strands-project/scitos_apps

very low speeds with undocking

Closed this issue · 10 comments

I have calibrated the docking, and betty docks normally. However, when i tell her to undock, it seems to start with normal negative speeds, but then it gets veryc close to zero and the robot doesn't move anymore. here is the firt few published velocities:

linear: 
  x: -0.5
  y: 0.0
  z: 0.0
angular: 
  x: 0.0
  y: 0.0
  z: 0.0

---
linear: 
  x: -0.5
  y: 0.0
  z: 0.0
angular: 
  x: 0.0
  y: 0.0
  z: 0.0

---
linear: 
  x: -0.5
  y: 0.0
  z: 0.0
angular: 
  x: 0.0
  y: 0.0
  z: 0.0

---
linear: 
  x: -0.5
  y: 0.0
  z: 0.0
angular: 
  x: 0.0
  y: 0.0
  z: 0.0

---
linear: 
  x: -0.00750212669373
  y: 0.0
  z: 0.0
angular: 
  x: 0.0
  y: 0.0
  z: 0.0

---
linear: 
  x: -0.00750212669373
  y: 0.0
  z: 0.0
angular: 
  x: 0.0
  y: 0.0
  z: 0.0

---
linear: 
  x: -0.00750212669373
  y: 0.0
  z: 0.0
angular: 
  x: 0.0
  y: 0.0
  z: 0.0

---
linear: 
  x: -0.00750212669373
  y: 0.0
  z: 0.0
angular: 
  x: 0.0
  y: 0.0
  z: 0.0

---
linear: 
  x: -0.00750212669373
  y: 0.0
  z: 0.0
angular: 
  x: 0.0
  y: 0.0
  z: 0.0

It keeps publishing these very slow speeds afterwards.

Maybe the obstacle detection?

Probably obstacle detection as pointed out by @cdondrup . Can you send a rosbag with the head camera topics ?

I think Betty might be having issues with the canera topics because of this and some other issues with the calibration but I guess @gestom will know better

will send the rosbags tmrw. She's able to dock though. Is there any calibration we can run for the head cam? We did have to repare her head cam halo, so it might not be exactly the same as before

also, there are some network issues and the rate of the head cam on betty is reduced:

strands@betty:/media/strands/PLANSIG2013/bham_integration/semantic_maps$ rostopic hz /head_xtion/rgb/image_raw
sh: 0: getcwd() failed: No such file or directory
subscribed to [/head_xtion/rgb/image_raw]
average rate: 12.678
    min: 0.078s max: 0.080s std dev: 0.00044s window: 12
average rate: 12.670
    min: 0.078s max: 0.080s std dev: 0.00053s window: 24
average rate: 12.669
    min: 0.078s max: 0.080s std dev: 0.00052s window: 37
average rate: 12.669
    min: 0.078s max: 0.080s std dev: 0.00052s window: 50
^Caverage rate: 12.670

@bfalacerda: My zip version complaints that the files are incomplete, can you check ?

yeah, there was something wrong with them. Please try again

Originally, the depth image data encoding was UINT16, now it's FLOAT32. This was not reflected when calculating obstacle distance during undocking. Should be solved in #154.

It still works on Linda using the new openni2 launch files though. Are we sure this is not just a Betty issue?