This is the second post I have made that is an update to my initial project design to make my fully autonomous aircraft. Our team has made a fair amount of decision making sense my previous post. Just for reference, the official name of our project is now "FlySpy".Airplane Selection:We have chosen the Easy Glider Pro from multiplex. Although it does not have as much internal space as the Easy Star, we fell that ailerons were a must for our application which Easy Star doesn't have. We have completely assembled the airplane and have flown it once. (No one on our team has experience flying RC Airplanes so we have an official pilot from a local UAV Imaging company. It would be nice for someone on our team to have experience flying but we do not want to use our project as a test dummy.) We are using the Power Pack and Servo pack that are made for the Easy Glider Pro. The motor seems beefy enough to support a flight with our components added to its weight so we do not intend on buying a bigger one.Battery : EVOLITE Li-Po 7.4V 2500mAh BatteryAutopilot Module Component Selection:Microcontroller - Microchip PIC24FJ256GA1103 Axis Accelerometer - Analog Devices ADCL3302 Axis Gyro - InvenSense IDG 300Barometer - VTI Technologies SCP 1000Range Finder - LV-MaxSonar-EZ1 Sonar Range FinderGPS Module - uBlox LEA-4P and getting Sarantel SL 1201 (Geohelix P2)SD Card Slot - TBDI have obtained a Niko Cool Pix s7c from a friend to use as an onboard camera. I am still working on decoding its push button pins so our Micro can turn it on/off and control the shutter.Topics we are contemplating:Powering Autopilot Devices - Everything that we are using runs off a 3.3V Vdd. We are uncertain if it would be okay to use the power rail coming out of the speed controller and sending it to a voltage regulator. We estimate our devices drawing about 157mA when everything is operating. Is this the usual solution or do we need to run off of two battery sources?Gyro & Accelerometer Data - I am still in the process of researching what are the best algorithms to fuse the data between the gyro and accelerometer to get a good orientation reading.If you are interested in looking at our progress, you can always look at our team website. Or if you want to go more in detail about my day to day discovery, you can look at my notebook.
Heey guys I have a question since your case looks related to mine. My ultrasonic currently works fine with a 5 V supply (hard wired to USB), however, in a 3.3V Lipo battery powered case, the ultrasonic conversion leads to significant errors. Currently I accept Analog input from the ultrasonic to the ardupilot. Is there a correct conversion arduino code for the ultrasonic (Maxbotix LV-EZ1) sensor when powered by a 3.3 V Lipo battery? I do not need the program for RX/TX, or PWM. I need it for the analog input, not for anything else.
If you want to have the ailerons moving together to increase drag for landing it would be usual to have them going up rather than down - "spoilerons". This reduces the risk of tip-stalling on a slow approach (ie one wingtip stalling before the rest of the wing, which can cause a plane to flick sideways into the ground).
@Jeremy: I think your intention of flapping ailerons down on your easyglider will be fine as long as you do not push them down to the servo arm maximum deflection. The more you'll push them down, the less aileron control you'll have left! I experienced aileron as flaps on my easyglider pro. At one point I had them too low and had not enough aileron control to avoid a tree...
You'll also have to mix in some down elevator to compensate for the bigger lift. But with the flaps, you may do a slow 25-30 degrees approach.
I also have an EZ1 range finder with which I intend to auto land my Kadet Senior UAV. My landing approach will be split in two parts: GPS altitude control for lining in approach and then range finder altitude control for final. For sure I'll log data from ground range readings and GPS in manual to tweak my code before attempting auto landing.
if you keep the nose up at 6m to touchdown on a easyglider you should have a nice landing.
there should be some groundeffect to, that reduces vertical speed a bit.
currently i land my "uav" from a height of about 20meters without any control input. (remote is sitting on the ground)
The fma-copilot keeps the plane level or slightly nose up. the landings are much better than most of my landings with my other aircraft.
just give it a try....
I would expect that trying to use the ailerons as flaps would give all sorts of problems. When they are deployed as flaps the drag will increase and a stall could quickly result unless you gave some throttle to compensate for the drag increase and then a touch of elevator to trim the attitude.
I am interested in your choice of rangefinder sensor. Have you tried it on a model and over differing types of terrain (ie long grass, short grass, sand etc.)?
Using a pressure sensor for height will be fine as long as the pressure does not change at ground level. The performance of pressure sensors are ok but I would not expect to get much better than about 5ft accuracy over a period of 20minutes for all sorts of reasons. Even a gust of wind could affect the pressure. Also to get the high reolution the measurement time may possibly have to be quite long.
After I while I wanted to smack some wheels to the bottom of this thing and have a smooth airplane type landing. So I would give it the starting and ending coordinates of a landing strip and it would descend to a altitude of about 6 meters above the ground and use the ailerons as flaps thereafter. Has anyone tried to do this?
Don't I need to know my height above the ground to start landing if I want to land at an exact position. I don't think I should just turn the engine off without knowing how long it is before I hit the ground.
Cool. You should get really nice readings from the rangefinder whether you use the analog, PWM, or serial output. I just finished using an EZ0 on a robot and it worked remarkably well. You might want to double check the range of the EZ1 vs. the EZ0, since I think the EZ1 might have a range of more like 15 feet. There is also the question of how much control authority you have over your altitude and whether a closed loop system will be effective. But then again, that's half the fun of experimenting with the UAV! Before closing the loop, I would be sure to do a bunch of manual landings with the rangefinder readings and the control surface deflections begin logged at a high rate to a data file. The ability to develop a useful control law will be greatly simplified if you have this data at your disposal. Good luck!
Comments
Heey guys I have a question since your case looks related to mine. My ultrasonic currently works fine with a 5 V supply (hard wired to USB), however, in a 3.3V Lipo battery powered case, the ultrasonic conversion leads to significant errors. Currently I accept Analog input from the ultrasonic to the ardupilot. Is there a correct conversion arduino code for the ultrasonic (Maxbotix LV-EZ1) sensor when powered by a 3.3 V Lipo battery? I do not need the program for RX/TX, or PWM. I need it for the analog input, not for anything else.
You'll also have to mix in some down elevator to compensate for the bigger lift. But with the flaps, you may do a slow 25-30 degrees approach.
I also have an EZ1 range finder with which I intend to auto land my Kadet Senior UAV. My landing approach will be split in two parts: GPS altitude control for lining in approach and then range finder altitude control for final. For sure I'll log data from ground range readings and GPS in manual to tweak my code before attempting auto landing.
there should be some groundeffect to, that reduces vertical speed a bit.
currently i land my "uav" from a height of about 20meters without any control input. (remote is sitting on the ground)
The fma-copilot keeps the plane level or slightly nose up. the landings are much better than most of my landings with my other aircraft.
just give it a try....
I am interested in your choice of rangefinder sensor. Have you tried it on a model and over differing types of terrain (ie long grass, short grass, sand etc.)?
Using a pressure sensor for height will be fine as long as the pressure does not change at ground level. The performance of pressure sensors are ok but I would not expect to get much better than about 5ft accuracy over a period of 20minutes for all sorts of reasons. Even a gust of wind could affect the pressure. Also to get the high reolution the measurement time may possibly have to be quite long.
Good luck and keep us posted on progress.
regards Peter
just keep the plane level or sligtly nose-up, kill the engine and wait...
-
1
-
2
of 2 Next