fix atama tracking

[Mobilizes]

TODO :
try publishing /measurement/status, since atama needs robot yaw
when robot looks at ball, kansei breaks and robot freezes. probably something to do with atama needing raw image

[Mobilizes]

update, publishing button stops the detection for some odd reason

[Mobilizes]

update, after doing this sequence:

1. running kansei (kansei is obsolete since rpy subscription is disabled), akushon, aruku, soccer and web using shell script
2. set mode to pan tilt
3. run tachimawari

and added debug prints in soccer, tweaks in process vision, soccer_node did receive the ball information.

however, after running tachimawari, soccer_node stops subscribing to ninshiki_cpp/dnn_detection

it did move the head to last known ball position in webots, albeit the tilt is off.

note : tilt in the terminal is 0.0, will to look into that

[Mobilizes]

update, soccer node is not the only thing thats off, its the entire soccer, and this MIGHT be the cause

looking into this

update again, yes it is the cause.

[Mobilizes]

update, the initial walk ready on webots is stuck, so if i overrode the action with anything, it CAN detect the balls now.
however, the ball coords are changing, but the pan is not moving

[Mobilizes]

update, did some more debugging on action, turns out the interpolator keeps initializing after action ends, repeating the process over and over again, this is due to behavior's is_action_running does not update

run action until finished stops soccer_node, which in turn does not update the is_action_running in behavior, causing it to keep playing walk ready.

really confused because in the robot, the soccer node (supposedly) keeps running, or maybe its my understanding that is lacking? going to the lab to check, after this, then i will check actual reason why pan is not moving so i can get real progress

[Mobilizes]

update, running soccer test proves that soccer CAN change the pan and tilt in webots, it just wont in the soccer run, i hope i can finish this issue by the end of this week. i've been procrastinating for too long. man i should really have a consistent schedule

[Mobilizes]

update, to make the robot 'track',

1. simulation_mode = true in config
2. run soccer, no matter if test or not
3. open web in soccer settings
4. wait until robot walk ready
5. change mode to pan tilt

problem :

• robot has this "NONONONOONNONONO" head motion (shaking head horizontally) when ball is to its right, not even tracking when ball is to its left
• robot only moves its head because of button is not none (when robot reset orientation and move by angle to 0.0 and 0.0 in process mode behavior). when controller button is none, the pan and tilt will have min value
• orientation in robot is discretely changing from 0 to above 100 degrees (unrelated to point above, which makes no FUCKING sense)
• from echo set joints topic, it is found that pan is only negative and never went to positive, which is why it couldnt track to the left

[Mobilizes]

update, kansei is the reason why the orientation goes to 0 discretely, the tracking is still problematic, so there are two things that may be why :

1. RPY that robot sends are wrong right from the start
2. tracking is not compatible in robot