return data type in get_delay()

This is a spin-off from #59 (closed).

The function get_delay() returning parameter represents a delay in nanoseconds while it is defined as a float data type. This return parameter is then rounded to an integer and converted to int32_t data type.

In practice, it does not happen in the current framework that a delay in nanoseconds is not an integer value making the current process complicated and confusing.

I thus propose to change the return data type of function get_delay() from float to int32_t and simplify its calling.