I dropped the dual cell approach because it didn't really offer any benefit.
You don't have to read this. I warn you!
To my knowledge, dual cell setup can increase sensitivity. This Wheatstone bridge is a setup with two 1k ohm 'resistors', one been almost stable and the other subject to tension and compression (they are equal components but with different placements in the material itself). This is done to prevent temperature induced variations (temparature does change pressure and compression state of the materials) and the cell will read it!
Take the attached image. With
just one cell, it can take the place of R2-R4 or R1-R3. It's indifferent.
Suppose we choose R2-R4, and R2 is the variable element (R4 stays stable with pedal pressure).
Points A and B and at equal potential without pedal pressure. Sensor output will be zero.
As soon as you apply pressure, R2 changes making a difference between point A and B. That's what you measure. Sensor output will not be zero. The electronis are just an amplifier (in fact, a differencial amplifier).
Now, suppose
dual cell setup. This second cell will take the place of R1-R3.
In this case, you want to have R3 as variable resistor, not R4, or else theycancel each other.
This way, while R2 is, let's suppose, pulling point B up, R3 will be pulling point A down. The difference between A and B will be double, hence more sensitivity.
The advantage of this setup is you already have the four resistor to make the Wheatstone bridge (with one cell, you should have two normal resistors to complete the configuration).
Is all that necessary? I suppose not. You can increase sensitivity by just making the cell thinner (take of some material from the region below the white thing (on the down side). It will bend more => more output from the same pressure.