We will process the camera images to obtain the current angle of the pole. We will use this information in a feedback system to move the cart under the pendulum and stop its motion. Using modern control theory, we will be able to do this in the most optimum way, minimizing undesirable behavior and stabilizing it as quickly as possible. If we can obtain some estimates of the amount of process noise (wind and such) and the amount of measurement noise (how inaccurate our sensors are), we could use Kalman filtering to optimally stabilize it in the presence of this noise.
Sadly enough, many of the papers we have read for the class are completely irrelevant to this robot.
It performs no mapping — so it will not use Monte Carlo Localization [Dellaert], nor other Markov-based probabilist navigation models [Simmons], nor DERVISH's relatively simpler model of the world [Nourbakhsh]. Furthermore, it will not be navigating in any significant sense, so [Roy]'s coastal navigation techniques will not assist us.
While it would be interesting to attempt to use evidence grids [Martin] in an attempt to determine where the pendulum is, this unorthodox solution would probably be more trouble than it's worth.
Our pendulum-balancing robot will be so simple as to essentially have one mode of behavior — balancing the pendulum — which makes subsumption architecture [Brooks] irrelevant.
While Polly [Horswill] accomplished many things, balancing a pendulum was not among them (nor will our robot give tours).
Applying Bayesian techniques might be an interesting exercise for a future project, but [Leonhardt]'s article does not help us much now.
We're having enough trouble getting a solitary pendulum-balancing robot to work; it's unlikely that we could add multiple robots that could work in concert. As such, [Navarro-Serment]'s Millibots and [Yim]'s PolyBot will not help us.
As our robot will not be humanoid, we will be unable to apply the exemplar-based learning approaches seen in [Drumwright].
While our robot differs substantially from [Vaughan]'s "sheepdog" in almost every respect, we too are using the trick of special coloring of objects to aid a vision system.
While failure recovery is always a good thing ([Toyama]), it is not likely that we will actually be able to implement such a thing in our robot. Once our pendulum falls below horizontal, it seems unlikely that any movement of the robot can right it.
There is a paper that will benefit us, however. As it did in the last project, [Mouravec]'s trials and tribulations with the Stanford Cart will remind us that, however bad things may get, they could get much worse.