CL-IC-001: Intelligent Connected Vehicle Algorithm and Simulation Platform　

I. Training Platform Introduction

The Decision Planning Simulation Training Platform is based on traffic scenario simulation software. It constructs predetermined traffic environments within a computer and carries out intelligent vehicle behavior decision-making and route planning simulations within the provided software environment. Students can design and conduct simulation experiments themselves, aimed at learning core concepts such as intelligent vehicle behavior decision-making, constraints, and route planning. At the same time, through the Decision Planning Simulation Training Platform, students become familiar with the composition, operation, and application of autonomous driving systems. The linkage function of the training platform connects sensor training platforms, by-wire chassis training platforms, intelligent connected training vehicles, and intelligent connected passenger car development platforms, achieving richer teaching, demonstration, and experimental functions. The Decision Planning Simulation Training Platform, intelligent connected training vehicle, and intelligent connected passenger car development platform use the same set of autonomous driving solutions. After teaching through the training platform, students can quickly get started with the practical application of autonomous driving systems.

II. Training Platform Configuration

Operation Desk

o Dimensions: 12001500630mm, sheet metal casing, includes a teaching panel for equipment principle instruction, and double drawers for storing teaching accessories.

1. Monitors

o Two 23.8-inch high-definition monitors, resolution 1920x1080, 99% sRGB color gamut.

2. Computer Platform

o CPU: Not less than 6 cores and 12 threads, base frequency not less than 2.9G, L3 cache not less than 12M.

o GPU: Base frequency 1550MHz, CUDA® core count 2304, memory type not less than 8G DDR6.

o Memory: Not less than 16GB LPDDR4x 2666MHz.

o Storage: Solid-state drive, 250GB.

I. Equipment Functionality

1. Provides Virtual Simulation Software, which includes pre-built traffic scenario environments. Users can also build traffic environments for simulation experiments using scenario construction software. It is possible to simulate any weather conditions such as sunny, rainy, or snowy days, under different lighting conditions for testing; generate random or preset traffic flows, and output various sensor data.

2. Offers an Autonomous Driving System consistent with the training vehicles. The autonomous driving system includes modules for perception, localization, planning, and control, which can operate continuously and be demonstrated; control of the simulated vehicle in the simulation software can be achieved through point cloud maps or high-precision maps.

3. Provides Map Making Tools, enabling control of the simulation vehicle in the virtual simulation system to drive in simulated scenes. Point cloud maps can be created using the simulation system's LiDAR data; high-precision maps can be made based on constructed scenes.

4. Offers Various Experimental Control Models, allowing modification or custom design of control models for various experiments.

5. Relies on Equipment for Interlinking with Other Platforms. In linked state, the following functions can be achieved:

1) Interlinking with By-Wire Chassis Training Platform

a. In autonomous driving mode, the Decision Planning Simulation Training Platform can control the by-wire chassis platform to execute driving, steering, and braking actions. The by-wire chassis platform needs to synchronize with the simulated vehicle's actions in the simulation platform, allowing direct observation of the by-wire chassis's driving state under autonomous system control.

b. In manual driving mode, the simulated vehicle in the simulation platform can be controlled by the by-wire chassis platform.

c. For comparison with real vehicle functions, manual operation of the by-wire chassis platform can enable the autonomous driving system to enter the autonomous driving state, taking over control of both the by-wire chassis platform and the simulation software's simulated vehicle.

2) Interlinking with Sensor Training Platform

a. Using various algorithm software in the sensor training platform to perceive and recognize various traffic participants in the virtual simulation scene, and conduct functional training in visual recognition, LiDAR recognition, and recognition fusion.

3) Interlinking with Intelligent Connected Training Vehicle

a. Sensor data from the virtual simulation system can be output to the intelligent connected training vehicle, testing the intelligent connected training vehicle's autonomous driving functions in an indoor environment, serving as a precursor course to the intelligent connected training vehicle's real vehicle testing.

b. The Decision Planning Simulation Training Platform can control the intelligent connected training vehicle's by-wire chassis, understanding its working principles and by-wire protocols, serving as a precursor course to the intelligent connected training vehicle's real vehicle testing.

II. Training Projects

1. Establishing Simulation Projects within the Virtual Simulation Software for Training

2. Building Simulation Scenarios for Training

3. Operation and Control of Simulated Vehicles in the Autonomous Driving System for Training

4. Autonomous Driving Training with High-Precision Maps

5. Various Autonomous Driving Function Simulation Testing for Training

6. Point Cloud Map Creation and Simulated Vehicle Tracking for Training

7. Joint Working Training of Intelligent Connected Vehicle's By-Wire Chassis and Autonomous Driving System