[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
步兵training spl
2019-1-30 11:53:49 显示全部楼层

马上注册,玩转Robomaster!

您需要 登录 才可以下载或查看,没有帐号?立即注册

x
本帖最后由 training spl 于 2019-1-31 15:09 编辑

NEWS: Online Text Tutorial Session For ICRA AI Challenge Is Coming This Thursday!

Beijing Time (UTC+8) 10 am 31st Jan 2019. Charles Yang from RoboMaster R&D will bring you to live text session about: Tips to prepare the ICRA challenge. Please make yourself convenient to watch the live text in this post and ask questions during the session.

If you have any questions regards to how to better develop robots, feel free to leave your question in below, the host will organize those questions and answer it correspondently during the live session.

Agenda:
10:00 - 10:10 am - General discussion about technical proposals of teams submitted.
10:10 - 10:20 am - Tips to prepare the ICRA challenge.
10:20 - 11:00 am - Real-time question answering by RoboMaster AI team.

Host: @Training Spl

Guest Speaker:

- Chief Engineer of AI Challenge - Charles Yang - Forum ID: charles.yang

- Engineer of AI Challenge - Vic Wen - Forum ID: vvwen

- Engineer of AI Challenge - Kevin Li - Forum ID: KevinLAD

During the live session, the host and the guest speaker will only type English letters, regardless of which language to describe questions are asked.

---------------------------------------------------------------------------------------------------------

Contents for the live session:

After receiving all teams’ proposal, to sum up, we find most of the teams already had a good command of the robot system design for the game scenario. However, there are still some general problems existed that may cause a negative effect of preparing the competition:

1.        Here are a few basic modules you must implement in the game:
a)        Localization
b)        Motion planning and control
c)        Visual (armor detection) gimbal servo control with projectile trajectory prediction
d)        Projectile supply (from navigation to visual servo control)
e)        Decision-making system

Besides, enemy robot detection, game field segmentation can also be applied to the robot.

2.        Pay much more attention to the NAVIGATION system for localization and obstacle avoidance motion planning. It is the basic element of your robot. Only to know how to run open sourced package cannot guarantee the performance of your robot. The penalty of collision is strict without any concession.

3.        Be sure your solution is PRACTICAL and FEASIBLE with more engineering consideration.
For example,
a)        if you use model-based filtering algorithm, please build your sensor model accurately enough or your algorithm robust enough in the game scenario, especially for UWB and IMU.
b)        If you use learning based algorithm, take care of the computing source for onboard PC so that all algorithms can process in acceptable and steady frequencies at the same time.
c)        If your algorithm highly depends on visual features. Think more about the light condition and possible accidents for enemy robots in the game scenario.
d)        If you want to set up a wireless link for communication between robots, take the transmission quality and delay into consideration.

4.        A stable system is a time-consuming work in the competition preparation period. It should take more than half of your preparation period to TEST instead of DESIGN. As the whole system is a combination of different modules, the buckets effect will have an obvious influence on the system performance, especially for visual servo control. Accurately targeting and hitting the armour of enemy robots needs to fine-tune detection, projectile trajectory prediction and gimbal servo control. In addition, decision making in the simulation is quite different from that in the real game scenario. Partially observable decision process needs testing to find the appropriate action space.

Now it is time for you to implement your solution on the real robot platform. Here is some advice about tuning the robot:

1.        When you received the robot, firstly read the 《ICRA 2019 DJI RoboMaster AI Robot User Manual》. Here is the link to download it:
ICRA 2019 RoboMaster AI Challenge Important Documents:
https://www.robomaster.com/en-US ... ype=announcementSub

a)        Check your circuit topology to make sure each module is connected in the right way.
b)        You must know how to use the remote controller to control your robot, including chassis and gimbal motion, turning on the friction wheel and launching the projectiles under manual mode.

2.        Get onboard PC ready for powering on with the proper voltage transformer and connected with MCU through a USB cable.

3.        Test the robot driver on the onboard PC, make sure the robot is placed in the OPEN and SAFE area:
a)        According to the tutorial, install the software dependencies, compile the robots packages and configure the udev rules on your onboard PC.
b)        Stay on manual mode for remote controller, run the roborts_base_node with Rviz under ROS. Use the remote controller to control the robot, and check if you can get the correct odometry and transform topics.
c)        If everything goes well, publish topics for gimbal or chassis control and switch remote controller to autonomous mode. Check if the robot moves correctly.

Here are some new information for the competition:

1.        In the mid-February, we will release detailed mechanical drawings about game field including projectile supplier and protective perimeter wall. And the wall of the projectile supplier will be attached with the certain marker.
2.        RoboMaster organization will not provide a standard simulator for the game this year.
3.        The communication between robots through the referee system is still under development and is supposed to finish on Mid-March. So teams had better build your own wireless communication link.
4.        We will be glad to hear from teams for the definition of robot illegal behaviours that would take advantage of loopholes in the rule but no more help the inspiration and development of technology. The rules will forbid such behaviours by common consent.

-----------------------------------------------------------------------------------------------------------------------------

Q&A

Q1. 关于补给站的问题:1. 能知道补给站的出弹速度吗?2. 手册里说:一分钟能接受任意一个机器人的两次补弹指令,指的是一个机器人两次,一共四次吗,还是一共两次?谢谢啦
A:1.  Yes, currently the time of supply for 50 projectiles is about 3 seconds.  
2.  2 times in total in 1 minute no matter how many robots get supplied.

Q2. 关于比赛规则:在比赛规则的获胜条件中有一条:“一局比赛时间耗尽时,双方机器人尚且存活,伤害高的一方获胜”。伤害高的一方获胜是指剩余血量高的一方获胜?还是命中次数多的一方获胜?伤害来源不只是射击伤害,还有其他的扣血机制。这里伤害是指(初始血量-对方剩余血量),还是只指射击命中伤害?
A:The initial value minus the remaining blood is the damage caused to the enemy.


Q3. 关于比赛规则:在淘汰赛中,上局比赛结束到下局比赛开始的时间间隔是多少?
A:After one game ends, it will jump to Five-Minute Setup Period of next game.


Q4. 关于比赛规则:补给区的触发条件?补给时长?如果由补给装置发生卡弹造成子弹补给不足如何解决?
A: After the robot sends the command to the referee system, the supply station will start replenishing.

Q5. Would mind to give some tips about when can we purchase manifold2 and referee system? And, how much are they?
A: Manifold2 will be on sales roughly in the middle of Feb. The referee system will be ready at the early March. Please pay attention to our latest announcement.

Q6. 场地周围的围栏是是什么颜色的 高反光白色?
A: In the mid-February, we will release detailed mechanical drawings about game field including projectile supplier and protective perimeter wall. And the colour of the protective perimeter will not influence the game.

Q7. 关于硬件设备:针对去年比赛时购买的UWB在最终比赛场地时,标签标号设置问题导致UWB信息不可靠问题(貌似是大家采用相同标签标号导致),今年是否还会有这个问题?
A. This is an inevitable problem unless every robot out of battlefield turns off its UWB module.


Q8. Should all the processing take place on the robot itself?
A. You can use your own computer out of the game field for computing and build a wireless link for data transmission. But the computer should be placed in the operator area.


Q9. Can we know the exact size of the Manifold2, worrying about the choice of other onboard computers?
A. Please refer to the robot size chart


Q10. When you mention "Visual Servo Control" for the Re-supply what does it mean? What is the precise mechanism that opens the small valve with 50 bullets? Being honest this doesn't look like require any kind of "vision" system.
In the mid-February, we will release detailed mechanical drawings about game field including projectile supplier and protective perimeter wall. And the wall of the projectile supplier will be attached with the certain marker.


Q11. 1. 关于模拟器的设计,我们发现实际上只需要不多的几个参数就可以描述整个游戏,比如用云台的俯仰角,底盘的角度和机器人的坐标就可以表示机器人的空间状态。所以我们打算这样来设计模拟器,就是先假设其他组的可以获得这些信息,然后我们来决策。这样简化了模拟器的设计,请问这样的设计会有什么问题吗?
2.关于强化学习的,实际比赛的时候,我们不能直接获得敌方的比赛信息,血量,枪口热度之类的,但在模拟器里训练的时候可以获得,我们想在训练网络的时候输入实际可以获得信息,但奖励函数使用这些信息,也就是网络的输入与奖励函数的参数是不一样的。这样可以吗,会有什么问题吗?提前谢谢了
1. The abstraction level of observation space and action space will definitely have an impact on the performance. And an end-to-end system is not that stable in the real world. So you may need to train the network separately for the high noise model, or just substitute high noise model with the human-designed model.
2. A new firmware that would provide enemy information for training is in our future plan, but during the competition, you are not able to know enemy information. So reward shaping may be tough work.


Q12. Is it possible to know the specific about the given IMU on the robot? Thanks
A. Development board type A: MPU6500+IST8310
          Single axis gyro module: ADXRS620


Q13. "A:The initial value minus the remaining blood is the damage caused to the enemy."
Therefore, if we have one robot against two, we have double health?
A. No


Q14. 补弹指令具体指什么呢?
A. It is a protocol command sending to Referee system (from STM32 MCU to Referee system). But in the later robots version, we will provide it in the ROS sdk also. So no worry about this question.


Q15. 请问比赛中什么样的冲撞会判犯规?只要碰到围挡或者障碍就算吗?双方机器人相互碰撞如何判罚?
A. Refer to Rule Manual 3.5.3.1 Attack and Compete (Collision Penalty)


Q16. 您好,我们队伍购买了官方车,但是买来的机器人云台是不可控的,请问车上的云台板是完全留待自己开发的吗,还有和github上的官方程序是底层板的程序还是云台板的程序?thank you very much!!
A. Make sure your remote controller is operated in the right mode (refer to AI Robot user manual). And STM32 program is open sourced (RoboRTS-Firmware: https://github.com/RoboMaster/RoboRTS-Firmwareicra2019_dev branch). Onboard ROS program is open sourced(RoboRTS: https://github.com/RoboMaster/RoboRTS ros branch)


Q17. A question about the "Projectile reservoir". I know that we should extend it, but the big question is: Is there any limit on how much bigger can it be? I assume that as long as we respect the 600x600x500 constraints we are fine?
A. Only need to meet the referee system inspection specifications,Refer to Rule Manual Appendix Four - Pre-match Inspection Form


Q18. In the current ROS API, how can we control the firing action? What is the topic for that?
A. /cmd_shoot (roborts_msgs/ShootCmd)
service command of projectile launching(including mode, freq and number) Refer to https://robomaster.github.io/Rob ... k_docs/roborts_base (English version will be released in days)


Q19. 初始50发弹丸如何计算数量?是否会根据上供弹无法完全发射供弹链路内全部弹丸而考虑实际载弹量?
The projectile in the projectile tube will not be cleared when clear the projectile reservoir.



跳转到指定楼层

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
梯队队员Anderson Sean
2019-1-30 18:34:10 来自手机 显示全部楼层
关于补给站的问题:1. 能知道补给站的出弹速度吗?2. 手册里说:一分钟能接受任意一个机器人的两次补弹指令,指的是一个机器人两次,一共四次吗,还是一共两次?谢谢啦

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
梯队队员happysf128
2019-1-31 10:51:27 显示全部楼层
About one robot against two, if we managed to kill one enemy robot, then there is no way the other team can beat us. Is this understanding correct? Thanks

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
梯队队员happysf128
2019-1-31 10:30:37 显示全部楼层
"A:The initial value minus the remaining blood is the damage caused to the enemy."
Therefore, if we have one robot against two, we have double health?

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
步兵training spl
 楼主| 2019-1-31 10:27:54 显示全部楼层
问题三、关于比赛规则:在比赛规则的获胜条件中有一条:“一局比赛时间耗尽时,双方机器人尚且存活,伤害高的一方获胜”。伤害高的一方获胜是指剩余血量高的一方获胜?还是命中次数多的一方获胜?伤害来源不只是射击伤害,还有其他的扣血机制。这里伤害是指(初始血量-对方剩余血量),还是只指射击命中伤害?
A:The initial value minus the remaining blood is the damage caused to the enemy.

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
梯队队员arosendo
2019-1-31 10:25:52 显示全部楼层
Stiven Ding, you should read about SLAM. It is possible to use a LiDAR sensor to localise yourself in the map.

I have a question: When you mention "Visual Servo Control" for the Re-supply what does it mean? What is the precise mechanism that opens the small valve with 50 bullets? Being honest this doesn't look like require any kind of "vision" system

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
步兵training spl
 楼主| 2019-1-31 10:25:16 显示全部楼层
问题一、关于补给站的问题:1. 能知道补给站的出弹速度吗?2. 手册里说:一分钟能接受任意一个机器人的两次补弹指令,指的是一个机器人两次,一共四次吗,还是一共两次?谢谢啦
A:1.  Yes, currently the time of supply for 50 projectiles is about 3 seconds.  
2.  2 times in total in 1 minute no matter how many robots get supplied.

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
步兵training spl
 楼主| 2019-1-31 10:06:26 显示全部楼层
For example,
a)        if you use model-based filtering algorithm, please build your sensor model accurately enough or your algorithm robust enough in the game scenario, especially for UWB and IMU.
b)        If you use learning based algorithm, take care of the computing source for onboard PC so that all algorithms can process in acceptable and steady frequencies at the same time.
c)        If your algorithm highly depends on visual features. Think more about the light condition and possible accidents for enemy robots in the game scenario.
d)        If you want to set up a wireless link for communication between robots, take the transmission quality and delay into consideration.

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
梯队队员20133222
2019-1-31 10:00:02 显示全部楼层
请问补给过程如何触发补给指令?官方有没有补给结束的指示?

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
步兵快拆小分队
2019-1-30 11:58:23 显示全部楼层
全英分享(我来打乱节奏)

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
步兵training spl
 楼主| 2019-1-30 14:06:50 显示全部楼层
怎么都没人理我,天呐

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
梯队队员Anderson Sean
2019-1-30 18:40:44 来自手机 显示全部楼层
不好意思,这好像不是关于如何更好开发机器人的问题,可以问吗?

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
梯队队员Anderson Sean
2019-1-30 19:18:38 来自手机 显示全部楼层
再问两个的问题哈,我是做智能决策的。1. 关于模拟器的设计,我们发现实际上只需要不多的几个参数就可以描述整个游戏,比如用云台的俯仰角,底盘的角度和机器人的坐标就可以表示机器人的空间状态。所以我们打算这样来设计模拟器,就是先假设其他组的可以获得这些信息,然后我们来决策。这样简化了模拟器的设计,请问这样的设计会有什么问题吗?2.关于强化学习的,实际比赛的时候,我们不能直接获得敌方的比赛信息,血量,枪口热度之类的,但在模拟器里训练的时候可以获得,我们想在训练网络的时候输入实际可以获得信息,但奖励函数使用这些信息,也就是网络的输入与奖励函数的参数是不一样的。这样可以吗,会有什么问题吗?提前谢谢了

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
步兵training spl
 楼主| 2019-1-31 09:43:02 显示全部楼层
hi, everyone

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
步兵training spl
 楼主| 2019-1-31 09:43:23 显示全部楼层
welcome joining to the first live session!

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
梯队队员happysf128
2019-1-31 09:45:28 显示全部楼层
Should we use English?

[高校AI挑战赛] ICRA2019 AI Challenge Live Text Tutorial Session with Q&A

[复制链接]
梯队队员Anderson Sean
2019-1-31 09:45:57 来自手机 显示全部楼层
hi, I am watching.
您需要登录后才可以回帖 登录 | 立即注册

本版积分规则

触屏版 | 电脑版

Copyright © 2023 RoboMasters 版权所有 备案号 粤ICP备2022092332号

快速回复 返回顶部 返回列表