More of the same

Citation

Loughlin, C. (1999), "More of the same", Sensor Review, Vol. 19 No. 2. https://doi.org/10.1108/sr.1999.08719baa.001

More of the same

More of the same

An optical shaft encoder comprises radial black lines on a transparent disc. With a single optical sensor detecting the lines you can tell how fast the shaft is rotating at any moment in time. Add a second sensor and you can tell not only its direction of movement but also keep track of how far it has rotated relative to its starting position. Add a third sensor that detects a reference mark and you can determine the absolute position of the shaft as well as its direction of movement, speed and acceleration. The three sensors therefore combine to give you all the information that you need about the shaft's movements. Further sensors may be added to perhaps improve the resolution of measurement but basically three sensors have the problem solved.

The important point of the above is that while the single sensor gave you the instantaneous speed of rotation, as more of the same type of sensor was added you actually gain different information.

The combination of similar sensors and combination of different sensors is the theme of this issue. Both can give you the fortuitous result that the whole is greater than the sum of its parts.

If you are setting out to design a sophisticated sensor with very precise characteristics then a number of possible routes will present themselves. The first is to develop a single sensor that performs exactly as required, which is immune to other stimuli such as temperature, vibration, light and pressure. A second is to develop a sensor which compensates for these other stimuli, for example my measuring the temperature and scaling the sensor's output accordingly. And a third will be to use a number of sensors with different but well known characteristics and compare their outputs to arrive at a more refined measurement.

If you have three sensors and if sensor A responds to apples and pears, B to bananas and pears and C to oranges and pears, then if all three give an output when sniffing a fruit sample you know you are looking at a pear. Or at least you hope you are because all three might also respond to the aftershave of the person who last handled the fruit. Add in an extra sensor that responds to all known brands of aftershave but not to fruit, and you are back in business.

Gas sensors provide an easy example of multi-sensor systems and "noses" have been covered before and are updated in this issue. But the same idea can readily be adapted for other sensing needs and the attraction of the approach is that you may be able to sense something for which no sensor exists simply by using a combination of well established and low cost simple sensors.

So next time you are asked to detect the lateral position of a chocolate on a conveyor belt, see if you can do it with two simple light beams or if you really need it, structured lighting and a machine vision system. Answers on a postcard please.

Clive Loughlin