Closed BenedicteLC closed 10 years ago
You are free to choose any conditional probability distribution (CPD), as long as you achieve the goal of this assignment, which is to implement a particle filter and get it (the particle cloud and the resulting pose estimate for the robot) to track the real robot. If you believe that l2 norms will work, then by all means use them.
Make sure to provide a short write-up to explain your design and implementation choices, and provide screenshot(s) to demonstrate that your solution works.
Hint: a CPD must output a probability, thus your function must satisfy all the conditions for a measure of probability.
I'm wondering how to compute the probability of z given x_i. Can we use the l2 norm of the distance between the real scan and the simulated scan for each particle in our calculations?