– Such filter is such a general and powerful tool for combining information in the presence of uncertainty.
– You can use a Kalman filter in any place where you have uncertain information about some dynamic system, and you can make an educated guess about what the system is going to do next.
– And it can take advantage of correlations between crazy phenomena that you maybe wouldn’t have thought to exploit!
– Kalman filters are ideal for systems which are continuously changing. They have the advantage that they are light on memory (they don’t need to keep any history other than the previous state), and they are very fast, making them well suited for real time problems and embedded systems.
– We’ll say our robot has a state xk (vector), which is just a position and a velocity:
xk→=(p, v )
– – Note that the state is just a list of numbers about the underlying configuration of your system; it could be anything. In our example it’s position and velocity, but it could be data about the amount of fluid in a tank, the temperature of a car engine, the position of a user’s finger on a touchpad, or any number of things you need to keep track of.
– But of course it doesn’t know everything about its motion: It might be buffeted by the wind, the wheels might slip a little bit, or roll over bumpy terrain; so the amount the wheels have turned might not exactly represent how far the robot has actually traveled, and the prediction won’t be perfect.
– The GPS sensor tells us something about the state, but only indirectly, and with some uncertainty or inaccuracy. Our prediction tells us something about how the robot is moving, but only indirectly, and with some uncertainty or inaccuracy.
– So, here we are trying to model inefficiencies in the robot’s motors to actually move on the terrain based on the GPS given data only. Here, the uncertainty of the env must be taken into account and modeled as closely as we can.
– Let’s look at the landscape we’re trying to interpret. We’ll continue with a simple state having only position and velocity. x(column vector) =[p, v]
– Kalman filters assume the parameters are Gaussian distributed. So, each variable has an associated mean and variance.
– Mostly the parameters maybe uncorrelated, but sometimes they might have meaningful correlations from which we can reduce the uncertainty we have. For eg. If velocity was high, our position would be more distant. This information is contained in the covariance matrix.
– So we need our best estimate, covariance matrix at each time “k”.
– Now, we have our values at time “k-1” and want to predict for time “k”
– We need to have a prediction matrix as well as way to update the covariance matrix
–