Path planning AD*/A*

[Ninatove]

Hello!

We have nearly any Labview experience, and got trouble trying to plan a path for our autonomous mobile robot.

We based our program on an already existing example on AD* path planning.

This example was found in NI Example finder: Robotics/path planning/Anytime D Star on an occupancy grid.

We deleted the part outside the first while-loop, and replaced it with a self-created defined map.

A problem occurs as the program plans the path. The program itself is running but we dont get the planned path in the front pannel. The only thing that occurs are the start-positions. We also get the error code "1556".

The same error code (1556) occurs in the original example-program, but at least the planned path is visible.

If any of you have more experience using A*, we're also happy to recieve some advice about how to get the path reference after planning the path.

Hoping for answers 🙂

[RoboticsME]

We actually have a discussion forum for Robotics Module specific questions: http://forums.ni.com/t5/forums/searchpage/tab/message?submitted=true&type=message&q=robotics&&page_s...

You may be able to get better support there since we have support engineers monitoring the forums.  There are also other support options at: http://www.ni.com/support/

I took a quick look at your code, and it was missing GreenPrint.vi and BluePrint.vi.  If I deleted the missing VI's and just wired through the double array, I didn't get any errors.

Make sure the array you create is big enough and that you don't have zero cost values.  I believe if you have a cell as "0" in cost, the algorithm may not work as you expect.  There should be some small cost associated with even "object free" cells to represent the cost associated with travel.  In other words, if there is no cost to move from one cell to the next, then I can move back and forth an infinite number of times and not accumulate any cost. 

Since it looks like you're converting booleans in to doubles, you'll have cells with zero cost.  I would suggest adding a constant (say 0.01) to all the cells in the array.

I dont' know about the error you're getting, but post your question on the link above to get help from an Application Engineer.

Cheers,
Karl

[Ninatove]

Thank you Karl,

I really appreciate your advice.

Cheers

Nina

Date: Tue, 29 Mar 2011 09:07:41 -0500

From: web.community@ni.com

To: nina_u_89@hotmail.com

Subject: - Re: Path planning AD/A

Community

Re: Path planning AD/A

created by RoboticsME in LabVIEW Robotics - View the full discussion

We actually have a discussion forum for Robotics Module specific questions: http://forums.ni.com/t5/forums/searchpage/tab/message?submitted=true&type=messag e&q=robotics&&page_size=50&board_id=170&sort_by=-date

You may be able to get better support there since we have support engineers monitoring the forums. There are also other support options at: http://www.ni.com/support/

I took a quick look at your code, and it was missing GreenPrint.vi and BluePrint.vi. If I deleted the missing VI's and just wired through the double array, I didn't get any errors.

Make sure the array you create is big enough and that you don't have zero cost values. I believe if you have a cell as "0" in cost, the algorithm may not work as you expect. There should be some small cost associated with even "object free" cells to represent the cost associated with travel. In other words, if there is no cost to move from one cell to the next, then I can move back and forth an infinite number of times and not accumulate any cost.

Since it looks like you're converting booleans in to doubles, you'll have cells with zero cost. I would suggest adding a constant (say 0.01) to all the cells in the array.

I dont' know about the error you're getting, but post your question on the link above to get help from an Application Engineer.

Cheers,

Karl

Reply to this message by replying to this email -or- go to the message on Community

Start a new discussion in LabVIEW Robotics by email or at Community

[二丶喵]

Has this problem been solved?