autowarefoundation / autoware.universe

https://autowarefoundation.github.io/autoware.universe/
Apache License 2.0
1.01k stars 652 forks source link

Test Autoware.Universe Modules in a University Campus Environment #8114

Open ismetatabay opened 4 months ago

ismetatabay commented 4 months ago

Checklist

Description

To observe the current capabilities of Autoware, we plan to test it in real-world conditions using the Yildiz Technical University Campus environment and the Leo Drive's Autonomous Robo-Taxi vehicle.

Yildiz Technical University (YTU) Campus Environment

Yıldız Technical University (YTU) is located in Istanbul, Turkey, and encompasses multiple campuses. One of these campuses is the Davutpaşa Campus, where we operate autonomous vehicles.

Here is some general information about the YTU Davutpaşa Campus:

1. Slopes and Terrains The YTU Davutpaşa Campus has a varied topography. Some areas are flat, while others have gentle or steep slopes. These features can pose significant challenges for perception, planning, and control modules. 2. Greenery and Trees The YTU Davutpaşa Campus is typically landscaped with various plants and trees, providing aesthetic appeal, shade, and contributing to the overall environment. These plants and trees can be beneficial for testing perception and planning behaviors. 3. Building structures The Davutpaşa Campus features a diverse range of buildings, varying in size from small-scale structures to larger ones. These buildings are suitable for localization using LiDAR-based methods (NDT).


Map Files: PCD and Lanelet2 map Link


Leo Drive's Autonomous Test Vehicle

To perform autonomous driving tests in the YTU Campus Environment, the Leo Drive Robo-Taxi vehicle will be used during these tests.

This vehicle includes six LiDAR sensors and one GNSS/INS device for autonomous driving stacks. The internal IMU of the GNSS/INS device is used as the IMU sensor.

The LiDAR sensors consist of:

The eight Lucid Vision cameras are used for camera-LiDAR fusion in the perception pipeline, and the SBG Ellipse D is utilized as the GNSS/INS device for the sensing and localization pipeline.

ECU of Test Vehicle

The ECU specs of our test vehicle are as follows:

Complement Product
CPU AMD Ryzen Threadripper PRO 3975WX 32-core, 64-thread
Memory 256 GB RAM
GPU 3x NVIDIA RTX A4000 (TensorRT YOLOX doesn't support multi-GPU usage, we plan to add it.)

Purpose

The main purpose of this task is to observe and assess Autoware's capabilities in real-life usage and address issues identified during tests. As mentioned above, the YTU test environment provides challenging conditions for autonomous driving vehicles, such as trees, slopes, and bumpy and rough road conditions.

Possible approaches

Definition of done

stale[bot] commented 1 month ago

This pull request has been automatically marked as stale because it has not had recent activity.