Now, run your open loop controller on a Unicycle with both process & sensor noise. Your open$-$loop control sequence probably does not drive

the system to the goal anymore. This is to be expected! It's like trying to park a car in a parking spot with your eyes closed. We will implement a

much smarter strategy next.

$[]$ np$.$random. seed $(0)$

max$\_$num$\_$steps $=200$

env $=$ gymnasium. make$("$gym$\_$neu$\_$racing$/$NEUEmptyWorld$-$vo$")$

# Add both sensor noise & process noise to make the problem more challenging

env $=$ NoisyStateFeedbackWrapper$($env$)$

env. unwrapped. motion$\_$model $=$ NoisyUnicycle$()$

obs, $\_=$ env.reset$()$ # initial observation of the system's state vector

observation$\_$history $=[$obs$]$

# query your open loop controller based on the initial state of the system

control$\_$sequence $=$ open$\_$loop$\_$control$\_$policy$($obs$,$ env.goal, num$\_$steps$=$max$\_$num$\_$steps$)$

# Now that we've computed the control sequence, run the system forward until

# we reach max$\_$num$\_$steps or the system's state is close to the goal

for step in range$($max$\_$num$\_$steps$)$:

action $=$ control$\_$sequence $[$step$]$

obs, $,\text{'}$ terminated,, $,\_=$ env.step$($action$)$

observation$\_$history.append$($obs$)$

if terminated:

break

# Generate a simple plot of the path taken

plot$\_$path$($env$,$ observation$\_$history$)$

if terminated:

print$(\text{'}$Success$!$ Your open$-$loop controller drove the system to the goal.$\text{'})$

else:

print$(\text{'}$Your open$-$loop controller did not successfully drive the system to the goal.$\text{'})$