn control connected devices

5.3 MEMS

(micro-electro-mechanical systems)

This idea of seamlessness has propelled the development

of integrated sensor–transducer products toward the incor-poration of computational intelligence Smart materials and microtechnology had adhered to parallel, albeit close, research and development tracks In a curious crossover from the information world to the physical world, a ubiquitous material emerged as the means to merge the two worlds Silicon, the workhorse material of semiconductors, revolutio-nized the communications and electronics industries when it was introduced due to its rather spectacular electrical proper-ties But silicon may well be even more compelling as a mechanical material Three times stronger than steel, but with

a density lighter than that of aluminum, silicon also has the near ideal combination of high thermal conductivity with low thermal expansion On a dimensionless performance level, silicon outperforms all other traditional mechanical materials Unique about silicon, however, was that there was an entire

fabrication industry already tooled for the manufacturing and machining of silicon components at the micro-scale Remarkable structures could be directly machined on a silicon chip, including microscopic gears, levers, drive trains and even steam engines! Millions of elements could be combined

s Figure 5-15 Two views of a spider mite crawling across the surface of MEMs devices The top view is of a comb drive, the bottom view is of a gear chain (Images Courtesy Sandia National Laboratories, SUMMiT

TM Technologies, www.mems.sandia.gov)

in a device no larger than a postage stamp Thus was born a micro-machine that had simultaneous electrical and mechan-ical functions

The term micro-electro-mechanical systems (MEMS) has come to describe any tiny machine, but the more precise definition is that a MEMS is a device that combines sensing, actuating and computing The earliest MEMS tended to lean toward one or another aspect, rather than equally addressing all three, such as the sensing primary accelerometers for air bag deployment, the actuating primary ink jet printers, and the computing primary analyzers for chemical analysis Many

of these early applications did not truly exploit the true potential of MEMS as a smart system, rather the fabrication capabilities simply allowed for miniaturization of more con-ventional equipment Today’s MEMS have much higher expectations, as they are being developed for unprecedented capabilities: navigation and control of unmanned flight, remote evaluation of the changing characteristics of environ-mental hazards such as forest fires, and implanted analysis and control of biological processes

One of the most interesting directions is the development

of micro energy systems A common problem among all electronics, systems, machines and any material with an electrical need is the provision of power Regardless of how small, how direct and how distributed a component may be, electricity must still be supplied When any device is miniaturized, its power needs, in terms of both voltage and current, are greatly reduced, but our traditional power supplies can not be correspondingly reduced in size Batteries have become a fairly standard accompaniment, often dwarfing the component being powered Micro-machines can perform almost any task on a smaller scale than a full size machine can do on a larger scale as long as the rules for kinematic and dynamic scaling are adhered to Smart dust could be part of a MEMS device with a rotor, and thus be able to fly to desired locations And unlike many tasks that require large amounts of force or power and cannot be scaled down, such as an automobile drive train, the electron movement inherent in electricity has no such large-scale needs A MEMS device may need just a few milliwatts of power, which can be easily achieved with tiny generators Labels associated with building-size HVAC equipment are now routinely associated with MEMS energy devices – we now have micro-compressors, chillers, heat pumps, turbines, fuel cells and engines While much of the early impetus for micro energy systems was for the replacement of batteries, there is growing interest in exploiting the energy transfer capabilities

of these devices directly For example, one of the defined goals for micro-power was to replace the heavy battery needed for the portable, albeit unwieldy, cooling systems that soldiers wear in extreme heat conditions If the power supply could be miniaturized, then why not the cooling system itself? Scaling of thermal behavior is much more difficult than that of kinematic behavior, nevertheless there are large research efforts currently proceeding in this area (These energy systems will be discussed in greater detail in Chapter 7.)

5.4 Sensor networks

If remote or local power generation will allow systems to become more autonomous, networks and webs are intended

to make them more interactive Smart dust enables the wide distribution of sensors and devices, but there needs to be a corollary effort in how the information they gather is processed and then acted on Obviously, a single particle of smart dust could combine all these activities into a single MEMS device known as a ‘mote’ with each one communicat-ing to a central station through wireless technology But if we were to push the idea of ‘smartness’ to its fullest potential, then the motes would communicate with each other and collaboratively decide which one needs to take action This is precisely the tenet behind the smart sensor web that many agencies are developing, including NASA and the Department

of Defense

For smart sensors to be effective in monitoring environ-mental or battlefield conditions, tens of thousands of them would need to be distributed Addressing each sensor individually would flood data banks with generally useless information that must be sifted through NASA tackled this problem in their design of the Mars Rover, when they recognized that fully networked communication between all the components, far from being usefully redundant, actually increased both the opportunity for failure and the severity of the consequences If, however, clusters of sensors commu-nicated among each other, decisions could be made locally and directly There are many models for this, from treating sensors as individual nodes in a cluster with each having a decision-making capability for events in that cluster, or the assignation of a master node that delivers instructions to neighboring nodes Regardless of the model, the intention is leagues beyond surveillance, as it will essentially allow for

‘remote control’ of our surroundings David Culler, at the University of California, Berkeley, perhaps has best described

Fan to remove excess heat

Peltier device at basal location on back of neck Recharging connection

s Figure 5-16 This ‘personal cooling and

heating’ device based on Peltier

technolo-gies is worn around the neck

the intent of smart sensor networks: ‘just as the Internet allows computers to tap digital information no matter where it is stored, sensor networks will expand people’s ability to remotely interact with the physical world.’1

5.5 Input/Output models

At this point it is useful to summarize the different input/ output models for the range of different approaches that have been discussed In more complex sensor/actuator systems not involving smart materials, the first model discussed was the

s Figure 5-17 Wireless sensor network Top image shows prototype for

a distributed sensor pod (NASA JPL) Bottom image shows ‘smart dust’ that functions as chemical sensors at the nano-scale level (Frederique Cunin, UCSD)

direct mechatronic model Here a sensor responds to external stimulus field, its output signal is then normally transduced, and there is a direct response by the actuator The actuator response is controlled directly by the sensed inputs In the direct mechatronic model, sensors, transducers and actuators are normally discrete components (others may exist as well, e.g., transmitters, receivers) The enhanced mechatronic model next discussed contains the same features as the basic mechatronic model but now contains a logic component normally reflected through a computational environment of one type or another Output responses are controlled by sensed inputs but can now be logically manipulated or controlled Note that this model can become quite sophisti-cated The introduction of a logic controller allows great control over actions This kind of environment even allows the introduction of ‘learned behaviors’, which in turn opens up the world of complex robotic actions, artificial intelligence and other sophisticated approaches

When smart materials are introduced, significant changes occur Here we see that the property-changing characteristic

of Type 1 smart materials means that the response itself is dependent upon both the input stimuli and the properties of the material The output response remains direct, resulting in

a constitutive model I When energy-exchanging smart materi-als are used, the sensor/transducer function of the basic mechatronic model are combined into a single internal action, and, in some cases, the whole sensor/transducer/ actuator function is combined into a single function An enhanced constitutive environment, or constitutive model II, results that is based on energy-exchanging materials It would normally include a logic function, which in most current applications would remain external to the material itself but would nonetheless control output responses

So what might all of these advances ultimately aspire to achieve? Perhaps an end aspiration might well be an emulation of the biological model itself We have briefly characterized the basic human sensory environment, and the actuation capabilities of humans and other biological forms are obvious The ‘center box’ for a biological model now becomes the neurological system itself Here there is complete internalization of all functions and logic controls into the ultimate seamless or one-part entity Clearly, we are an enormous distance away from this model, but the aspiration level is interesting Perhaps the basic question is that of the following: if the biological model represents the basic aspiration, why should we approach it via an emulation approach that fundamentally utilizes non-living materials?