ArduCopter-3.0 ready for wider use

ArduCopter 3.0 is ready for widespread use. To make the transition easier, this time we are asking people to voluntarily upgrade from 2.9.1b for the next month or so before we make 3.0 the default firmware downloaded by the mission planner. The new version can be found in the Mission Planner's Beta firmware's link, firmware.diydrones.com, GitHub and the new Downloads Area.

Warning #1: A bug was found in the FENCE in which if you lost GPS it could lean at extreme angles as it tried to maintain an invalid position. This is fixed in AC3.0.1-rc1.

Warning #2: This release has not been fully tested on Traditional Helicopters

Warning #3: GPS glitches can cause sudden and aggressive position changes while in loiter mode. You may wish to reduce the WPNAV_ACCEL to 100 and Loiter PID to 0.2 (from 1.0) to reduce aggressiveness.

Improvements over 2.9.1b include:

Inertial Navigation for Loiter and Auto meaning much more accurate control (Randy,Leonard,JonathanC)
3D navigation controller follows straight lines in all dimensions between waypoints (Leonard,Randy)

WPNAV_SPEED, WPNAV_SPEED_UP, WPNAV_SPEED_DN, WPNAV_ACCEL allows configuring speeds and acceleration during missions

"compassmot" to compensate for interference on compass from the pdb, motors, ESCs and battery. (Randy,JonathanC) (Set-up video here)
Safety improvements:
- simple Tin Can shaped Geo Fence
- pre-arm checks to ensure all calibration has been performed before arming (can be disabled by setting ARMING_CHECK to zero). (video description here)
- GPS failsafe - switches to LAND if GPS is lost for 5 seconds
- stability patch improvements to stop rapid climbs in very overpowered or overtuned copters
Circle mode improvements including "panorama" when CIRCLE_RADIUS set to zero (Randy,Leonard)
SONAR_GAIN parameter added to allow better tuning of sonar surface tracking
CH8 auxiliary switch (same features as CH7)
works on PX4 (some minor features still not available) (Tridge,PatH)

How to upgrade:

1. Make sure you are using Mission Planner 1.2.59 or newer.

http://ardupilot.com/downloads/?did=63

2. Click on the MissionPlanner's Firmware screen and click the "Beta Firmwares" link on the bottom right. The version numbers should update to "ArduCopter-3.0.1rc2", then click the appropriate frame icon and it should upgrade as per usual.

3. Reduce the Loiter and Alt Hold PIDs if you have modified them from the defaults. The modified PID values for the 3dr frame can be seen in the image below.

Note: Nav parameters have been combined with Loiter so do not be concerned if you can't find them.

4. If you purchased an APM prior to March of 2013, update your PPM encoder to the latest firmware.

http://copter.ardupilot.com/wiki/common-ppm-encoder/

http://copter.ardupilot.com/wiki/common-ppm-encoder-apm2x-atmega32u2/

5. Try out the new version in stabilize mode first, then alt-hold, then loiter and finally RTL and Auto.

Special Thanks to Marco, DaveC and the rest of the beta testers for putting their copters at risk during the extended testing period. Some of their videos can be found here, here, here, here, here and here. Thanks also to MichaelO for the MP changes required for this release.

All feedback welcome. Please put your questions, comments (good and bad!) below.

Added by Craig:

Please watch Randy's videos on setting up and flying APM-Copter 3.0 before you go flying. They are excellent!

3DR Quad Set-Up Suggestions http://www.youtube.com/watch?v=33Tl_rjhQEE

AC 3.0 "Live" Compass Calibration http://www.youtube.com/watch?v=DmsueBS0J3E

AC 3.0 CompassMot Setup http://www.youtube.com/watch?v=0vZoPZjqMI4

AC 3.0 Pre-Arm Checks http://www.youtube.com/watch?v=gZ3H2eLmStI

AC 3.0 Fence https://www.youtube.com/watch?v=HDnGdo54o-4

AC 3.0 Maiden Flight Checks http://www.youtube.com/watch?v=NOQPrTdrQJM

I thought it might be worth trying to put an overall perspective on the "New" GPS warning.

There are basically 3 reasons for the GPS warning.

A. We have just figured out the potential greatest magnitude of the problem which is entirely related to GPS and not basically "fixable" except with military grade (externally ground source secondary transmitter) compensation. - (and that seriously isn't going to happen)

B. It is definitely worse during the high solar activity, more common for the next 2 years than normal, but still quite rare.

C. Because of the ability to respond faster, we have tweaked the response rates ~ 2.5x (which can be reset back to their previous levels) removing the aggressively part of the warning so you have time to respond.

The increase in response rates lets the copter hold position more tightly especially in gusting wind conditions and lets auto fly successfully and more accurately in conditions where it could not before.

Whatever you say about this system, it is going to be reflected identically in all systems relying on GPS for position determination including airplanes in flight (except the military and really big jets who can access real time compensation (with really expensive equipment) as well).

In fact, dialing down the nav response rate back towards the original values really removes the "panic" from this problem and will give you enough time to compensate for a wild GPS position, the only negatives are that it won't maintain as tight a loiter in gusty conditions and will be a little "sloppy" waypoint following.

The bottom line is that this is a pernicious problem relating to depending on GPS for absolute position maintenance and the only really satisfactory solution is to have an alternate and dependable method for position determination.

Possibly a fixed object that you could determine your position and motion relative to.

Nobody else has got this yet either (well my guess is the military does by now, but they certainly aren't talking about it.)

In fact, the Optical Flow camera is designed to do exactly that: show your ground relative motion.

Although they haven't talked about it much, the thing that could help Optical Flow the most is a stabilizing platform for keeping it pointed straight down regardless of airframe attitude.

Think about it, if the camera is bouncing all over the place, you then have to compensate by continuously shifting the image: computational nightmare or at least a time intensive pain in the program.

For your use for now, If you are uncomfortable with mastering the 1 second switch back to stabilize move, I would just dial back the nav rate parameters then you should have the advantages of inertial in loiter with the only drawback that it won't compensate as quickly in gusty conditions and might be a bit sloppier.

(It will still compensate better than the older versions, just not as quickly as with the rates set higher.)

Until we can incorporate a secondary position or motion determination system (not GPS) this is a fact of life for the APM / PX4 and every other GPS dependent flight controller out there.

Notice that even the new Naza system went to a lot of trouble to not actually use the GPS to actually do anything.

And the WooKong has exactly the same problems in this regard we do, remember New Zealand, that was a WooKong GPS pod sticking up and GPS multipath killed it deader than a door nail.

Optical flow could go a long way to providing an answer in the long run (fog could be a problem though).

Replies

palo sestak June 19, 2013 at 11:54am

I have just returned from field.

Have tested ACM 3.0.0. From beginning I was pleased - just glitches of motors that could be because of vibrations - I have no anti-vibration setup. Then I was going to switch to loiter. Just before I touched the switch quad dropped down suddenly 2-3 meters. I remained in STAB mode and then came next drop down. Fortunately had enough hight so didn't hit the ground.

Anyway level of adrenaline jumped too high. Rest of the flight was the fighting with unexpected level drops.

If enyone can help me what I'm doing wrong would appreciate it. Here is the log.

2013-06-19 20-03 2.log

https://storage.ning.com/topology/rest/1.0/file/get/3692749541?profile=original
Francis Ignacio June 19, 2013 at 10:52am

Randy and Dev Team,

Just need clarification regarding INS_MPU6K_FILTER setting.

I noticed that in 2.9.1b release the INS_MPU6K_FILTER setting was changed to 20hz as per the release notes.

Now with this release 3.0 in the Mission Planner, in the Advance Parameter List under the description for INS_MPU6K_FILTER, it reads that this setting should be set 42hz for Arducopter and 20hz for Arduplane. However, there is no mention of the INS_MPU6K_FILTER setting being changed back to 42hz in the 3.0 release notes.

Which is the correct INS_MPU6K_FILTER setting for the 3.0 release for Arducopter, 42hz or 20hz?

I commend you all for this release.

Thank you!
Gary McCray June 19, 2013 at 10:51am

I thought it might be worth trying to put an overall perspective on the "New" GPS warning.

There are basically 3 reasons for the GPS warning.

A. We have just figured out the potential greatest magnitude of the problem which is entirely related to GPS and not basically "fixable" except with military grade (externally ground source secondary transmitter) compensation. - (and that seriously isn't going to happen)

B. It is definitely worse during the high solar activity, more common for the next 2 years than normal, but still quite rare.

C. Because of the ability to respond faster, we have tweaked the response rates ~ 2.5x (which can be reset back to their previous levels) removing the aggressively part of the warning so you have time to respond.

The increase in response rates lets the copter hold position more tightly especially in gusting wind conditions and lets auto fly successfully and more accurately in conditions where it could not before.

Whatever you say about this system, it is going to be reflected identically in all systems relying on GPS for position determination including airplanes in flight (except the military and really big jets who can access real time compensation (with really expensive equipment) as well).

In fact, dialing down the nav response rate back towards the original values really removes the "panic" from this problem and will give you enough time to compensate for a wild GPS position, the only negatives are that it won't maintain as tight a loiter in gusty conditions and will be a little "sloppy" waypoint following.

The bottom line is that this is a pernicious problem relating to depending on GPS for absolute position maintenance and the only really satisfactory solution is to have an alternate and dependable method for position determination.

Possibly a fixed object that you could determine your position and motion relative to.

Nobody else has got this yet either (well my guess is the military does by now, but they certainly aren't talking about it.)

In fact, the Optical Flow camera is designed to do exactly that: show your ground relative motion.

Although they haven't talked about it much, the thing that could help Optical Flow the most is a stabilizing platform for keeping it pointed straight down regardless of airframe attitude.

Think about it, if the camera is bouncing all over the place, you then have to compensate by continuously shifting the image: computational nightmare or at least a time intensive pain in the program.

For your use for now, If you are uncomfortable with mastering the 1 second switch back to stabilize move, I would just dial back the nav rate parameters then you should have the advantages of inertial in loiter with the only drawback that it won't compensate as quickly in gusty conditions and might be a bit sloppier.

(It will still compensate better than the older versions, just not as quickly as with the rates set higher.)

Until we can incorporate a secondary position or motion determination system (not GPS) this is a fact of life for the APM / PX4 and every other GPS dependent flight controller out there.

Notice that even the new Naza system went to a lot of trouble to not actually use the GPS to actually do anything.

And the WooKong has exactly the same problems in this regard we do, remember New Zealand, that was a WooKong GPS pod sticking up and GPS multipath killed it deader than a door nail.

Optical flow could go a long way to providing an answer in the long run (fog could be a problem though).
James W Powell June 19, 2013 at 9:40am

I cannot do the compassmot, I always get this error: Failed! Compensation disabled. Did you forget to raise the throttle high enough? I have raised it to 75% for 5 seconds and quickly pulled the stick down, to 80%, 90%, 100%, even tried just 50% , I always get the same result. I followed the instructions per this video to the T: https://www.youtube.com/watch?v=0vZoPZjqMI4

Any help/advice would be appreciated :)
palo sestak June 19, 2013 at 6:21am

By descriptions it promises a lot over versions till 2.9. Will going to test it in the evening.

But what I understood from Randys writing when I touched some param it will stay untouched by loading new firmware. I have touched some and want to start testing ACM 3.0 with defaults. Do I need to reset them somehow or can somebody upload here the default params ?
theo van niekerk June 19, 2013 at 5:13am

looking good guys thx....willtest soon
Waladi June 19, 2013 at 5:10am

Hi Randy and Dev team.

Finally.. it's definitely comparable with Naza now, thanks for the hard work.

On last RC5, I notice that if I set waypoints (without any RTL on it) and click on "Write WPs" button then my RTL_ALT parameter will be set to 10000 (100m) even if I had it on zero before and it's same value as Default Alt on Flight Planner. Is it bug ?

Thanks in advance.
Виталик June 19, 2013 at 4:39am

Team Land. It was he who sits at the penultimate point waypoint or point of land if there are its coordinates.
Peter Koek June 19, 2013 at 3:51am

I have programmed my APM 1 with version 3.0.0 through Arduino and performed a factory reset. All is well, until I enable the optical flow sensor in mission planner, after which the helicopter simply stops responding to a regular connect.

It's a little bit different for CLI, as it keeps functioning. However, when I try to enable the optical flow sensor through CLI, it end up rebooting a few times followed by nothing. Reopening the terminal and rechecking the optical flow settings show it as disabled! Testing the flow sensor outputs only zeros.

I have currently switched back to 2.9.1b, which does not have any of the aforementioned issues. Optical flow works fine with this version.
Itzik Ronen June 19, 2013 at 3:01am

Hi

Thanks for all the info, its help me to solve the problems with my quad.

where can I download the Mission planner 1.2.53 mav 1.0?

where I define the fence radius. Is the radius is always from starting point?

Thanks

Itzik

of 28

This reply was deleted.

Discussions

ArduCopter-3.0 ready for wider use

Replies

Activity