Minimal sensing: the target counting problem

Prof Simon Dobson –  26 June, 2017 


A lot of effort in sensor networks goes into building more capable
sensors, so we can collect more data more easily to support more
analysis But it’s interesting to ask what we can find out using only
minimal sensing. A motivating example for this kind of work is the
target counting problem: given a set of sensors that can count (but not
identify) targets in their vicinity, how many targets are there? What
makes this difficult, of course, is the possibility that many different
sensors are seeing the same target, and in this talk we’ll discuss a
couple of methods for addressing this (including one we’re working with
quite heavily in St Andrews) as a way of exploring the uses of even
minimal sensors.