Obstacle avoidance with laser or kinect data
I wish to have a basic obstacle avoidance code when a robot is moving (without a planned path). This is the goal: I have three turtlebots following a leader bot. The leader uses navigation stack to create and follow a path in the map. The three followers need to follow the leader in a formation and avoid obstacles. They change their formation if they encounter obstacles.
One way is by using data available on /turtle01/scan. Using a threshold value, I can look for the minimum values, compute their corresponding angle and make the robot move away from that direction. However, that seems like a very computationally expensive and ineffective technique. Is there a package available for this?
I know the question is very similar to previously asked questions but I haven't found anything useful yet. Feel free to redirect me if the answer exists. Thank you.
I suggest that you break this into multiple questions. I see at least three different questions. You'll get better, more focused answers that way.
I have edited and re-organized the question. Thanks!
Thanks for the edit, but I meant for you to create multiple questions, as in delete two of the questions here and create a new question (post) for the each of the others. Multiple questions are, many times, difficult to answer at once. Sorry for the confusion.
I thought the edit would help to see them as different viewpoints for the same question so that someone could tell me which way is the best. But I get your point, it just makes the question a little vague to understand. I removed it now. Thank you.