contact: Maxime Bouton, boutonm@stanford.edu
This repository consists of different driving scenarios formulated as POMDPs. It provides a generative model for computing policies. A few of them have explicit transition and observation models.
To install this package, first add the SISL registry and the JuliaPOMDP registry such that all dependencies are automatically installed. You can run the following in the julia REPL:
using Pkg
Pkg.add("POMDPs")
using POMDPs
POMDPs.add_registry()
Pkg.Registry.add(RegistrySpec(url="https://github.com/sisl/Registry"))
Pkg.add("AutomotivePOMDPs")
Run docs/Urban POMDP tutoial.ipynb
for a visualization of the simulation environment.
This package exports the following POMDP Models:
SingleOCPOMDP
: Occluded crosswalk with one single pedestrian. Discrete states and observations with explicit transition and observation model.SingleOIPOMDP
: Occluded intersection with one single car. Discrete states and observations with explicit transition and observation model.OCPOMDP
: Occluded crosswalk with a flow of pedestrian. Generative model implementation with continuous state and observationsOIPOMDP
: Occluded T intersection with a flow of cars driving in multiple lanes. Generative model implementation with continuous state and observationsUrbanPOMDP
: Occluded T intersection with crosswalks, flow of cars and pedestrians. Generative model implementation with continuous state and observationsThese models are defined according to the POMDPs.jl interface. To see how they are parameterized, toggle the documentation using ?
or
use the function fieldnames
if documentation is not yet written.
Video of the UrbanPOMDP
scenario:
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。