Refactor mdr_find_people

[minhnh]

As discussed in #145, we need further discussion in how to handle mdr_find_people.

Yeah but mdr_detect_person can be extended to have 3D processing, instead of creating a new package. OpenPose and other methods can also be added later there.

mdr_find_people, I think, will potentially use mdr_detect_person for detection but most likely also involve navigation and moving the head, which may need to be made into a more complex skill. It may also be redundant with find_object skill, depending on how we implement that.

With that in mind, I think it's a good idea to merge the detection and 3D calculation into mdr_find_people. There's a portion which handles inserting the person's info into the knowledge base, however, that may remain here in mdr_find_people.

_Originally posted by @minhnh in https://github.com/render_node/MDIzOlB1bGxSZXF1ZXN0UmV2aWV3VGhyZWFkMTc0NjAzMzE5OnYy/pull_request_review_threads/more_comments