Commit Graph

44 Commits

Author SHA1 Message Date
Lucas De Marchi 82d210144b AP_Math: remove warnings from constrain_value()
Return type is T which can be an integral type, float or double. By
dividing by 2 we avoid float operation on the first case and do the
right thing on the second and third.
2017-01-27 12:15:19 -08:00
Lucas De Marchi 2605c7265b AP_Math: remove warnings from safe_asin()
Return type is float, so operate on float types everywhere.
Fixes this warning while building for PX4:

../../libraries/AP_Math/AP_Math.cpp: In instantiation of 'float safe_asin(T) [with T = double]':
../../libraries/AP_Math/AP_Math.cpp:56:48:   required from here
../../libraries/AP_Math/AP_Math.cpp:44:11: warning: implicit conversion from 'float' to 'double' to match other operand of binary expression [-Wdouble-promotion]
     if (v >= 1.0f) {
           ^
../../libraries/AP_Math/AP_Math.cpp:47:11: warning: implicit conversion from 'float' to 'double' to match other operand of binary expression [-Wdouble-promotion]
     if (v <= -1.0f) {
           ^
2017-01-27 12:15:19 -08:00
Lucas De Marchi 4f8d2059f8 AP_Math: use right epsilon for is_equal()
We are calling fabsf(), which returns a float. We should use the epsilon
from float type, not from the argument type passed to fabsf().

On the other hand when the double version is instantiated we do want to
use the std::numeric_limits<double>::epsilon() value.

This adds a branch to the function, but it's removed when the function
is intantiated by the compiler since the type is known at compile-time.

Fixes this warning when building for PX4:
../../libraries/AP_Math/AP_Math.cpp: In instantiation of 'typename std::enable_if<std::is_floating_point<typename std::common_type<_Tp, _Up>::type>::value, bool>::type is_equal(Arithmetic1, Arithmetic2) [with Arithmetic1 = double; Arithmetic2 = double; typename std::enable_if<std::is_floating_point<typename std::common_type<_Tp, _Up>::type>::value, bool>::type = bool]':
../../libraries/AP_Math/AP_Math.cpp:23:66:   required from here
../../libraries/AP_Math/AP_Math.cpp:17:29: warning: implicit conversion from 'float' to 'double' to match other operand of binary expression [-Wdouble-promotion]
     return fabsf(v_1 - v_2) < std::numeric_limits<decltype(v_1 - v_2)>::epsilon();
                             ^
2017-01-27 12:15:19 -08:00
murata e903cb9945 AP_Math: Change mask value to hexadecimal number. 2017-01-17 10:20:15 -08:00
Pierre Kancir da49149d19 AP_Math: is_equal correct comparison for integer as epsilon doesn't exist.
Credit to Kwikius for the right solution
2016-12-20 14:09:04 +00:00
Andrew Tridgell a8d10e8c2c AP_Math: added get_random16() 2016-12-02 09:49:38 +11:00
Ricardo de Almeida Gonzaga 481e3a2af6 AP_Math: Fix typos 2016-05-13 19:20:06 -03:00
dgrat 6d3b491c02 AP_Math: Replace is_equal with a type safe template function
It makes sense to consider also other floating point types.
2016-05-10 11:41:26 -03:00
dgrat 503867b7dc AP_Math: Replace safe_sqrt() by template function 2016-05-10 11:41:26 -03:00
dgrat 5deb0e8e03 AP_Math: Replace safe_asin() by template function 2016-05-10 11:41:26 -03:00
Lucas De Marchi 846b4927ec AP_Math: use if/else chain instead of 2 ternary operators 2016-05-10 11:41:26 -03:00
dgrat 174f899a29 AP_Math: Replace the constrain_* functions by a single template
Besides being simpler this reduces ~4k in the binary size for PX4.
2016-05-10 11:41:26 -03:00
Lucas De Marchi 880f130670 AP_Math: fix loss of precision on float addition
When using wrap_180_cd() we are adding a small float (180 * 100) to a
possibly big number. This may lose float precision as illustrated by the
unit test failing:

    OUT: ../../libraries/AP_Math/tests/test_math.cpp:195: Failure
    OUT: Value of: wrap_180_cd(-3600000000.f)
    OUT:   Actual: -80
    OUT: Expected: 0.f
    OUT: Which is: 0
2016-05-10 11:41:26 -03:00
dgrat 76362caee0 AP_Math: Replace wrap_* functions with template versions 2016-05-10 11:41:26 -03:00
Andrew Tridgell 6bff07397e AP_Math: added linear_interpolate() function 2016-04-02 22:44:47 +11:00
dgrat 5148e41c1a AP_Math: Cleaned macro definitions
Moved Definitions into a separate header. Replaced PI with M_PI and
removed the M_PI_*_F macros.
2016-02-27 02:51:33 -03:00
dgrat 7c4c8ea579 AP_Math: Remove ROTATION_COMBINATION_SUPPORT
This function is not used.
2016-02-27 02:51:33 -03:00
Valmantas Palikša d4daf19151 AP_Math: Move simple math function implementations to header for better
compile time optimization

Functions like sq() are better moved to the header file as inline.
Compiler can then optimize these out when used in code, this saves cpu
cycles with stack push, pop during function calls.
2015-09-09 09:57:51 +10:00
Randy Mackay 46c652e42f Math: maxf and minf functions 2015-08-19 16:44:34 +09:00
Andrew Tridgell 77a2b4acf6 AP_Math: removed fast_atan 2015-05-05 13:57:22 +10:00
Tom Pittenger 4ec2fb3a9c AP_Math: Compiler warnings: nuke fast_atan2()
per Randy's suggestion, fast_atan2() is no longer necessary over atan2() because only copter uses it and copter is no longer supported on future builds of APM

ccd578664f (commitcomment-11025083)
2015-05-05 13:27:02 +10:00
Tom Pittenger 6e6f481ecb AP_Math: compiler warnings: apply is_zero(float) or is_equal(float) 2015-05-05 13:26:56 +10:00
Andrew Tridgell eca675c556 AP_Math: fix for HAL_SITL rename 2015-05-05 09:45:55 +10:00
Randy Mackay 0392292489 AP_Math: inline is_equal, add is_zero 2015-04-28 16:19:01 +09:00
dgrat 726d7df710 AP_Math: add is_equal to compare floats 2015-04-28 16:18:59 +09:00
Tom Pittenger a8dda9f2ed AP_Math: fix compile warnings re float constants 2015-04-24 14:03:54 +09:00
Jonathan Challinger 5f7480b740 AP_Math: change fast_atan2 to use atan2f on fast CPUs 2015-02-09 22:24:09 +09:00
David Dewey 17374ff5e8 AP_Math: fast_atan2
This is 126us per call vs 199us on the AVR.  it is accurate to about
0.28 degrees

Committed by rmackay9 but contribution is from David Dewey
2014-06-06 18:50:41 +09:00
Randy Mackay d81b7b507d Math: add Leonard's fast tan function 2013-05-30 18:24:32 +09:00
Andrew Tridgell ba83950fc4 libraries: replace constrain() with constrain_float()
this makes the type much more obvious. Thanks to Tobias for the
suggestion.
2013-05-02 10:25:40 +10:00
Andrew Tridgell 83dc7dbc92 AP_Math: handle NaN in constrain(), returning average
this makes it less likely a NaN will propogate
2013-04-15 14:28:51 +10:00
James Bielman 5631f865b2 Update floating point calculations to use floats instead of doubles.
- Allows use of hardware floating point on the Cortex-M4.
- Added "f" suffix to floating point literals.
- Call floating point versions of stdlib math functions.
2013-01-16 13:52:01 +11:00
Andrew Tridgell 38062bbb6d AP_Math: support some more rotation combinations 2013-01-13 17:32:48 +11:00
Andrew Tridgell a072afa223 AP_Math: expand some macros into functions
this saves some flash
2012-12-20 14:52:38 +11:00
uncrustify 533860dd73 uncrustify libraries/AP_Math/AP_Math.cpp 2012-08-21 19:03:34 -07:00
Andrew Tridgell 58a68c89d9 Math: added a function to combine standard rotations
this will allow us to have an overall board rotation plus a per-sensor
rotation
2012-03-19 17:29:02 +11:00
Andrew Tridgell 4422486d44 AP_Math: made rotation matrices more C++
thanks to Adam for the suggestion
2012-03-11 15:37:07 +11:00
Andrew Tridgell 9f0cb78f08 AP_Math: re-work quaternion functions to be more C++ like
thanks to Adam for the suggestion!
2012-03-11 15:37:07 +11:00
Andrew Tridgell 4d65cda0ea AP_Math: added quaternion->matrix and earth frame routines
this will be used for 3d acceleration
2012-03-10 10:34:31 +11:00
Andrew Tridgell 8abbbe5713 AP_Math: better way of handling safe_sqrt()
better to test the result, than predict it
2012-03-10 10:34:31 +11:00
Andrew Tridgell c7d5f06b21 AP_Math: added quaternion helper functions and a test suite 2012-03-10 10:34:30 +11:00
Andrew Tridgell 16ed8bc8f6 AP_Math: added rotmat <-> euler functions
these will make the dcm matrix manipulation easier to understand
2012-02-24 11:52:55 +11:00
Andrew Tridgell 17d16bc8c6 AP_Math: added safe_sqrt() function
this function will never return NAN. It will return zero for negative
numbers.
2012-02-24 11:52:55 +11:00
Andrew Tridgell 502fbf6e17 AP_Math: added a safe_asin() function
this adds range checking to asin()
2012-02-24 11:52:55 +11:00