Home > Research > Sensors and Actuators
Sensors and Actuators
In the shrinking world of MEMS, devices get smaller, better, cheaper, and more amazing all the time. Improvements have been made in materials, machining techniques, and testing capabilities. But one important piece lags the rest—the part of the device that acts as a transducer. In fact, according to Professor Ilene Busch-Vishniac, in virtually all measurement and control systems the sensors and actuators account for the bulk of the cost, the limitations on size, and the majority of system failure situations. Her recently published book, Electromechanical Sensors and Actuators, aims to remedy that situation by taking a novel approach to the theory and modeling of these devices.
A transducer, for the less mechanically-minded, is a device that takes energy in one form (mechanical, electrical, optical, magnetic, chemical, or thermal) and converts it into another form. A telephone has two transducers. A microphone in the mouthpiece takes the sound energy from your voice and translates that into an electrical signal. The earpiece converts the electrical signal coming down the wire into a mechanical sound wave. Transducers in a clothes dryer sense when the clothes are dry and switch the dryer off automatically. Transducers are typically separated into sensors and actuators: sensors monitor something about a system, ideally without altering the system in the process, and actuators impose a state on a system. Typical actuators are motors, pumps, and force heads. Sensors measure parameters such as temperature, humidity, flow velocity, pressure, or acceleration. Sensors and actuators are often used together, as part of a measurement and control system. Cars are full of them. In a passenger side airbag, for example, there is a tiny accelerometer that is connected to a deployment mechanism for the airbag. If the acceleration exceeds some threshold, then the airbag is deployed. The thermostat in the engine, the antilock brake system and keyless door locks are all systems governed by sensors and actuators.
To improve the performance of sensors and actuators, argues Busch-Vishniac, we need to revisit the way we think about them. Traditionally, engineers study transducers by categorizing them into what they do: temperature sensors, accelerometers, motors, pumps, and so on. She argues that it makes much more sense to look instead at the fundamental coupling mechanisms that link the electrical and mechanical domains, rather than at specific sensors and actuators that are already in use. She categorizes transducers by the material or structural behavior that leads to transduction. Often a single principle can be applied in many ways to achieve various different sensing and actuating outcomes. Looking at just one of those outcomes, say, humidity sensing, limits the potential use of the same mechanism for other applications.
Busch-Vishniac takes a “systems dynamics” approach, centered around the energy in the system and the parameters that can be varied to translate that energy into different domains. Instead of the traditional circuit models used to describe energy flow through transducers, she takes a unique modeling approach known as Bond Graph Modeling. These bond graph models don’t assume linearity, and they are capable of describing causal relations as well as conservation equations. Similarities and dominant effects stand out, and they give a powerful visual picture of the way a system works, without solving all the equations involved. Because they can show causal relations, bond graph models identify immediately the information needed to create the system state equations much more accurately than typical circuit models.
The design of transducers hasn’t kept pace with the incredible advances in electronics in the past few decades, indicating that the current approach is limited and needs reevaluation. The demand for automated sensing devices, and the increasing importance of electronic devices in our everyday lives, gives an added impetus to the problem. Busch-Vishniac’s book provides a much-needed fresh look at the issue.