programmer-man's profile - activity

Yup, I can ask it there. However, I previously got feedback that ROSAnswers is the more appropriate forum for this type

Why not include a velocity profile in the controller input path? In nav2_core::Controller the input path is a nav_msgs/P

Why not include a velocity profile in the controller input path? In nav2_core::Controller the input path is a nav_msgs/P

Would it not make sense to have the same planner with multiple interfaces? I.e., one for NAV and one for Moveit.

I have another question about the design choices of nav2. Specifically, I'm wondering why the controller plugins (nav2_core::Controller) don't have a mechanism for failure. Consider the following example as motivation.

Suppose we have a very forgiving progress checker (our robot is slow and unreliable and so we have a 10 second timer for it to raise an error) and suppose we have a simulator in our controller plugin that can detect problems. Then in the event that our robot detects a problem in simulation, we will have to wait an additional 10 seconds for the control server (nav2_controller::ControllerServer) to quit. This seems far too long since we were able to detect a problem in the controller plugin 10 seconds early.

Looking through the code, I'm wondering why nav2_controller::ControllerServer doesn't allow for dynamic goal checker plugins like it does for the controller plugins. Take the following examples as motivation.

Scenario 1: We want the robot to interface with a control panel and so we give it a path that ends with it facing the panel (orientation is important, thus for this scenario the goal checker needs to ensure we have achieved the correct pose).

Scenario 2: We want the robot to communicate with another robot and so we give it a path that brings the robot within range of the other robot (we don't want it to go out of it's way to orient itself since in this case orientation doesn't matter).

Of course, these two scenarios would likely require different planners and or controllers but wouldn't it be ideal that our controller returns true if we achieve the desired result (the correct location and orientation for scenario 1 and just the correct location for scenario 2)?

PS. I would think we would also allow for dynamic plugins for the progress checker, because you never know.

Hey guys,

I'm trying to understand the intent behind some of the design choices of nav2_controller::ControllerServer and nav2_controller::SimpleGoalChecker. Specifically, it looks like the goal checker will return true when it is in the vicinity of the target pose. Subsequently, the controller seems to constantly check if we have reached the goal. The result being that we can't give the controller a path that loops over itself, i.e., it will exit and return successful the first time it reaches the goal.

Example, suppose I want to have the robot drive 10 laps around a closed track and so I provide a path that goes around the track 10 times, but when I give the path to the controller it will exit after the first lap, in fact it might immediately exit since our initial pose and goal pose are the same.

Of course we can go with a different goal checker but the interface only provides the query pose, the target pose, and the velocity as input, thus it makes it hard to extend it to track the path until the robot has driven the entire path. Of course we could have it subscribe to the controller's action messages but that seems less than ideal, if this information was important then it should be provide it via a method call. Alternatively we could use the velocity to skip the other poses but that too seems to have limitations.

How was the above scenario intended to be solved? Thanks

Hello,

There seems to be some overlap between ros navigation and moveit. Navigation seems to be more concerned with mobile bases, while moveit is concerned with manipulators but like I said there seems to be some overlap. Is there a document that explains the scope and use cases of these two frameworks/systems?

Thanks

Nav2 vs. Moveit Hello, There seems to be some overlap between ros navigation and moveit. Navigation seems to be more c

https://answers.ros.org/question/9831/subscribing-to-action-servers-feedback-topic/

PS. I seen the other question/answer about this topic but the answer didn't seem relevant to my question.

How to subscribe to an action server's feedback What is the recommendation for subscribing to the feedback of an action

PS. Where should I be asking these questions? Here? Discourse? Or somewhere else? Thanks.

Thanks Steve, I didn't see the try and catch, I was just looking at the API. PS. Thanks for all of your other answers a

Why doesn't the nav2 controller plugin API have a method for failing? I have another question about the design choices o

Why is SimpleGoalChecker so simple? Hey guys, I'm trying to understand the intent behind some of the design choices of

Why not allow for dynamic goal checker plugins? Looking through the code, I'm wondering why nav2_controller::ControllerS

Why is SimpleGoalChecker so simple? Hey guys, I'm trying to understand the intent behind some of the design choices of

And what would you say is the normal division of labour? Do you think my summary is accurate?

What is the scope of a local controller? Hello, I'm trying to nail down the scope of local controllers, local planners,