Does any open source autopilot (e.g. Pixhawk) use GPS/INS (using GPS signals to correct or calibrate a solution from INS)? For what kind of flight mission and what case will the usage of GPS/INS become important and show significant advantages over the other systems?
You need to be a member of diydrones to add comments!
Replies
Hi Anna,
Could you perhaps be more specific in your question? I think you are referring to what may be known as 'robust NAV', ie, a system implementation and hybridization of the various available sensors to provide continuous albeit degraded modes of navigation and attitude control as various sensors start to fail.
But your question is to broad and open to know....If not robust nav, then perhaps your question is simpler - in which case, ALL 'low cost' sensors based IMU's used in an INS implementation will require additional sources of correcton data to correct for the sensor drifts, especially in highly dynamic platforms where accelerator outputs are highly dynamic.
A high accuracy INS may have sensor drifts of the order of tenths of a degree per hour, or better, while the typical MEMS Gyro rate sensors in the 'DIY' IMU may have drifts around hundreds of degrees per hour. the combination of sensors, hybridized by means of a Kalman filter or other method, then compensates for this, to produce higher accuracy.
Maybe you can ask your question again?
Regards
Joe
Sorry. I'm a newbie to this area. My question is not clear enough. I'm asking about GPS aided INS which is explained by the link below.
http://en.wikipedia.org/wiki/GPS/INS
How many degrees will 3DR's Pixhawk's Gyro rate sensors drift per hour? Does that mean if I let Pixhawk fly fixed wing UAV for more than one hour, its Gyro won't be accurate at the end of the flight so it is for the UAV to crash?
Hi Anna,
What you are describing is something that is common in commercial aviation. I don't know the exact answer to your question about drift, but in general I would guess that the position gyro's of the pixhawk (if it has position INS) would not be comparable to a full up nav solution used in commercial aviation. Even ring laser gyro's used on airliners would not likely have the total precision to be that accurate (drones require extremely high accuracy to RTL for instance) without GPS updating into a navigation solution that has a mathematical degree of confidence. In this case, the confidence would be very low. Civil aviation uses a system called RNP - or required navigation performance, that determines position with 95% confidence all of the time. The latest update to the pixkawk has algorithms for EKF - Kalman filtering which enhances GPS accuracy considerably.
The pixhawk does have INS, gyros and a compass that is quite accurate. The INS is mostly used for determining the rate of acceleration on the the three axis of the vehicle which assists in control accuracy, not navigation accuracy. GPS doesn't provide control accuracy, only navigation.
PS. This is all my speculation from working with the pixhawk. I'm not a developer and have made some assumptions based on my experience.
Best,