I'm going to expand on
@jack action's lead and go into some detail on the nuts and bolts of rolling resistance and how exactly a piezo cell could be made to impact it.
Rolling resistance has 3 somewhat inter-connected components:
1. Inelastic deformation of the tire.
2. Kinetic friction due to slippage due to deformation.
3. Inelastic deformation of the ground.
Obviously, #3 is the one we are concerned with. In general/for a quality road, it is a couple of orders of magnitude smaller than the other two. But it works identically to #1. Here's how it works:
As described, an elastic object is like a bunch of little, connected springs. Apply a force and the springs compress. Release the force and the springs expand back to their original length. For the most part, energy is conserved in this process and the spring gives back almost all of the energy it absorbed. Applied to a car, the tire and road are literally shoving each other out of the way,
but then they both rebound and push each other back, recovering the energy it took to shove each other out of the way. In other words, we said the car is basically climbing a hill all the time. Well, that's what the front of the tire is doing. The back of the tire is traveling
down that hill (er - the previous hill).
But springs, roads and tires are not perfectly elastic. There is a component of inelasticity, which means that when they spring back, they do so imperfectly: with less force than it took to compress them. That means the hill they are rolling down isn't as steep as the hill they are driving up. The rolling resistance is the
difference between the compression force and restoration force (adjusted for the angle of the contact to get it in the direction of motion of the car).
A simple piezoelectric device you can buy and play with for a couple of bucks is just a strip of plastic, to be used as a vibration sensor. Flick it and it vibrates - and it has a certain inelasticity that dissipates the energy and causes it to eventually stop vibrating. But otherwise, like the road rebounding, this is a pretty elastic situation.
Now, when you hook the device up to a circuit, you change its mechanical properties. How exactly depends on the circuit, but I'll give a simple example that dovetails well with what we are discussing. Let's say that for the deflection part of the cycle you leave the circuit open. At max deflection, you have a certain voltage and charge available and at that moment you shunt it to something that can harness it. What you just did is to siphon-off some of the elasticity of the device, making it rebound with less force and lower amplitude.
Applied to a road, the connection should be obvious: When you implant the device in the road, you can say the elasticity is the same as the concrete you replaced, if you choose it that way. But when you turn it on to extract the energy, the impact is to not rebound as well, opening a new gap between the compression and the expansion, creating new rolling resistance.