Read object positions from ROS <-> gRPC bridge

### Description

Right now, the object abstraction for RosGrpc is just a mock. We should pass the current objects poses from the ROS to our learning environment.

### References

_No response_

Start by locating the RosGrpc object abstraction and the ROS-to-learning-environment entry points; the issue says this abstraction is currently a mock. Trace where current object poses could cross the bridge, and consider the work done when the learning environment receives those poses instead of mock objects.

由索引模型根据 Issue 内容生成。