Physics equation for optimal robot path from point A to point B

[Kura_kai]

I am new here and i am a java programer and i needed a physics equation. I need to get an equation that can allow a robot to calculate how to get from point a to b the best way. I have been thinking of something like that but keep drawing blanks. Any help

 

[Davorak]

You are not giving use much to go on. Is it a straight line? does the robot have a map or does it need to observe it surroundings?

Graph theory has many common algorithms to determine the lowest cost path. Is this what you are looking for?

 

[Kura_kai]

It is going in a straigh path the distance is going to be like 50 feet or something.

 

[ZapperZ]

It is going in a straigh path the distance is going to be like 50 feet or something.

.. and the simple equation of a straight line like y = mx + b just doesn't cut it?

Zz.

 

[Gokul43201]

If the co-ordinates of the 2 points are A(x1,y1) and B(x2,y2), the equation of the line passing between them is :

$$y - y_1 = \frac {y_1 - y_2}{x_1 - x_2} ~ \cdot ~ x - x_1$$

 

[Davorak]

Please tell me there is something more too it then this, there are bumps on the ground, there is a hill in the way, or some sort of obstacles. It is not flat ground for 50 feet is it?

 

[dextercioby]

Another question:is this a geodesic or a brahistochrone problem...?

Daniel.

 

[ZapperZ]

Please tell me there is something more too it then this, there are bumps on the ground, there is a hill in the way, or some sort of obstacles. It is not flat ground for 50 feet is it?

I vote for the speed bump...

Zz.

 

[Kura_kai]

It is a straight path for 50 feet. The equation is meant to find figure out how big the steps have to be so it doesn't fall forward or something other than that there is not going to be a sudden unexcpected force accting on the robot.

 

[ramollari]

That's too ambitious kura_kai to be handled by a simple Java program. There are two things you can consider to drastically improve you robot design: 1. use the reactive agent architecture instead of symbolic reasoning (or at least hybrid), 2. Incorporate some form of learning to your robot, so that the latter itself decides which are the most optimal movements and step sizes. Sth like reinforcement learning.
Good luck!

 

[ramollari]

BTW have a look at this link: http://hebb.mit.edu/people/russt/ro...le/russt/robots/Toddler_v5/learningMedley.wmv.

Besides, to make the robot deal with dynamic environments, or ones with bumps, map percepts directly to actions with priorities without trying to calculate the coordinates. You could put for example a radio wave signal at the destination and make the robot move uphill the signal.