Forward and Reverse KL Divergence

[OscarRhoades]

p is the distribution you are calculating from real samples, q is the distribution from the particle filter, i is the grid cell number -- so you are calculating the divergence between all the grid cells and summing. Ignore the integral, in this case, it is a sum.

[vpham1897]

x is a region

p(x) given a specific region, what's the odds that the searched car is in that grid.

• Sum all cars in that grid that meet the criteria and divide that by all the cars of that criteria
• I'll organize it so that grids only have cars of a certain criteria

q(x) is a variable that's gotten from Oscar