autoVSimulator

This is a simulator for autonomous driving.

To run the demo,

python visualize_ngsim.py
python first_person_view.py
python flash_view.py

To start in Windows:

Install 32-bit Anaconda2 and set its build-in Python as the default Python.

Install Panda3D and set the Anaconda Python as the connected Python: (1)Uncheck the python2.7 during installing (2)Simply create a "panda.pth" file inside your copy of Python, containing the path of the panda directory and the bin directory within it on separate lines (for example C:\Program Files\Panda3D-1.2.3 and C:\Program Files\Panda3D-1.2.3\bin).

Install numpy(using command): pip install numpy

Install cvxopt(using command): conda install -c https://conda.anaconda.org/omnia cvxopt

Visualization of intention prediction

Based on the simulator developed by Jianyu Chen and Changliu Liu in Berkeley MSC Lab in 2016.

Multicar

This "multicar" branch is an application built on Auto Vehicle Simulator, which demonstrates the Multi-car Convex Feasible Set algorithm. There are 2 built-in scenarios for display, a 2-car overtake and a crowded 9-car plot.

Environment Setup with Anaconda

conda create -n avsim python=3.6
conda activate avsim
pip install numpy scipy matplotlib pandas sympy nose
pip install panda3d==1.10.3
conda install -c https://conda.anaconda.org/omnia cvxopt
pip install cvxpy

How to Run the Demo

1. Planning mode Configure self.scenario in main.py for different scenarios, leave self.replayFile as None for this mode. Use keyboard "c" to triger lane changing.
2. Replay mode Configure self.replayFile to a trajectories log file. The simulator will use the saved trajectories without live planning.

Scenarios

1. 2-car overtake

logged filename: "traj_log_2.npz"

2. crowded 9-car

logged filename: "traj_log_9.npz"

References

Huang, J., and Liu, C., 2020 "Multi-car Convex Feasible Set Algorithm in Trajectory Planning"