ROS pepper odometry

[Herman]

Hello,

Im trying to learn the Pepper robot to navigate through an environment on its own via deep reinforcement learning.
 I m locating the robot by using the the data received from the /pepper/odom topic.
But I dont know how i can reset the accumulated distance after each episode, as i dont want to close the simulation but just set the Pepper back to one of its starting points.

Is there a way to reset the /odom data in a python script?

I really would appreciate any help

Best regards
Hermann

[Daniele Proietti]

Hi Hermann,

I am wondering the same thing but in the real robot, do you know if there exists a way to reset the odometry data provided by the naoqi sdk? 

Best 

Daniele

[Herman]

Sorry but I dont work with the Naoqi sdk, so i cant help you with that

[HaSu]

Not sure how up-to-date this still is, but did you get it solved? You could publish a new localization pose on the specific topic. Wouldn't that reset the odometry also?

We have a node that solves some localization issues by re-localising the robot on a regular basis, but I have never tried it on Pepper.

Did your approach work? Pepper's navigation capabilities in large open spaces are very bad unfortunately. :(