High-altitude video and barometer calibration testing

Krzysztof Bosak July 28, 2009 at 2:12am

The manual is being finished.
There are generic hardware specifications useful mostly for integrators.
The photos of the PCB are not recent for a good reason.
http://www.aerialrobotics.eu/flexipilot/flexipilot-brochure-en.pdf

BTW just for you:
http://www.aerialrobotics.eu/flexipilot/flexipilot-command-set-en.pdf

Installing FMA would theoretically shorten the time of recovery from DIVE, however will worsen everything else, namely the max achievable navigational dynamics and route following precision, decrease reliability (rain, fog), increase drag, increase control lag.
IMO FMA-based systems are good for recovery. however you are not supposed to do loops with camera onboard - because the wings became U-shaped. The autopilot know perfectly the roll even in a deep spiral - and it recovers from it. It also is preventing the spin as it commands the yawrate with 32Hz rate - but in easystar this is not usually a problem.] When the winds rise and the groundspeed is low, the IMU takes total control of the heading. This way the UAV survives to fight another... minute.
The autopilot uses tightly-coupled GPS and IMU and is good at reliable navigation. Adding random components around will worsen the reliability and performance for the task it has been tuned (aerial photography, heavily overloaded foamy glider-pusher). Note with different set of control laws you can achieve different autopilot applications. When I talk autopilot at the moment, I talk about the 'public' version for EasyUAV.

brakar July 28, 2009 at 1:34am

..., and for stabilization of a bigger airframe, would someting like the FMA co-pilot be a good idea?

brakar July 28, 2009 at 1:32am

Thank you for the update Krzysztof. Is there a datasheet or manual for the autopilot yet?

T3

Krzysztof Bosak July 28, 2009 at 12:39am

RVOSD is the official telemetry (bought separately). The whole autopilot is tuned, self-adjusting and uses relative waypoints. This way you need the laptop once per 1h of flight in order to view the logs. The project is intended to be minimal, thrusted solution for small scale aerial photographers (300g payload).
The autopilot supports the serial port input, online WPT change and collective feedback system reset, remote settings of all variables (require inflight reset) etc. HOWEVER you must connect directly at 921bps 8N1, order special version without Serial2USB converter onboard, and send the commands from your microcontroller which is the bridge between your modem/protocol and the autopilot.
Since the UAV should be officially flown at the line of sight and within RC range, there isn't much you could do with GS you couldnt do by yourself.
Second point, is that for telementry, it is not really useful as hard-realtime datalog guaranted to flush every 1/16s at 1/32s datarate max 9the reate is selectable). This is IMO superior to any telemetry since the data stored are at the order of 49kbps. This is overwhelming for most modems, at that datarate they choke later or sooner during heavy maneuvers.
Another story is why you want to put another strong TX and maybe its own supply in a limited airframe. I think the weight can be better used for putting a larger camera. As you see, I am very mission-oriented with this product, however the autopilot is general in its capabilities. I am willing to provide custom versions but preferably for research groups facing challenges, as they tend to look at the book at least once before contacting the vendor. There are some 'very secret' development projects based on this autopilot, but none available for public (price tag).

T3

Krzysztof Bosak July 28, 2009 at 12:36am

RVOSD is the official telemetry (bought separately). The whole autopilot is tuned, self-adjusting and uses relative waypoints. This way you need the laptop once per 1h of flight in order to view the logs. The project is intended to be minimal, thrusted solution for small scale aerial photographers (300g payload).
The autopilot supports the serial port input, online WPT change and collective feedback system reset, remote settings of all variables (require inflight reset) etc. HOWEVER you must connect directly at 921bps 8N1 and send the commands from your microcontroller which is the bridge between your modem/protocol and the autopilot.
Since the UAV should be officially flown at the line of sight and within RC range, there isn't much you could do with GS you couldnt do by yourself.
Second point, is that for telementry, it is not really useful as hard-realtime datalog guaranted to flush every 1/16s at 1/32s datarate max 9the reate is selectable). This is IMO superior to any telemetry since the data stored are at the order of 49kbps. This is overwhelming for most modems, at that datarate they choke later or sooner during heavy maneuvers.
Another story is why you want to put another strong TX and maybe its own supply in a limited airframe. I think the weight can be better used for putting a larger camera. As you see, I am very mission-oriented with this product, however the autopilot is general in its capabilities.

brakar July 28, 2009 at 12:09am

Interesting project Krzysztof, I have bookmarked www.aerialrobotics.eu and will be following the further development closely. Btw, will it be possible to add telmetry and GS to this autopilot?

T3

Krzysztof Bosak July 26, 2009 at 12:35pm

Thanks everybody for encouragement. You can notice the flight took place in the middle of nowhere, away from air trafic routes, and the traffic on local airfield is 2 per day. For anybody doing such flights, contacting local aeroclub and issuing NOTAM message is a must, but it can be done. Otherwise, without notifyilng the authorities, reasonnable rules are within LOS and max 200m AGL (ultralight rules). Hopefully, most photos are likely to be done from 200m AGL. This flight was exceptional.
This foamy setup weights 800g and loses the fight with almost anything from general aviation (it is not overloaded with metal objects, except the motor and the camera).