My Second Experiment Ready! Your Opinion is Welcome.

[JavaRN]

I have just posted a video on you tube to show my second experiment. In this experiment the Robot has to look for the red box again but this time it is hidden in a small "home" consisting of two rooms. In the video I've chosen to place the box in the second room so that it takes longer for the robot to find it. To better understand how things work, this is a screen shot of the controller program (written in Java):



On the top left hand side of the window are the Sensor Readings, the program uses these readings to check where the robot is and to calculate the next move.

The Robot Data section shows the current move the robot is performing.

In the Object Information section details of where the box is is displayed - in this case the object is near the robot on his left hand side. Of course this section changes according to where the object is currently found from the robot's point of view.

I use the Application Log for debugging, the application stores in it details of the program execution - sensor readings, movements, image recognition notes etc.

On the right hand side of the window there are two views - the top one is a real-time view of what the robot sees. The lower one is for image recognition (you can see a black border box surrounding the red object).

If you are interested in the video, click on the link below:


http://www.youtube.com/watch?v=xpcLjWfUtIw



Charles

[DirtyRoboto]

Very good work my Robobretheren. You are taking the RN to new heights.

What is the decision tree for whether to cartwheel or walk?
Do you have the layout parameters in the program or does RN feedback wall position/distance?

What is the decision tree for search method, ie. where to search first?

Nice work and keep it up mate.

Marcus.

[JavaRN]

What is the decision tree for whether to cartwheel or walk?

My robonova is equipped with four IR sensors one n front two on the sides and one on the back if the left or right sensors show a reading > 70 then a cartwheel is possible, other wise he has to slide his way.

Do you have the layout parameters in the program or does RN feedback wall position/distance?

It is both, the robot has an idea of the layout of the maze, but the moves are not predefined (only two moves are), the program uses the sensors to determine the next move.

What is the decision tree for search method, ie. where to search first?

The box can be in one of the two rooms. As you might have noticed I placed the box in the second room to make it more difficult in fact he looked in the other room first. (may be I should post another video with the box in the first room).


Charles[/code]