A simulation environment used to train a network to predict a trajectory given an image of a scene. Includes generating maze simulations where a robot moves through a particular environment, or "maze". Robot dynamics are based on proximity to objects, generating a reasonable trajectory as it moves across the course. These are then used to train a model on how a robot would normally move, so that at inference time we can provide just the image of the environment and receive a trajectory prediction.
-
Notifications
You must be signed in to change notification settings - Fork 0
Predict a robot's trajectory in a simulated environment using only an image of that space, part of some 2020 DPhil research.
License
roberto-aldera/speed-from-image
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
About
Predict a robot's trajectory in a simulated environment using only an image of that space, part of some 2020 DPhil research.