Adaptive Controller for HVAC in different Cars

Good evening,

I'm trying to program a cntroller for a cars HVAC system, which can adapt to defferent systems.

What do I have so far?

1) I have a GT-Suite Simulation of the cars HVAC system and Interieur. We know, that this simulation should fit approximately, but due to the fact, that the cars are not fully developed yet and the same controller will be built into different cars with different inner volume and interiour, I need a Controller, that can adapt to these conditions. The GT-Simulation can include Simulink models for Co-Simulations

2) I have a beaglebone black with a linux server distributuion which will be used as controller for the HVAC. We developed a board with all the electronics needed and the beaglebone will only be the "brain". It gets multiple Input-Data like temperatures, relative humidities, CO2 concentrations etc and has to set multiple Outputs like Compressor-Power, Fan Speed, Additional heating-power, air flap positions,... The Cntroller will at the end be implemented in C Code. I can include Python Code for e.g. NNs in my python Code.

My question now is, which way to go? I do have basic knowledge about linear controls, but almost nothing about nonlinear ones.

I found, that a fuzzy-controller or a neuronal network controller could fit best. What do you think: Which one can do the job best.

I would really like to have the same dataset for alle controllers in the different cars, which adapts itself to the conditions.

The controller only needs to be able tokeep the temperature and humidity on a specific value and the CO2 concentration below a specific value. No Multi-Zone or so.

Can you give me a first push, how to solve this problem?

Thank you!

Greetings

Simon