[ros-users] [Discourse.ros.org] [Humanoids] Humanoid_msgs Definitions

[Alexander Stumpf]

Hello,

in our footstep planning framework, we are using nested msg types to model the level of detail required by the different processing steps for 3D planning:
https://github.com/team-vigir/vigir_footstep_planning_msgs/blob/master/msg/StepPlan.msg
https://github.com/team-vigir/vigir_footstep_planning_msgs/blob/master/msg/Step.msg
https://github.com/team-vigir/vigir_footstep_planning_msgs/blob/master/msg/Foot.msg

The most difficult part is to cope with the distinct data required for the motion/walking software. Depending on the actual implementation of the walking algorithms, very different data is requested. From my experiences, I've seen that BDI wants to know the "lift" height in _step mode_, IHMC could handle foot contact area information and ROBOTIS THORMANG allows even to set timings for COM/ZMP trajectory calculation. We spend a lot of time to implement a framework which allows for handling this issue.

We have solved the variation of requested data by adding an uint8 array, to encode custom robot specific data. Furthermore, our step messages contain attributes for operator interfaces in order to enable supervised planning ("interactive planning").

>From previous comments, I can already see a lot of different use cases. I guess we should first collect a clear list of requirements by the community and then work out a proper design for the messages, so everyone could work with it.

At this early stage, we should already strongly consider to design the message in way that it is usable for n-legged robots.

@dkotfis: Yes, would like to use common messages in our footstep planning software. But the current humanoid message does not provide the required features.





---
[Visit Topic](https://discourse.ros.org/t/humanoid-msgs-definitions/2119/12) or reply to this email to respond.