0

I am trying to generate C code for controller and plant models which are continuous in time, so the code shall be generated with floating point support. But the my real-time target (RTlinux kernel x86) does not support floating point processes. So, I have to come up with a solution of generating code with fixed point processes.

In first place, is it possible to do so? if yes, then could anybody post some possible suggestions.

--

regards

Eshwar Khatri
  • 1

1 Answers1

1

You can use the "Fixed-Point Designer" available as a Simulink toolbox: "Fixed-Point Designer lets you convert floating-point algorithms to fixed point." If the toolbox is not available, I still recommend you to read its documentation since it contains useful hints and issues to be considered in a "manual" floating-point-to fixed point conversion process.

Captain Future
  • 329
  • 1
  • 5
  • thanks. but I have already tried it with "Fixed-Point Designer" and the problem is: it doesn't support the fixed point advisor for _**library(system/subsystems)**_. In my case I am using continuous PID and Plant blocks. I have other option of using hardware Floating Point Unit (FPU) in Linux kernel, but this the last option I would opt if I don't get any other solution. I am using x86 pc and linux-3.18.22. – Eshwar Khatri Aug 10 '16 at 15:06