From Maslab 2011
|(4 intermediate revisions not shown)|
Latest revision as of 13:49, 5 January 2011
tl;dr: Team 13 is operating at half-power for the first week; Kevin and Bayley will join us starting next Monday. Dima and Rebecca are working extra hard to meet the Checkpoints and be ready for Mock Contest 1.
Current progress: Absentees aside, we nonetheless completed Checkpoint 1 by the deadline. Pegbot is now capable of 1) moving in a circle or in a straight path and 2) communicating with eeePC via Bot Client.
Future direction: Strategy discussion. Rebecca will study what former groups have done, what has worked, what hasn't, etc. Our choice of strategy will likely dictate (and in turn, also be dictated by) the hardware details of our robot. Team 13 would like to receive input from all members, so efforts are being made to hold a real-time discussion with Kevin and Bayley. Dima hopes to have finished our robot by the end of the week.
tl;dr: Successful analog input (specifically, a test voltage) reading from the orcboard. However, IRRangeFinder still eludes us. The value for range hovers around an apparently arbitrary ~0.12 for no good reason.
Current progress: Our robot is going to operate on a basic principle that it picks up as many balls as it can find via the use of a roller, elevates them a small distance by means of an Archimedes screw (inspired by Team 3 from 2003) and store them in a hopper where they will all be docked at the first goal/yellow-wall that the robot finds.
The software architecture consists of several states - explore/random motion, wall follow, scan surroundings, and three separate action states to handle balls/goals/yellow walls - that will be controlled by timers. If the robot has spent a certain amount of time on one task unsuccessfully, it may be stuck and the timers are intended to protect against that. In addition, we plan on coding very robust explore/random and wall follow programs, relying on the camera vision coding heavily only when the robot stops intermittently to scan its surroundings. Having located something of interest, it then needs to drive in a straight line, which we can hopefully achieve by specifying velocity rather than WPM for the motors/wheels.
As far as Checkpoint 3 is concerned, after a lot of confusion with ADC ports, it transpires that we have been specifying the wrong port, and hence have received no meaningful feedback from the AnalogInputs instantiated. Now we are getting a voltage reading that does respond accurately when we hook it up to a test circuit.
Future direction: Kevin and Bayley are still AWOL. Team 13 does not panic yet, however (although Team 13 is beginning to worry/wonder where the rest of the team is). Rebecca is going to spend tomorrow drafting a code for finding and approaching walls. Dima plans to finish CADing and ordering parts by tomorrow night, at the latest. Team 13 will take the pegbot into lab tomorrow afternoon and maybe get some help with the IR sensors. Probably will also set up and start working on the webcam and vision coding tomorrow night, after the Checkpoint.