Hello all and msf_developers(@simonlynen , @markusachtelik)
Thank you for your countributions to msf algorithm.
I have two questions regarding use of MSF.
My system uses two vision sensors for drone navigation. One for april tag information and the other for ORB_SLAM2. And a Pixhawk IMU is used for system update. April tag gives absolute position and ORB_SLAM2 gives relative position and orientation with drift.
First, I use a pose-position_msf, ORB_SLAM2 for pose sensor handler and april tag for position sensor handler. However since orientation of ORB_SLAM2 drifts during long term test, I want to integrate it with orientation measurement from tag. I know that #80 issue deals with pose-pose-msf but I failed to make it.
So is there someone who made example or documents for pose-pose-msf?
Secondly, when ORB_SLAM2 worked, it only gave 7~8Hz measurements and result of msf showed saw- shaped graph which is not good for drone flying stabilization. I guess it is a cause of low frequency measurements. Is it a normal result? Or do I need to tune some parameters to have a better result?
Hello all and msf_developers(@simonlynen , @markusachtelik) Thank you for your countributions to msf algorithm. I have two questions regarding use of MSF. My system uses two vision sensors for drone navigation. One for april tag information and the other for ORB_SLAM2. And a Pixhawk IMU is used for system update. April tag gives absolute position and ORB_SLAM2 gives relative position and orientation with drift.
First, I use a pose-position_msf, ORB_SLAM2 for pose sensor handler and april tag for position sensor handler. However since orientation of ORB_SLAM2 drifts during long term test, I want to integrate it with orientation measurement from tag. I know that #80 issue deals with pose-pose-msf but I failed to make it. So is there someone who made example or documents for pose-pose-msf?
Secondly, when ORB_SLAM2 worked, it only gave 7~8Hz measurements and result of msf showed saw- shaped graph which is not good for drone flying stabilization. I guess it is a cause of low frequency measurements. Is it a normal result? Or do I need to tune some parameters to have a better result?
Thank you