Accuracy test incase of robot kidnap problem

[mohdwaseem27]

I conducted same accuracy test for a both particle filters (AMCL and optimal) in a robot kidnap scenario by manually relocating the robot, then moving it linearly for 5 meters without explicit localizing manually (neither moving around nor giving estimate from 2D pose estimate from rviz), and subsequently computing the Root Mean Square Error (RMSE) between the ground truth pose and the filter pose

In case of optimal filter, error was around 1.5m (150cm)