I have been working on making my flights more autonomous and so far I have had a great deal of success. The setup shown in the video has made eight successful fully automated flights with no operator input from takeoff to landing. This flight was different, and I would appreciate some assistance trying to determine what went wrong.
I planned a simple mission with a takeoff to 50 meters, three standard waypoints, a RTL, and land. The quad ascended about 2 meters and made a hard bank into the trees. Wind was about 20kts to the SE, so the quad was flying almost into the wind. Below is a recreation of the actual flight path the aircraft took. The KML appears to show the aircraft traveling in the opposite direction.
Note: The "Home" location shown below is from the second time the motors were armed when I flew out of the tree.
Hmmm...I had a very similar thing happen today. I performed a flawless autonomous flight and after landing and waiting for a few minutes tried the exact same flight. The copter took off and headed at top speed in almost the exact opposite direction. I hit RTL to no avail and tried to go to manual but it was too late. It is a minor miracle that mine was not damaged more than it was. I lost a video transmitter antenna, some landing gear, a battery, and a couple of props...
Logs are attached and video is here: http://www.youtube.com/watch?v=tWPeH7VqeQQ
The telemetry is baffling to me as it shows the copter knew exactly where it was supposed to go and that it new it was heading in the opposite direction....
The battery was not changed, it was fresh before the first/successful flight (takeoff, go to 25 meters, hit 3 quick waypoints all within a few 100' of each other then land) which was less than 2 minutes long (normal flight times for my setup are around 10 minutes before first warning). I do not know what happens after you successfully complete a set of waypoints having never done the same one twice before. My assumption was until you wrote new waypoints to the APM the previous waypoints would stay... The telemetry data makes it look like it knew exactly where the waypoints were.
OK, Tried the something similar yesterday. Not an expected outcome.
The unit was decending at its last (second last) waypoint as expected (land command on 5th wpt) after 4 waypoints and auto take off.
Pressed restart mission about 10 meters alt and the unit took off to wpt 2 at low level and descended slowly over 30 meters traveled so I took over.
WPT 1 is the takeoff waypoint.
I should have placed wpt 1 over wpt 4 so that it takes off and lands at the same spot?
And why the low altitude departure?
I had the external compass installed 180º wrong. So it was doing the same as your copter. I just put it right and now it works great!