Obstacle avoidance
from Malaysia  [14 posts]
7 year
Im unable to solve this type obstacle avoidance;
Kindly can you share the idea to make the robot avoid the obstacle using roborealm

I attach the exact video where the robot perform obstacle avoidance :

Steven Gentner from United States  [1446 posts] 7 year

I'd first try to find some video seen from the actual robot as this will be very useful for testing. If you cannot, you will need to create a mock arena that you can test on.

This appears quite straightforward. As the obstacles are all black one can even just use the average intensity of the image to avoid obstacles. I.e. connect the camera, add the Color_Statistics module and then check the MEAN_COLOR_G. If that value falls below some threshold (say 40) then you are about to hit a black object ... so move the robot accordingly until that value is above 40 again.

A lot of this will be based on what camera you use and how your robot is constructed ... but because of the arena layout, this isn't a very difficult vision challenge but will be more about how well you can control your robot.

from Malaysia  [14 posts] 7 year
thanks steven,
i got your concept by using the color_statistic and observe the mean_color_g.

so, if the value mean_color_g <40,

how i do perform the object counting based to the black box ?
Steven Gentner from United States  [1446 posts] 7 year
"how i do perform the object counting based to the black box ?" - Not sure what you are asking here. Do you need to count the black boxes? To avoid them you don't ...

from Malaysia  [14 posts] 7 year
Because the obstacles are place in random. My strategy is to determine the robot path by facing the block by counting ( counting enable when the robot are near to black block), so the robot follow the path by read the sequence of counting variable. So, what your opinion about the concept?

Steven Gentner from United States  [1446 posts] 7 year

Thanks for your image, your definition of "counting" was a bit confusing. What you are really talking about is robot state or a state transition diagram. This allows you to execute different actions based on past experiences. There are many ways to do this but you can start with the State Machine module:


If this is still confusing you can search the web for "state transition diagram" or "finite state machine" and gets lots of material about that.

Each time you encounter a black block you can transition into a another state that moves you right or left for a number of seconds, encoder ticks or until the camera is ok with what it sees (i.e. no black block). Then that state would transition into a move forward state until the next black block is encountered.

Start with just a single block and figure out how to get around that. Once you have done that, the next new states are easier.


This forum thread has been closed due to inactivity (more than 4 months) or number of replies (more than 50 messages). Please start a New Post and enter a new forum thread with the appropriate title.

 New Post   Forum Index