imitation-learning
imitation-learning copied to clipboard
question about datasets
hi, I found that the value of the throttle in the dataset is basically 0.5 or 1, but the speed is basically uniform. Why? Assuming the throttle is always 0.5, the speed should always increase. I'm so confused. Looking forward to your reply~
@felipecode, I have the same question here. Would it be possible to give some information in this regard?
Throttle 0.5 is like a minimum to beat the inertia in carla physics.
@felipecode, I was running some experiments and lower throttle values of around 0.2 already beat the inertia.
Thanks for the question BTW. @paulaurel Yeah, but it goes very very slow right ? On 0.5 it like already input some constant speed. Inertia was not the right term. I haven't analyzed that properly to be fair. But there should be sometimes were the throttle is zero right ?
@felipecode yes it is relatively slow. It reaches approximately 14 km/h. Do you have a model to map the throttle values to corresponding acceleration? If not is there a model of the Carla's car drivetrain? This would be great.
@felipecode Thanks for clarification.
Why is there negative numbers of speeds in the data set?
speed command steer throttle brake
count 657600.000000 657600.000000 657600.000000 657600.000000 657600.000000 mean 17.935953 3.356726 0.001634 0.504136 0.174912 std 14.463256 1.173195 0.155913 0.330658 0.379184 min -18.739027 0.000000 -1.154329 0.000000 0.000000 25% 10.102288 2.000000 -0.025604 0.500000 0.000000 50% 18.982355 3.000000 -0.000035 0.500000 0.000000 75% 22.477591 4.000000 0.029197 0.531509 0.000000 max 82.729424 5.000000 1.198919 1.000000 1.000000