Simulation program for reviewing self-driving

Simulation for verification of diverse driving environments and vehicle sensors

The driving simulation program was created to allow all data that can be gained from operating an actual self-driving vehicle in a virtual environment in a more efficient, real, and safe manner.

The virtual data provided is identical to the protocols of actual sensors and cars, and image processing, path planning, and other algorithms for self-driving can be verified virtually and applied quickly to actual vehicles.

The data is almost identical to actual physical data from LiDAR and cameras. In addition, custom scenarios allow for driving tests in various virtual environments with varying traffic situations, road conditions, weather, climate, nigh/day time, etc.

ERP42 Self-Driving & Obstacle Avoidance

ERP42 Self-driving Simulator for You

Self-driving simulation software

  • null

    Driving algorithm (Control), sensor interface (LiDAR, GPS), vision processing (Image Processing), deep learning (object detection) are available. As for HW platforms, the software is applicable to WeCAR (educational)/ERP42 (for competitions), and ERP42 MINI (for educational research).

  • null

    There is the academic version (Base, Full Version) and the research version (Professional, Developer Suite).

    • A virtual environment features modeling of vehicles, driving environments, sensors, etc. and allows self-driving car development in a virtual environment.
    • The simulator outputs vehicle data and sensor data, and takes speed and steering commands to move the car.
    • The output of the simulator is identical to actual sensor protocols (Velodyne, GPS nmea, Camera, etc.)
    • HILS fuction: simulation function includes hardware (vehicle, controller). Communication between self-driving software and the simulator can be performed via serial or UDP.
    • A virtual car model identical in size and shape to ERP-42 and with similar dynamics is provided. The same serial communication protocol is provided.
    • UDP provided for virtual car control (Protocol exclusive to the simulation, serial option is also available)
    • 250 m straight track provided, has object placement feature
    • Provides ground truth regarding objects and vehicle status

WeGo ERP-42 / ERP-42mini Simulator function / price comparison/detailed explanation of each item

연구/기업 Professional

Research / Corporate Professional

연구/기업 Standard

Research / Corporate Standard

학생 Full

Student Full

학생 Basic

Student Basic

Trial

FEATURES

설명

Description

O O O O O

Vehicle Model

Vehicle Model

실제 차량의 크기와 외형이 같은 3D 모델 및 실제차량과 통신 프로토콜이 동일한 가상의 차량 제공

A virtual vehicle with a 3D model identical to the size and appearance of the actual vehicle that uses the same communication protocol as the actual vehicle is provided

O O O O O

Ground Truth Data

Ground Truth Data

정답 데이터(주변 물체의 위치, 가상공간에서의 정확한 차량 위치 등)

Truth data (The location of surrounding objects, the exact location of the vehicle in the virtual space, etc.)

O O O O X

Sensor Models

Sensor Models

기본 센서 모델 (Camera, Lidar, GPS, ...)

Basic sensor models (Camera, Lidar, GPS, ...)

상용 센서와 유사한 규격을 따르는 센서 모델

Sensor models follow specifications similar to commercial sensors

Option X X X X

커스텀 센서 모델 (e.g. Velodyne, Ublox,...)

Custom sensor models (e.g. Velodyne, Ublox,...)

사용자가 요청한 실제 센서들의 모델링

Modeling of user-requested sensors

O O O O O

Maps

Maps

직진 트랙 (종방향 제어 테스트)

Straight track (longitudinal control test)

종방향 테스트가 가능한 가상 공간

Virtual space where longitudinal testing can be performed

O O O O X

원형 테스트 트랙(경로 추종, 장애물 회피)

Round test track (path tracking, obstacle avoidance)

주어진 경로 추종 및 장애물 회피 테스트가 가능한가상 공간

A virtual space where tracking of set paths and obstacle avoidance can be tested

O O O X X

K-City

K-City

경기도 화성의 K-City를 그대로 모사한 앱(가상의 GPS값과 실제 GPS값이 같다)

An app that accurately models K-City in Hwasung, Gyeonggi-do (virtual GPS values and actual GPS values are identical)

O O O X X

Sensor Calibration환경

Sensor Calibration Environment

센서의 성능 테스트 및 Calibration이 가능한 앱

App that allows sensor performance testing and calibration

Option X X X X

커스텀 트랙

Custom Tracks

사용자가 요청하는 트랙 제작 (실제 도로 구간도 제작 가능)

User-requested tracks are created (actual road sections can be created)

Option X X X X

기타기능

Other functions

Ghost 차량 배치 기능

Ghost car placement

가상의 환경의 데이터를 실제 차량에 데이터를 전송하여다양한 실험이 가능한 기능(상세 설명은 문의)

This functions allows various experiments to be performed via sending virtual environment data to actual vehicles (please send an inquiry for more information)

Option X X X X

다중 클라이언트 환경

Multi-client environment

하나의 트랙에 여러 클라이언트가 접속이 가능한 기능

Allows many clients to connect to a single track

Win10
Linux
Win10
Linux
Win10
Linux
Win10
Linux
Win10

시뮬레이터 실행 환경

Simulator execution environment

시뮬레이터를 실행하는 환경

Environment in which simulator is run

O O O X X

기술 지원

Technical support

시뮬레이터 기술지원

Simulator technical support

WeGo ERP-42 / ERP-42mini Simulator function / price comparison/detailed explanation of each item

Type : refers to sensor’s communication method

  • UDP : UDP communication method. IP/Port must be set
  • Serial : Serial communication method. COM Port Number must be input at Port Num

* Serial communication method is identical to actual ERP42 communication method

K-City (Hwasung self-driving test city) map composition applied effectively

Hardware pictured is applied in the WEGO-ERP42 standard platform

  • Rendering Quality – sets the simulator’s current rendering quality
  • Location Setting – Moves the vehicle according to each scenario.

Obstacle selection

20 or so obstacles can be added to the simulation as necessary for the test.
Obstacles placement mode can be turned on/off(obstacles can be selected with the mouse from a list)

LiDAR sensor

HardLike with Veloview, this is a LiDAR test demo using LabVIEW; it can be applied to Python, OpenCV, etc.
ware pictured is applied in the WEGO-ERP42 standard platform
Demo – LabVIEW Veloview

LabVIEW Lidar example

WeGo simulator

Type : refers to sensor’s communication method

  • UDP : UDP communication method. IP/Port must be set
  • Serial : Serial communication method. COM Port Number must be input at Port Num

Lane Detection applied

LabVIEW Vision module

Python OpenCV

Jupyter notebook

20 or so obstacles can be added to the simulation as necessary to test it.
Obstacles placement mode can be entered/exited (obstacles can be selected with the mouse from a list)