[stabilization] helicopter indi controller #1752
Conversation
| _out[i] = _in1[i] - _in2[i]; | ||
| } | ||
| } | ||
|
|
There was a problem hiding this comment.
There seems to be some functions that have this functionally, check sw/airborne/math/pprz_simple_matrix.h.
If you have time, you could add a math file which has the wrappers for the matrix computations for vectors. I think these functions would be interesting for others not just for indi, you would just make the INDI_DOF the input t the general function.
If you don't have time (or the interest) maybe make an issue for to do this.
There was a problem hiding this comment.
Ok I can change the matrix thing to MAT_MUL_T
What would be the best location to add the vector functions (with length as variable), pprz_algebra.h?
There was a problem hiding this comment.
I think in the pprz_simple_matrix.h. Or maybe in a pprz_vector.h. @flixr?
There was a problem hiding this comment.
There are n-dim float_vect_x functions in pprz_algebra_float.h, so I guess a fixedpoint equivalent should go into pprz_algebra_int.h
|
I did not use MAT_MUL because it assigns a float zero. I'm not 100% sure about the double pointer stuff here. |
| static inline void indi_apply_measurement_butterworth_filters(int32_t _out[], int32_t _in[]) | ||
| { | ||
| for (uint8_t i = 0; i < INDI_DOF; i++) { | ||
| _out[i] = update_butterworth_2_low_pass_int(&measurement_lowpass_filters[i], _in[i]); |
There was a problem hiding this comment.
Why have twice the same function with a different filter struct? can that not be an argument?
There was a problem hiding this comment.
Good point, will change that. Also applies for the notch filters.
There was a problem hiding this comment.
I removed duplicate code for notch filters. Its not really trivial to make the filter struct an argument without reducing the flexibility. Like it is now, its easy to replace one filter with another.
|
Ok this is weird, from the failed build log: |
…heli_indi_pr Conflicts: conf/airframes/TUDELFT/tudelft_walkera_genius_v2.xml
|
That happens sometimes in travis, no idea why... maybe a caching problem on their build farm... |
|
I think this is now mergable |
| //_out[INDI_THRUST] = _in[INDI_THRUST]; | ||
| } | ||
|
|
||
| static inline void indi_apply_notch_filters(struct SecondOrderNotchFilter *filter, int32_t _out[], int32_t _in[]) |
There was a problem hiding this comment.
all these apply functions should not be inline, since you want to use pointers to these functions later...
|
Thanks! |
This is an INDI controller for a small helicopter. It was flight-tested on a walkera genius cp v2. When it is merged, I can add documentation to the wiki on how to use it. It is quite different from the other INDI controller, so therefore I suggest to not combine it with the other one.