Robot Commander - Find object info needed


#1

When I run the robot commander app I see there is a “find (target)” option.

What are the list of “targets” and can I add some of my own as well?


#2

If you look at the documentation here:

You can see that this uses the inbuilt DNN coupled to the camera to find objects in the room. It can recognize about 20 different pre-trained objects namely:

aeroplane, bicycle, bird, boat, bottle, bus, car, cat, chair, cow, dining table, dog, horse, motorbike, person, potted plant, sheep, sofa, train, tv monitor

You can add your own but you’d need to re-train the DNN on the robot to do that - this is non-trivial.


#3

In all my attempts to find objects using the robot commander app I have yet to get Magni to move or find any object.

What gives?


#4

You need to have the dnn node (deep neural network based classifier) running on the robot before you can have this functionality work.

On the robot run:
roslaunch magni_demos dnn_rotate.launch

Rohan


#5

Is there a way to know whether the dnn node is running?


#6

You can run rosnode list and see if dnn_detect and dnn_rotate are running.