Participants will be asked to create a ROS2 package to carry out the scenario, and the following ROS2 package is provided in aichallenge2023-racing/docker/aichallenge/aichallenge_ws/src as sample code to serve as a base for this in this repository The following ROS2 package is provided.
- Please place the ROS2 package you created under aichallenge_ws/src/aichallenge_submit and follow the steps below to build and run it.
+ aichallenge_launch, sim_msgs change is not applied at submission
Sample Code
-
The following is a partial configuration under aichallenge2023-racing/docker/aichallenge/aichallenge_ws/src.
+
aichallenge2023-racing/docker/aichallenge/aichallenge_ws/src directory structure is below
aichallenge_launch
Contains the main launch file aichallenge.launch.xml. All ROS2 nodes are launched from this launch file.
-
aichallenge_scoring
-
-
aichallenge_scoring is responsible for the tasks required for scoring and ranking the AI Challenge participants.
-
-
-
aichallenge_scoring_msgs
-
-
msgs for scoring
-
-
-
aichallenge_scoring_result
-
-
Calculate scores from /aichallenge/score and /aichallenge/collision.
-
-
aichallenge_submit
All ROS2 packages implemented by participants should be placed in this directory, as only the contents of this directory will be submitted at the time of submission.
-
Please modify this launch file accordingly to configure the ROS2 nodes you have implemented to be launched. Please modify this launch file accordingly and configure it so that your ROS2 node will be activated.
-
initialpose_publisher
-
-
Since the default initial position estimation may result in initial position deviations, this function provides the ability to set the vehicle position at a pre-defined initial position.
-
-
-
pose_initializer_custom
-
-
The pose_initializer package in autoware.universe has been modified to discard the posture calculated by Monte-Carlo and NDT Matching and set it with the posture sent by the initialpose_publisher.
-
-
-
self_driving_controlloer
-
-
Provides the ability to set GoalPose and engage required to start automatic operation.
-
-
-
autoware_launch
+
Please modify this launch file accordingly to configure it to launch the ROS2 node you have implemented. Please modify this launch file accordingly to configure your ROS2 node to be launched.
+
autoware_launch, autoware_universe_launch
-
We have copied and partially edited the launch, config-related packages of Autoware; Autoware in the Docker image has removed the packages included here. You can modify the behavior of Autoware by editing here.
+
The Autoware packages included here have been removed from the Docker image of Autoware. You can modify the behavior of Autoware by editing here.
# In the aichallenge2023-racing/docker/evaluation directory
+
# In the aichallenge2023-racing/docker/evaluation directory
bash build_docker.sh
-
Launch AWSIM
-
Start the Docker container (After launch, autoware will start automatically, and autonomous driving will begin)
+
Launch AWSIM
+Start the Docker container (After launch, autoware will start automatically, and autonomous driving will begin)
-
#In the aichallenge2023-racing/docker/evaluation directory
+
GPU version
+
# In the aichallenge2023-racing/docker/evaluation directory
bash run_container.sh
-
-
Verify result.json
-Once the evaluation is complete, result.json will be stored in the output folder.
+
CPU version
+
# In the aichallenge2023-racing/docker/evaluation directory
+ bash run_container_cpu.sh
+
+
+
Verify result.json
+after evaluation、output folder has below
+
+
result.json
+
rosbag.db3
+
rviz_capture.mp4
+
autoware.log
+
+
Upload to the Online Environment
diff --git a/en/rule/index.html b/en/rule/index.html
index 37a2e0b..18f2478 100644
--- a/en/rule/index.html
+++ b/en/rule/index.html
@@ -80,61 +80,60 @@
Table of Contents
-
Rule
-
This page explains the rules and ranking system for the competition. Please note that the contents of this page may change during the tournament period.
+
Certainly! Here's the translation of the provided text in the same format:
+
Rule
+
This page explains the rules and ranking system of the competition. Please note that the contents of this page may change during the competition period.
Ranking System
-
The ranking of participants is determined based on the following two metrics:
+
Participants' rankings are determined based on the following two criteria:
Distance Points:
-
For participants who could not reach the goal within the time limit (10 minutes), their ranking is determined by the distance points (0-100) corresponding to the distance they could travel from the starting point within the time limit.
-
If the participant's vehicle completed a lap within the time limit, the distance points are 100.
-
In the case of disqualification, the distance points are 0.
-
If the distance points are the same, the person with less time added due to penalties will rank higher.
+
For participants who cannot reach the goal within the time limit (10 minutes), the ranking is determined by Distance Points (0-5700), based on the distance traveled from the start point within the time limit.
+
If the goal is reached within the time limit, the Distance Points are approximately 5700 points. (Details of the exact distance are omitted for brevity.)
Total Time:
-
For participants who reached the goal within the time limit, their ranking is determined by the total time, which is the sum of the time elapsed from start to finish and penalty time. The penalty time will be calculated based on violations during the race.
+
For participants who reach the goal within the time limit, the ranking is determined by the Total Time, which includes the time taken from start to goal plus any penalties.
-
Ranking Example
+
Example of Ranking
Distance Points
Time from Start to Goal
-
Time Added Due to Penalties
+
Time Addition due to Penalty
Total Time
Rank
-
100
+
5700
01:10
00:00
01:10
1
-
100
+
5700
01:30
00:10
01:40
2
-
100
+
5700
01:20
00:30
01:50
3
-
50
+
60
N/A
00:00
N/A
@@ -157,36 +156,98 @@
Ranking System
Penalty
-
Deviation from the course and collisions with other vehicles running alongside are considered violations. There are 3 types of penalty for violations:
+
Violations such as deviating from the course or colliding with other vehicles are penalized. The penalties for violations are:
-
Disqualification
-
10-second addition to the driving time
-
5-second addition to the driving time
+
End of run
+
Addition of 10 seconds to the running time
+
Addition of 5 seconds to the running time
-
The following lists each violation and the corresponding penalty.
-
Critical Violation (Disqualification)
+
Below are the violations and their corresponding penalties:
+
End of Run
-
Deviating from the course (for more than 2 seconds)
-
Deviating from the course (more than 5m from the course boundary)
-
Driving in the wrong direction on the course
+
Deviating more than 5m from the course boundary for more than 10 seconds
+
Being more than 100 seconds away from the course boundary
+
Completion of the run is considered when crossing the following goal point.
+Sample can be found at docker/aichallenge/aichallenge_ws/src/aichallenge_submit/autoware_micro/goal_pose_setter/config/default_goal_pose.param.yaml (Note: yaml file is not specified in the launch file)
+
Sevior Violation (Penalty: 10 seconds per occurrence):
+
Major Violation (Penalty: 10 seconds/occurrence):
-
Colliding with another vehicle (the collision state lasts for more than 2 seconds)
+
Collision with another vehicle that lasts more than 2 seconds
+
Deviating more than 5m from the course boundary
+
Deviating more than 2 seconds from the course boundary
-
Minor Fouls (Penalty: 5 seconds per occurrence):
+
Minor Violation (Penalty: 5 seconds/occurrence):
-
Deviating from the course (for less than 2 seconds AND less than 5m from the course boundary)
-
Colliding with another vehicle (the collision state is resolved within 2 seconds)
+
Deviating from the course (less than 2 seconds and less than 5m from the course boundary)
+
Collision with another vehicle (resolved within 2 seconds)
Note
-
In cases where another vehicle collides with the rear of your vehicle (e.g., being rear-ended by another vehicle), it is not considered the fault of your vehicle and no penalty will be applied.
+
In cases where another vehicle collides with the rear of your vehicle (e.g., rear-end collision by another vehicle), it is not considered your fault, and no penalty is applied.
+
Do not drive in the wrong direction.
+
Do not cut corners.
Submission
-
Participants are requested to upload the software they developed to the evaluation system via the submission page. Three simulations will be conducted for each upload, and distance points and total time will be calculated for each simulation. The best score out of the three simulation results will be reflected in the ranking.
-
How to check results
-
The details and results of the simulations are planned to be available for confirmation through a json file output by the evaluation system and videos of the simulations. Please wait for further announcements.
+
Participants are required to upload their developed software to the evaluation system via the submission page. Three simulations will be conducted for each upload, and distance points and total time calculations will be performed for each simulation. The highest score among the three simulation results will be reflected in the rankings.
+
How to Check Results
+
The score of the results will be sent to result.json.
+
Log Format for Results
+
The results will be output in the following format in ~/awsim-logs/result.json.
+
{
+ "rawLapTime": 72.77926,
+ "distanceScore": 457.0,
+ "lapTime": 302.779266,
+ "isLapCompleted": false,
+ "isTimeout": false,
+ "trackLimitsViolation": [
+ 19, # out of track less than 2 sec
+ 19, # out of track more than 2 sec
+ 2, # out of track less than 5m
+ 2, # out of track more than 5m
+ 0 # not used
+ ],
+ "collisionViolation": [
+ 0, # collision less than 2 sec
+ 0, #
+
+ collision more than 2 sec
+ 0, # not used
+ 0 # not used
+ ]
+}
+
+
Additionally, ~/awsim-logs/verbose_result.json will also be output in the following format.
CPU: Intel Corei7 (4 cores and 8 threads) or higher
+
GPU: NVIDIA Geforce RTX 1060 or higher
+
+
※"Please place the Autoware operation PC and the AWSIM operation PC within the same network. If they are properly configured, topic communication between PCs is possible without any additional settings. If, by any chance, topic communication is not possible, please disable the firewall or review its rules."
+
For Those Using CPU Only PCs
+
Minimum Hardware Requirements for Headless mode
+
"We may not be able to fully support the operational environment of PCs used in this competition. However, to encourage more people to participate in the contest and engage in active discussions, we have experimentally made it possible to participate with the following CPU-only configuration.
+
+
OS: Ubuntu 22.04
+
CPU: Intel Core i7-8650U (4 cores) with Intel HD Graphics (no NVIDIA GPUs)
+
Memory: 16 GB or more
+
Storage: SSD 16 GB or higher
+
+
+
other CPU only cases
+
+
+
CPU : AMD Ryzen 5 5600G with Radeon Graphics
+
+
Memory : 32GB
+
+
CPU AMD® Ryzen 7 pro 4750u with radeon graphics × 16
+
+
Memory 32.0 GiB
+
+
DELL XPS 13 9300
+
+
CPU: Intel(R) Core(TM) i7-1065G7
+
+
Memory: 16GB
+
-
Autoware and AWSIM PCs should be located in the same network.
-If they are located in the same network, topic communication between PCs is basically possible without any additional settings. In the unlikely event that topic communication is not possible, please disable the firewall or review the rules.
Environment Setup
AWSIM(Ubuntu)
-
Preparation
+
preparation
-
Install Nvidia drivers
+
(Skip this step if you are using a CPU only or already have NVIDIA Driver installed) Installing Nvidia drivers.
If the above method takes a long time or times out, you can use the following method.
-We have placed a tarball of the image at here. Please use the following command
-
docker load < autoware-universe-cuda_v1.tar.gz
+
Once confirmed, re-login
+
Check that the following display is visible in docker images
+
REPOSITORY TAG IMAGE ID CREATED SIZE
+ghcr.io/automotiveaichallenge/aichallenge2023-racing/autoware-universe-no-cuda latest 9601fc85f1bd 3 weeks ago 7.31GB
cd aichallenge2023-racing/docker/train
bash build_docker.sh
-
start rocker
+
Starting the Competition Docker Container
+
For GPU environments
bash run_container.sh
+
For CPU only environments
+
bash run_container_cpu.sh
+
-
Starting AWSIM in a Docker container
-
If you want to start AWSIM from a Docker container, please follow the steps below after installing a Docker image according to the Docker image preparation procedure (see below).
+
Certainly! Here is the translation of your text in the same format:
+
Starting AWSIM in a Docker Container
+
If you want to start AWSIM from a Docker container, follow the Docker image preparation steps (described previously) to install the Docker image, and then proceed with the following steps:
-
extract the executable file for the convention in aichallenge2023-racing/docker/aichallenge (Hereinafter, it is assumed to be located in aichallenge2023-racing/docer/aichallenge/AWSIM/AWSIM.x86_64)
-
launch the Docker container (please verify with docker container ls that container exists)
-Please confirm the existence of the following docker image using other terminal by using the docker image ls command. below is an example.
+
Extract the competition's AWSIM executable file into aichallenge2023-racing/docker/aichallenge (for the purposes of this guide, we will assume it's located at aichallenge2023-racing/docker/aichallenge/AWSIM/AWSIM.x86_64).
+
Start rocker
+Open a new terminal and confirm that an image like the one below exists with docker image ls.
-
aichallenge-train latest f5f05f758f55 2 weeks ago 14.9GB
+
REPOSITORY TAG IMAGE ID CREATED SIZE
+aichallenge-train latest 67a4d45d119d 16 minutes ago 7.37GB
-
Once confirmed, launch rocker with the following command.
+
Once confirmed, start rocker with the following commands.
+
For GPU environments
cd ./aichallenge2023-racing/docker/train
bash run_container.sh
-
Open a new terminal and confirm the existence of the docker container by using the docker container ls command. below is an example.
-
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
-fdbe7cb05782 1f3d763bc501 "/bin/bash" 15 minutes ago Up 15 minutes elegant_hellman
+
For CPU-only environments
+
cd ./aichallenge2023-racing/docker/train
+ bash run_container_cpu.sh
+
+
In the new terminal, confirm that docker exists as shown below with docker images.
+
REPOSITORY TAG IMAGE ID CREATED SIZE
+aichallenge-train latest 67a4d45d119d 16 minutes ago 7.37GB
-
execute the following in the container
+
Execute the following inside the container
cd /aichallenge
bash run_awsim.sh
Note
-
The messages for the topics that are being published and subscribed to from AWSIM are partially defined in /aichallenge/aichallenge_ws/src/sim-msgs. To handle these messages, please execute the following commands:
+
The messages published and subscribed by AWSIM are partially defined in /aichallenge/aichallenge_ws/src/sim-msgs. To handle these messages, please execute the following commands.
cd /aichallenge
bash build_autoware.sh
source /aichallenge/aichallenge_ws/install/setup.bash
-
AWSIM(Windows)
-
-
Download latest AWSIM_AIChallenge_Windows_v*.*.zip from GoogleDrive and unzip it.
-
double-click the file to start it
-Confirm that the following screen is displayed.
-
-
-
Copy map data (pcd, osm)
-
-
Map data is stored in AWSIM compressed files. Copy the osm and pcd files located in AWSIM_Data/StreamingAssets/kashiwanoha2023_integ to aichallenge2023-racing/docker/aichallenge/mapfile and arrange them so that the file structure is as follows:
+
Certainly! Here's the translation of the provided text in the same format:
+
Placement of Map Data OSM Work in Progress (WIP)
+
For this competition, we have allowed modifications to the lanelet2_map file. Please place it in /aichallenge2023-racing/docker/aichallenge/aichallenge_ws/src/aichallenge_submit/aichallenge_submit_launch/map/*, ensuring the file structure is as follows:
※ There are no constraints such as having to refer to the lanelet2_map for route planning (referring to a csv file to generate routes is also permissible).
+※ PCD files are included, but there is no need to submit them this time. Use them only if necessary for editing the map.
Autoware
-
Verify Autoware operation
-This section describes how to check the operation of Autoware using AWSIM.
+
Checking Autoware Operation
+Here we describe how to check Autoware's operation using AWSIM.
-
Start AWSIM.
-
Start Autoware.
+
Start AWSIM
+
Start Autoware
-
# In the Rocker container
-sudo ip link set multicast on lo
+
# Inside the Rocker container
cd /aichallenge
bash build_autoware.sh
-source aichallenge_ws/install/setup.bash
-ros2 launch autoware_launch e2e_simulator.launch.xml vehicle_model:=golfcart sensor_model:=awsim_sensor_kit map_path:=/aichallenge/mapfile
+bash run_autoware.sh
-
Confirm that the following screen (Rviz2) is displayed.
-
+
Ensure that a screen like Rviz2 is displayed as shown below
+
-
Confirm that self-position estimation is done. If it is not estimated correctly, select 2D Pose Estimate in the tab and drag the actual position of the vehicle.
-
+
Select 2D Goal Pose in the tab, specify the goal position by dragging. Confirm that the route is displayed as in the image and that Routing changes from UNSET to SET (it takes a little time after specifying)
+
-
Select the 2D Goal Pose in the Goal Pose tab and drag to specify the goal position. Confirm that the route is displayed and Routing changes from UNSET to SET as shown in the image (it takes a little time after you specify it).
-
+
-
Press the AUTO button in OperationMode and confirm that self driving is started.
-
+
+
Starting from the Second Time Onwards
+
+
Starting AWSIM
+
+
Start the container
+
+
cd ./aichallenge2023-racing/docker/train
+ bash run_container.sh or bash run_container_cpu.sh
+
+
+
Start AWSIM inside the container
+
+
cd /aichallenge
+ bash run_awsim.sh
+
+
Starting Autoware
+
+
Start the container
+
+
cd ./aichallenge2023-racing/docker/train
+ bash run_container.sh or bash run_container_cpu.sh
+
+
+
Build Autoware inside the container if necessary
+
cd /aichallenge
+ bash build_autoware.sh
+
+
+
Start Autoware inside the container
+
+
cd /aichallenge
+ bash run_autoware.sh
+
-
+
+
How to Record the Screen
+
The default rocker setup doesn't set the display driver to your own PC, so you need to add the following options:
+
#!/bin/bash
+rocker --device /dev/dri --x11 --user ... # For CPU version
+rocker --device /dev/dri --nvidia --x11 --user ... # For GPU version
+
These are the teams who participated in the preliminary round of this year's Japan Automotive AI Challenge 2023 Integration Competition.
-
-
-
-
-
Team Name
-
Photo
-
Message from the team
-
-
-
-
-
Team A
-
-
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Proin consectetur velit ac nisi condimentum fringilla. Sed ac rhoncus dui. In hendrerit sapien ut magna faucibus, et facilisis justo scelerisque. Cras ut egestas metus, sit amet faucibus nisl. Nulla non magna mattis, viverra eros at, porttitor tortor. Curabitur rutrum lacinia vehicula. Vestibulum rutrum orci a arcu iaculis venenatis. Nulla ornare tellus nec eleifend venenatis.
-
-
-
Team B
-
-
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Proin consectetur velit ac nisi condimentum fringilla. Sed ac rhoncus dui. In hendrerit sapien ut magna faucibus, et facilisis justo scelerisque. Cras ut egestas metus, sit amet faucibus nisl. Nulla non magna mattis, viverra eros at, porttitor tortor. Curabitur rutrum lacinia vehicula. Vestibulum rutrum orci a arcu iaculis venenatis. Nulla ornare tellus nec eleifend venenatis.
