Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Artificial Horizon - Android or iPhone Bubble/Level App?
496 views
Skip to first unread message
V1
Jul 15, 2015, 10:40:25 PM7/15/15
to
There are lots of bubble/level apps available for both Android and iPhone. I played with one on the ground (it combines a level and compass) including doing simulated steady state turns; it seems accurate and repeatable. I wondered if it could function as an artificial horizon in a pinch, like descent through clouds? I haven't tried it in flight (where it may react differently than on the ground), since it would be a distraction to do significant testing in anything other than a two-place.
Has anyone actually tried to use one in flight, without reference to outside horizon?
- Frank
Has anyone actually tried to use one in flight, without reference to outside horizon?
- Frank
threeu...@gmail.com
Jul 15, 2015, 11:42:56 PM7/15/15
to
I have been playing with some.
Including building one myself with a micro display (and flying it).
It depends on the implemented algorithms.
The ones I played with work well when stationary, but do not work under acceleration.
A coordinated turn is an accelerated movement.
Or in other words: In a coordinated turn, the angle of the acceleration vector wrt the body (i.e. airframe) does not change. Hence no roll indication.
You can easily test if the algorithm takes acceleration as the main input.
Take your phone and accelerate (move) it quickly linearly along all its axis, without turning (no gyro input)
If you move the phone forward and backwards (without pitching) and the display shows pitch, the algo uses the accelerometer over the gyro and it will not work in the air.
If you move the phone left and right (without rolling) and the display shows a bank, the algo uses the accelerometer over the gyro and it will not work in the air.
If you want to read more about the AHRS algorithms, there is a lot of info on the web in the RC drones area.
Just in a few words: The dilemma is that the accelerometers are noisy and the gyros have an offset. Integrating / accumulating the gyro offset over time creates drift. To avoid that drift, some algorithms emphasize the accelerometer over the gyro.
3U
Including building one myself with a micro display (and flying it).
It depends on the implemented algorithms.
The ones I played with work well when stationary, but do not work under acceleration.
A coordinated turn is an accelerated movement.
Or in other words: In a coordinated turn, the angle of the acceleration vector wrt the body (i.e. airframe) does not change. Hence no roll indication.
You can easily test if the algorithm takes acceleration as the main input.
Take your phone and accelerate (move) it quickly linearly along all its axis, without turning (no gyro input)
If you move the phone forward and backwards (without pitching) and the display shows pitch, the algo uses the accelerometer over the gyro and it will not work in the air.
If you move the phone left and right (without rolling) and the display shows a bank, the algo uses the accelerometer over the gyro and it will not work in the air.
If you want to read more about the AHRS algorithms, there is a lot of info on the web in the RC drones area.
Just in a few words: The dilemma is that the accelerometers are noisy and the gyros have an offset. Integrating / accumulating the gyro offset over time creates drift. To avoid that drift, some algorithms emphasize the accelerometer over the gyro.
3U
Darryl Ramm
Jul 15, 2015, 11:48:46 PM7/15/15
to
Artificial Horizons rely on gyroscopes not accelerometers. These inclinometer apps just use accelerometers to compare the phone inclination to the local gravitational field direction. Moderns smart phones do have gyroscope sensors as well as accelerometers and can drive "artificial horizon" toy apps. If survivign flight in IMC is of any real interest/concern you are not going to be dicking around with a smart phone.
Dan Marotta
Jul 16, 2015, 11:18:59 AM7/16/15
to
Without proper training and experience, you will survive inadvertent
cloud flying only by sheer dumb luck. Is it worth risking your life
on that basis?
Here are two little tidbits of information that could help you make up your mind:
Illusions
and
178 Seconds to Live
Take the time to look and consider...
Here are two little tidbits of information that could help you make up your mind:
Illusions
and
178 Seconds to Live
Take the time to look and consider...
--
Dan Marotta
Dan Marotta
lyle
Jul 16, 2015, 12:22:58 PM7/16/15
to
An AHRS (Altitude Heading Reference System) must use at minimum three sensors: Accelerometers, Gyros, and Magnetometers (or GPS).
Algorithms that use just the accelerometers to correct for the gyro bias will diverge (get a VERY wrong answer) over time if you are in a coordinated turn. (The filters in question try to observe the long term trend of which way is "down." Circling in a thermal will fool these fairly quickly.)
The magnetometer or GPS input must be used to observe the turn and prevent incorrect estimation of the gravity vector.
A fairly well-known university (with which I am not affiliated) discovered this about a decade back when they were commanding their fixed-wing UAVs to loiter by circling. After a couple minutes they would overbank, enter a spiral dive, and crash rather dramatically...
Algorithms that use just the accelerometers to correct for the gyro bias will diverge (get a VERY wrong answer) over time if you are in a coordinated turn. (The filters in question try to observe the long term trend of which way is "down." Circling in a thermal will fool these fairly quickly.)
The magnetometer or GPS input must be used to observe the turn and prevent incorrect estimation of the gravity vector.
A fairly well-known university (with which I am not affiliated) discovered this about a decade back when they were commanding their fixed-wing UAVs to loiter by circling. After a couple minutes they would overbank, enter a spiral dive, and crash rather dramatically...
0 new messages