Interpolation Curves/Surfaces in FRC

[apalrd]

So, I read another thread discussing methods of finding an equation for a given set of points, I wondered how or if other teams use interpolation tables as a quick and easy way of scaling data and getting a desired transfer function.

I attached two examples of 3d surfaces (both are from Kettering's FSAE engine software) and a 2d curve (from a robot).

In either case, the input breakpoints and output values at each cell are calibrated in a table. You can change things so the response is nonlinear, or if you have exact values at certain points you can type them in. The software then finds the correct cells bounding the input variables, and interpolates linearly to find the output.

I use them all the time, both at work and in FRC. Does anyone else use them. What are your thoughts?

[magnets]

When we need to find a function we just use excel and we get a polynomial line of best fit, then put the equation on the robot.

It's nice because you can play around with the inputs/outputs and watch how the graph changes.

We've used it for making functions for distance from goal vs. shooter rpm, distance from wall vs. desired robot speed, motion profiling for elevators/arms/climbers where we'd like to go different speeds at different locations, robot speed vs. intake roller speed (this is really nice for fast robots with roller intakes), and a bunch of other things. It's really useful.