Co-variances were being re-zeroed after being set. This meant that the initial declination learning was sensitive to measurement errors which could result in poor initial yaw accuracy.
Fixes bugs that prevented planes being able to reset yaw to GPS to recovery from takeoff with a bad magnetoemter.
1) If the velocity innovation check had not failed by the time the in-air transition occurred, then the yaw reset would not be performed
2) The velocity states were not being reset
3) The non fly-forward vehicle (copter) reset could occur first and effectively lock out the fly-forward vehicle (plane) yaw check.
Remember the mag bias and earth field states learned during flight when the vehicle lands.
This improves performance for vehicles that do multiple flight on one power cycle
Provide consistent overshoot of 5% across a wider range of time constants and prevent selection of larger time constants causing 'ringing' in the position and velocity outputs.
Fixes a problem observed in a flight log where rapid temperature change caused the accel bias to change faster than the EKF could keep up.
This allows the bias to be learned faster but with acceptable level of noise in the estimate
IMU data was being corrected before being used by the co-variance prediction, whereas the delta angles and velocities in the derivation were supposed to be uncorrected.
This patch creates separate variable for the corrected data
Volatile will provide protection to sequence re-ordering and guarantee
the variable is fetched from memory, but it won't provide the memory
barrier needed to ensure that no re-ordering (by either the compiler or
the CPU) will happen among other threads of execution
accessing the same variables.
For more info about this effect can be found on articles about
std::memory_order.
When using reserved(), the reserved memory cannot be read before it's
written, therefore we cannot update 'tail' until the caller of
reserved() is done writing.
To solve that, a method called 'commit()' was added so the caller can
inform that is done with the memory usage and is safe to update 'tail'.
The caller also has to inform the length that was actually written.
This solution was developed to work considering the usage context of
this class: 1 reader and 1 writer **only**.
Adds a method called `reserve()`, that will take a ByteBuffer::IoVec
array of at least two elements, and return the number of elements
filled out. 0 will be returned if `len` is over the total space of
the buffer; 1 will be returned if there's enough contiguous bytes in
the buffer; 2 will be returned if there are two non-contiguous blocks
of memory.
This method is suitable to be used with POSIX system calls such as
readv(), and is an optimization to not require temporary memory copies
while reading from a file descriptor.
Also modify the write() method to use reserve(), so that similar checks
are performed only in one place.
Modify ByteBuffer class to have a `peekiovec()` method, that takes in a
`struct IoVec` array (similar to `struct iovec` from POSIX), and a
number of bytes, and returns the number of elements from this array
that have been filled out. It is either 0 (buffer is empty), 1
(there's enough contiguous bytes to read that amount) or 2 (ring buffer
is wrapping around).
This enables using scatter-gather I/O (i.e. writev()), removing calls
to memcpy(). That's one call when no wrap-around is happening, and
two calls if it is.
Also, rewrite `ByteBuffer::peekbytes()` to use `peekiovec()`, so that
some of the checks performed by the former are not replicated in the
latter.
- Correctly sort includes and add missing AP_Math.h
- Use anonymous struct for trim_registers in _load_trim_values,
renaming its members so they don't start with underscore
- Don't change _dig* values when we failed to read from sensor
- Add some blank lines
- Make _dig_* members be inside a _dig struct
- Use constrain_int32 instead of if/else chain
- s/time_us/time_usec/
- Construct raw_field with a single constructor in _update()
- Add missing copyright notice
- Group methods together in declaration
this gives more control over throttle for petrol
helis. H_RSC_POWER_NEGC allows for a asymmetric V-curve, which allows
for less power being put into the head when landing or when sitting on
the ground. That can lead to significantly less vibration and chance
of ground oscillation. A heli not being flown with aerobatics does not
need to use high throttle at negative collective pitch.
The H_RSC_SLEWRATE allows for a maximum throttle slew rate to be
set. Some petrol motors can cut if the throttle is moved too
quickly. We had this happen at a height of 6m when switching from
ALT_HOLD to STABILIZE mode. It also lowers the chance of the blades
skewing in their holders with the sudden change of power when the heli
is disarmed. In general it is a bad idea to do instantaneous large
movements of a IC engine throttle.
Magnetometer bias states will subject to larger errors early in flight before flight motion makes the offsets observable and the state variances reduce.
Adds a check on state variances.
Replaces the parameter check with a check of the actual filter fusion method being used.
Allow different process noise to be set for body (sensor bias) and earth field states.
This allows a stable magnetometer bias estimate to be available at end of flight whilst still allowing for external magnetic anomalies during landing.
Adjust default values to give stable mag bias learning and fast learning of external anomalies.
Automatically use the highest gain consistent with a 5% overshoot to minimise RMS tracking errors.
Provide an alternative correction method for the position and velocity states that allows the user to specify the time-constant. This can be used to fine tune the output observer for for platform specific sensor errors and control loop sensitivity estimation noise.
The toilet bowling check during early flight has been removed. This check caused problems where bad compass calibration was the cause of the toilet bowling and resetting to the compass was a bad option. The handling of simultaneous failed mag and velocity innovations is already handled outside the EKF by the failsafe.
A check for yaw errors due to a ground based magnetic anomaly has been introduced.
The logic for in-flight yaw and magnetic field resets has been cleaned up and variable names improved.
Many of our SPI and I2C sensors define the protocol of setting the most
significant bit of the register address in order to perform a read operation.
Thus, enable the use of a "read flag" that is ORed with the register's address.
Since this is an abstraction for general devices, it's a good idea to have zero
as the default value for that flag.
While at it, add documentation to read_registers().
../../libraries/AP_HAL_Linux/Perf.cpp: In member function ‘void Linux::Perf::_debug_counters()’:
../../libraries/AP_HAL_Linux/Perf.cpp:85:36: warning: format ‘%llu’ expects argument of type ‘long long unsigned int’, but argument 4 has type ‘uint64_t {aka long unsigned int}’ [-Wformat=]
c.name, c.count);
^
this reduces elevator control when rolled over hard in fixed
wing. Using the elevator when on the side just caused earth frame yaw
and is counter productive. It can also prevent some aircraft from
recovering from inverted flight.
That fixed compilation issues and seems more semantically correct. Using array
of length 0 fails compilation because of -Werror=array-bounds in GCC 6.1.
That fixed compilation issues and seems more semantically correct. Using array
of length 0 fails compilation because of -Werror=array-bounds in GCC 6.1.
Add DEFINE prefix, since this macro is defining these operators and
remove the parameter since we will always use it to access a
union/struct as a byte array.
Test code for integration with another thread to pull data from internal
perf counters. Since we are using the timer thread here, there's no
retry mechanism and we only print that data can be corrupted.
Instead of creating a new object Perf_Lttng copying the necessaries
fields, just make a tighter integration with the internal perf counters
and re-use the same fields.
The idea is to leave the internal perf enabled all the time, like it is
in PX4, and then allow the integration with lttng on top. Next step
would be to runtime enable/disable only the perf counters we are
interested in.
This also changes the structure so it's easy to allow another thread to
pull data from the Perf object. A rw lock protects from addition of new
counters and an atomic unsigned int allows other threads to do a
lockless copy of the data.
In order for this to work the allocation was changed to use a single
memory pool instead of returning a calloc'ed data for each perf counter.
Since most of our counters are of ' elapsed' type, don't bother using a
smaller struct for the 'count' type