Skip to content

zoespot/CarND-Unscented-Kalman-Filter-Project

 
 

Repository files navigation

Unscented Kalman Filter Project Starter Code

Self-Driving Car Engineer Nanodegree Program

Self-Driving Car Engineer Nanodegree Program -Term2 Project2

My code built without error with cmake .. && make, and connected to GUI with ./UnscentedKF.

With fusion LASER and RADAR data, the px,py,vx,vy RMSE = [.0760,.0832,.3169,.2914], less than the requirement [0.09,0.10,0.40,0.30]. Parameters noise std_a(lonitudinal acceleration noise standard deviation) = 2 and std_yawdd(yaw acceleration noise standard deviation) =0.3.

Fusion with std_a= 2, std_yawdd =0.3

Its NIS (nomarlized innvoation square) are

Fusion with std_a= 2, std_yawdd =0.3

I also tried smaller process noise std_a=0.2 and std_yawdd=0.2. NeitherNIS nor RMSE improves much. So I kept std_a=2 and std_yawdd=0.3 as default parameters choices.

Fusion with std_a= 0.2, std_yawdd =0.2

Its NIS (nomarlized innvoation square) is

Fusion with std_a= 0.2, std_yawdd =0.2

About

Unscented Kalman Filter Project for Self-Driving Car Nanodegree Program

Topics

Resources

License

Stars

0 stars

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages

  • C++ 98.6%
  • C 1.2%
  • Other 0.2%