April 15, 2012 at 7:32 p.m.
May 1, 2011
Computer Science, Electrical Engineering
University of California, Berkeley
In this project, our goal is to design a configurable robot. A configurable robot is a robot that does not run a fixed code, and it can be configured by simply giving it a new command in English text. Therefore, there won’t be any need to change the code. Instead, updating the code will be done automatically.
In order to achieve this goal, we need to provide two sets of information to the robot controller: The first set is the input data, which consists of any sensor data that will be used. The next set
of information is the English command, or in other words, the desired behavior we will be expecting from the robot.
According to Kress-Gazit's work, this input can be categorized to four different groups:
1) Coverage: Go to room A, B, C, D in any order
2) Sequencing: First go to room C then go to room A
3) Conditions: If you sense left distance sensor, go to room A otherwise stay here.
4) Avoidance: Don't go to room B
We have collected our data through an optical tracking system called OptiTrack: Tracking Tools that uses passive markers to detect objects.
Login and complete your profile for commenting