Need to handle irregular geometric inputs

I’ve been working on a project using STM32 where I need to handle irregular geometric inputs (user-defined coordinates) and compute results efficiently on the device side.

While implementing this, I started exploring different approaches for calculating areas of irregular shapes, especially when dealing with limited resources and precision constraints typical in embedded systems. The Shoelace formula seems like a good fit, but I’m curious how others here handle optimization for such calculations on STM32 (e.g., fixed-point vs floating-point, performance trade-offs, etc.).

I also experimented with similar logic in a browser-based tool, like this irregular shapes calculator, mainly to validate the math and input handling before porting it to embedded code.

My main questions:

• What’s the most efficient way to implement polygon area calculations on STM32?

• Do you recommend sticking with float/double, or using fixed-point arithmetic for better performance?

• Any best practices for handling dynamic input sets (variable number of vertices)?

Would really appreciate insights or examples from anyone who’s worked on similar embedded geometry or computational tasks.

Thanks!