Team members:
- Ali Ozgur Abali
- Leonardo Carrasco
- David Cox
- Randy Direen
- Lisa Prince
We propose to build an autonomous robot with the capability of mapping out its surroundings. The robot we plan to build will have the capability of mapping out its environment using ultrasonic vision, and build a 2-D map based on what it sees. Using the map it has constructed in memory the robot will make decisions where to move by calculating which direction will be the easiest to move in. The purpose of such a machine would be the first step in developing a robot that could intelligently "learn" its environment and use what it has learned to make decisions for future tasks. Future tasks could involve labeling key objects the robot has found and having the robot navigating to those objects by command then performing some function on the object. We believe that this is a first step in developing robots that can help around the house and in industry.
The machine vision will be based on three equally distributed ultrasonic transducers. The Center transducer will send out a short ultrasonically modulated pulse and the two side transducers will receive the reflected pulse. Using a scheme antenna engineers use to increase pattern directivity, the two receiving transducers will have their signals modulated together so as to only pick up reflected signals directly in front of them. By putting the three sensors on top of a stepping motor the robot will be able to scan back and forth to create a 2-D image of what is in front of it.
The motion of the robot will be controlled by two motors connected to wheels. Using this two motor scheme the robot will be able turn in place which gives it more maneuverability. It also makes the construction and control easier than some other structures.
We also intend to control the robot from the computer using a wireless RS-232 interface that we've seen before. This will give us the ability to upload images of what the robot sees and send the robot basic commands.