marinbao/carla-gui

GUI tool for creating simulation scenarios with Carla simulator

Carla-GUI

GUI tool to help human-vehicle interaction researchers design and conduct traffic experiments with CARLA real-time driving simulator.

Instructor

• Prof. Paul A. Green

Team Members

Summer Team (created Version 0.00 in summer 2020)

• Grant Barry
• Shiji Liu
• Weixin Feng(Targy)

Fall Team

• Zhiyang Chen
• [Brandon Sapp]
• [Raeed Rasul]
• [Yi Wu]
• [Yifei He]
• Nihar Joshi
• Jihyeong Ko

Project Purpose

The CARLA simulator is a powerful open source tool developed to "support development, training, and validation of autonomous driving systems" (introduction, CARLA Simulator). To properly utilize the powerful functionality provided by the CARLA simulator, however, users need to have:

• an in depth understanding about the function provided by the simulator
• basic knowledge about vehicle control
• basic knowledge about python

Meeting all these requirements is definitely not easy for begining researchers in the field of human-vehicle interaction. To make life easier for those researchers, a simple GUI has been developed, which provides an interface for users to create and conduct experiments using the CARLA simulator.

Function introduction

Our simulator allows users to create both urban and freeway simulation scenarios.

The start page of the gui, with choices for editing both urban(intersection) and freeway simulation scenarios.

Urban simulation

The GUI for urban simulation is still under development, documents will soon be available.

Here is a demo of back end function for the Urban GUI:

add vehicles inside the urban simulation environment

simulation starts inside the CARLA enviroment

For more information about urban simulation scenario, see urban back end

To view the full process of creating and conducting an urban simulation, please refer to urban demo video

Freeway simulation

The GUI for freeway simulation is still under development, documents will soon be available.

Here is a demo of back end function for the Freeway GUI:

add vehicles inside the freeway simulation environment

demo of freeway simulation

For more information about urban simulation scenario, see freeway back end

To view the full process of creating and conducting an urban simulation, please refer to freeway demo video

Installation

• step 1: download this repository and put it to your desired location on your device (currently support windows and linux)
• step 2: follow instructions on PyQt5 website to install PyQt5, which is the tool we used to create the GUI. Note: if you have PyQt5 pre-installed, make sure the version is >= 5.15.0
• step 3: follow instructions on back end installation guide to install tools essential for the back end.

Usage

• open a Python terminal inside the gui_interface folder, enter

    python gui_test.py
  

• Some reminders: (1) In Freeway mode, remember entering any section at least once, otherwise FreeWayenv might fail to be detected. (2) In Intersection mode, remember to enter the first intersection and spawn ego vehicle. Otherwose some errors will happen.

Data Collection

For now, data collection feature can be enabled/disabled by changing boolean values 'RECORD_ENABLE' in intersection_backend.py and 'self.RECORD_ENABLE' in section_environment.py

Note for now data_collection feature might make carla slower. Some optimizations need to be made soon.

Tutorial video

• urban tutorial video
• freeway tutorial video

Poster

The poster briefly describing the project

-- SIM team