Design Patterns for Processing Time-Series Data that changes format and algorithm regularly

I'm trying to design a system for processing time series data from a variety of aircraft sensors. The main issue with the design process is that sensor data formats and calibration values change with time, as does the hardware configuration, meaning that processing algorithms and pre-processing steps need to be changed accordingly.

My current plan is to write software and use some form of a version control system to link files defining the sensor data formats and various system parameters for the sensors to a specific version of code, all based on the date they were written. When the user runs a makefile or some form of a script, the code I ideally want to write takes care of this automatically. My overall plan is to then launch each of these sub-programs in parallel such that data form any time period is processed without issue, provided that the code has been properly validated.

However, I'm not a very experienced programmer and am not knowledgeable in what software design patterns or VC methodologies are out there for these types of problems. Would there be any particular methods that could implement this plan efficiently, or some web/book reference which could help me out?