Closed MFernandezCarmona closed 6 years ago
I think I got the problem: I was using both "with_qsr_lib" and "with_qsr_prob_rep" as true. Using "with_qsr_lib"=true only, I get another error:
[ERROR] [1523467370.151489, 443.876000]: timeout exceeded while waiting for service /robot1/prob_rep_ros_server/pf/update
Seems that only "working" config is "with_qsr_prob_rep" as true. Thoughts?
The second error looks like it is trying to contact the particle filter but can't because it isn't running. So yeah, it needs with_qsr_prob_rep:=true
. Can you send me a rosbag with system output that breaks the particle filter? I'll try to recreate the error here and have a look.
And the models you are using as well, please.
Models are the default ones. Also, as it's a simulation, I think it's easier if you just clone https://github.com/LCAS/iliad.git and launch my simulation roslaunch hri_simulation iliad_sim_hri.launch gazeboGui:=false
Everything is remapped at /robot1
and usually that's the source of my problems.
It's a testbench repo, probably there will be many dirty things. Apologises in advance!
Ok, I think I found the problem. In file qsrrep_pf/probability_density_functions.py
At line 116
o = np.array(self.models[self.models.keys()[m]][self.key])[p.astype(int),x[0]]
That x[0]
index is not casted into an int. I changed to:
o = np.array(self.models[self.models.keys()[m]][self.key])[p.astype(int),int(x[0]])
And now I don't have the error...
Did you try any of these in kinetic/Ubuntu 16.04?
Sorry I didn't have time to look at it yet. Glad you found the error.
I never tried this in Kinetic, no.
Hello, I'm modifying strands human aware navigation stack to be used in multiple robots, and it's mostly working. I did minor modifications in several packages to support renaming topics, services, nodes and namespaces mostly. Currently I'm working with a gazebo simulation where a human walks towards the forklift. It's in package hri_simulation, iliad_sim_hri.launch.
But I'm get this error every time gazebo actor is detected as a tracked human. @cdondrup, Do you have any idea what may I be doing wrong? Thanks!