Some time back we took a look at Plessey’s EPIC sensor, which can monitor electric fields of various kinds and strengths. They’ve identified a number of applications that can interpret the field signal and – hopefully – make smart decisions.
They recently announced a new application that I certainly could have used over the summer as I went on a 7500-mile roadtrip: the ability to detect whether a driver is sleepy. It’s sort of amusing that they talk in terms of this being an improvement over a traditional ECG in the car: I can’t say as I’ve ever seen someone driving down the road with shirt off and wires hanging off of gel contacts. <Stops to reflect again…> Nope… pretty sure I never saw that.
That aside, the software required to process the signal and interpret what’s going on sounds every bit like sensor fusion software, except that it’s not fusing more than one sensor. Just as with AMS’s lightning sensor, they have to process the signal in a way that rejects what it considers spurious information and keep the good stuff. When fusing sensors, deciding that one sensor’s signal isn’t being helpful is where the data from a different sensor can come in to fill the gap – like a compass helping to correct a gyro. It’s just that here, there’s no plan B to fill in. So the useful signal has to be strong enough (or the algorithm’s ability to discriminate a weak signal has to be strong enough) to give high confidence.
They also talk about ways of minimizing noise based on sensor placement. In particular, focusing several sensors in the seat back lower down (to minimize noise from movement), putting the capacitive driven ground plane under the seat to hide it, and attaching the steering wheel to that ground. They’ve got an eval kit consisting of a six-sensor seat back and the ground plane for automotive folks to check out.
You can find more details in their release…