icra2018/adversarial-testing repository overview

⁠adversarial_testing

This is a python package for testing controllers for black box system in simulators

⁠Installation

To install,

python3 setup.py install --user

This should allow you to use the package anywhere in your current environment

⁠Tests

The tests folder has 3 files:

• test_sincos.py : This file shows the difference between modeling smooth and non-smooth functions using GPy
• test_car.py: This file implements a simple linear controller on a car for obstacle avoidance. This file shows how we use KernelPCA for reducing the input space.
• test_cartpole.py: This file tests a nearest neighbor controllder code submitted by user for the Open AI Gym environment Cartpole-v0

⁠How to Run with Docker

Tested on:

• Ubuntu 16.04.6 with Docker 18.06.1-ce⁠
• macOS Mojave 10.14.3 with Docker Desktop for Mac 2.0.0.3 (engine: 18.09.2)⁠
• Windows 10 Education with Docker Desktop for Windows 2.0.0.3 (engine: 18.09.2)⁠

1. Open a terminal in Linux/macOS or a PowerShell in Windows and run the command:

docker run --rm -p 8888:8888 icra2018/adversarial-testing:latest

1. Run a web browser and open the link: http://localhost:8888/lab/tree/README.ipynb⁠