People counter steps

Hi, I am looking at making a pair of capacitive (compressible) sensors for a 2 step environment to track people movement. I tried the optical approach but mounting the IR is a problem, can't do opposing tx/rx pairs as the 'doorway' swing interferes with the optic mounting. Reflective IR is also compromised by having to 'tune' the range to within a couple of inches to 'ignore' the door, which can be reflective. Tough to do reliably.

It's a 'small' bus , 25 seater, application, to track # of people onboard.

I need some help with the 'logic' in counting persons directionally using the step switches. The challenge is that as a person vacates a step , the following person can place a foot on it so the logic has to do with transitions as opposed to states.

What I have so far (footsteps)

1) Person 1 enters bus climbing steps A and B, followed by person 2 doing the same:

A free B free, A(1) B free , A(1) B(1), A free B(1), A(2) B(1), A(2) B(free) , A(2) B(2) and so on.

A similar but converse arrangement for footsteps exiting the bus.

Any advice on picking a particular transition that would reliably identify # of persons entering or leaving?

thx!

Going to be difficult as people can step two different ways: 1) each foot hitting every other step (normal stepping) and 2) each foot hitting every step (old or disabled people).

Not really, as the logic is the same....you still have to have weight on the other step before you leave the current step.