Sensing a Pattern

I was watching TV the other night when a promo came on for the police-procedural crime show, Criminal Minds. In the ad, Thomas Gibson, who plays the stoic FBI agent Aaron Hotchner, is looking at a white board when he says, “I’m sensing a pattern.” That’s not an uncommon phrase, but the detached, robotic way in which he delivered the line started me thinking about what we mean by the phrase and whether is actually has any implications for robotic design.

Whenever I hear the phrase used, it seems to mean that the person who is making the observation has chosen from a number of things he or she has seen and heard, a certain recurrence of events that led them to believe that there is a common cause for them or a causal relationship between them. Often, this realization is not the result of a conscious analytical process but rather, a subconscious one. The observer suddenly realizes there is a pattern of events that are related.

A key aspect of this pattern sensing is that the input on which it is based is often from a number of senses; something was seen or heard or even smelled or felt. By integrating this information from multiple sources, we are able to reach a conclusion that is greater than the sum of their parts.

Several years ago, the U.S. military began investigating what it called sensor integration. By this, they meant that input from several electromechanical sensors – acoustic, thermal, electromagnetic radiation, and so forth – would be used to determine the nature of a target, whether it was friend or foe. The pattern of this information could provide faster and more definite data on which to decide a course of action.

Today, we live in an era in which having multiple sensors in a handheld device is commonplace. In many cases, however, the information from those sensors is used singularly. For example, a Wii controller senses acceleration alone to determine a course of action. But, if we were to combine that acceleration information with other sensor inputs, say a thermal input from sensing the proximity of a hand to the handheld case, we’d know whether the device was being intentionally moved or had accidentally been dropped.

The opportunities for sensor integration are myriad, and the technology is now available to make such integration possible. By designing products with this in mind, engineers may be able to produce some very amazing and “intelligent” new devices.