Lidar Tracking using $2$D Kalman Filter

You are to build the $20$Kalman fites, i$.$e$.,$ to track vahiclo uing a $20$ lidaz. Use the motion model as

$\frac{d}{dt}{P}_x=v_x$

$\frac{d}{dt}{P}_y=v_y$

And express A$,$ B$,$C$,$ D$-0$ stzte space matrices for the state

In continuous time. Then, discretiae at $t=0\mathrm{.}1$ sec$.$ Aed obtain F$,$ G$,$ H matrices of of state space model. implemant

the model based perediction and lidar measurement update as ghown.

Asaume the varlances in $x$ and $y$ positions is ${0\mathrm{.}1}^2{m}^2$ and wiocities $2^2(\frac{m}{s}{)}^2.$ Lidar sencor has the variance of ${0\mathrm{.}2}^3{m}^2.$

Simulated the estimation for $10$ minuas and display the pesition similar to below fie.

Opendata Set Usage for Lidar Data

Using the basic

loop.py script provided in midterm illustrate the followings for each frame in the sequence file used

Print the at of whides

Display the froet viere camera

Print the

VFON

b$.$ the range image siav.

c$.$ pitch resolution.

d$.$ Max$/$min rangos

Visualise the rango imaper' range and intensity channels seporately

Convert range imane to point doud

Crop the point doud as sugested by complex Yolo paper SOmaSOmaten volume

Oecain eEv of the cropped point closd to focus object detoction and tracking effort

Print the Min$/$max intensty, tceal number of vehicles and difficult to track of all trames

Primt and display labolled $($ground theth$)$ bounding bows of the cars on top of the BEV image

Using the precomputad detections given in the results directory, overlay the prediction with red bow.

Compute precisions and recalis

After iterating over al of the frames, compute owerall precision$-$secall and plot it

Debug the errors and unaxpected results to male sure the functions working as intended.