Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

wander & multi agent training #9

Open
1 of 3 tasks
CzJaewan opened this issue Sep 4, 2020 · 5 comments
Open
1 of 3 tasks

wander & multi agent training #9

CzJaewan opened this issue Sep 4, 2020 · 5 comments
Assignees
@CzJaewan

Comments

@CzJaewan
Copy link
Collaborator

CzJaewan commented Sep 4, 2020 •
edited
Loading

  • ppo train in wander world

  • ppo train in wander & multi agent world

  • ppo train in multi agent world

비교
@CzJaewan
Copy link
Collaborator Author

CzJaewan commented Sep 4, 2020 •
edited by Geonhee-LEE
Loading

  • wander.py 만들어서 실행.
    꼬일 까봐 git을 하나 더만듦. 나중에 통합하겠음.

  • git : https://github.com/CzJaewan/RLCA_trainning

  • wander가 많으면 stageros segmentation fault 발생
    memory 관련 문제인 듯 하나 해결책 찾지 못함.

  • error msg

(base) nscl@nscl-lab:~$ rosrun stage_ros_add_pose_and_crash stageros -u /home/nscl/catkin_ws/src/RLseminar20summer/worlds/servingbot_agent2.world 
 [Loading /home/nscl/catkin_ws/src/RLseminar20summer/worlds/servingbot_agent2.world][threads 4]
[ INFO] [1599199342.362621085]: Found 1 laser devices and 0 cameras in robot 0
[ INFO] [1599199342.370267229]: Found 1 laser devices and 0 cameras in robot 1
[ INFO] [1599199342.376595537]: Found 1 laser devices and 0 cameras in robot 2
[ INFO] [1599199342.383256327]: Found 1 laser devices and 0 cameras in robot 3
[ INFO] [1599199342.390224147]: Found 1 laser devices and 0 cameras in robot 4
[ INFO] [1599199342.396946725]: Found 1 laser devices and 0 cameras in robot 5
[ INFO] [1599199342.404789802]: Found 1 laser devices and 0 cameras in robot 6
[ INFO] [1599199342.413297790]: Found 1 laser devices and 0 cameras in robot 7
[ INFO] [1599199342.423249841]: Found 1 laser devices and 0 cameras in robot 8
[ INFO] [1599199342.435438903]: Found 1 laser devices and 0 cameras in robot 9
[ INFO] [1599199354.067220492]: Resetting stage!
Segmentation fault (core dumped)
  • 관련 사이트

ros-simulation/stage_ros#55 *************** 시도
https://answers.ros.org/question/9217/segmentation-fault-on-running-stageros/
Acmece#2

@CzJaewan
Copy link
Collaborator Author

CzJaewan commented Sep 5, 2020

wander py는 lidar로 물체를 회피하기 때문에 라이다 사용하지 않는 직선 주행만 하는 wander 필요.

@CzJaewan
Copy link
Collaborator Author

CzJaewan commented Sep 6, 2020 •
edited by Geonhee-LEE
Loading

  • stageros.cpp 에서 posestamp 제거
  • world.py 에서 ramdom.uniform -> round(ramdom.uniform)

@CzJaewan
Copy link
Collaborator Author

CzJaewan commented Sep 7, 2020 •
edited by Geonhee-LEE
Loading 

*** Error in `/home/geonhee-ml/rl_ws/devel/lib/stage_ros_add_pose_and_crash/stageros': double free or corruption (fasttop): 0x00000000033ba380 ***
[ INFO] [1599484203.436257029]: index : 0
[ INFO] [1599484203.436397108]: index : 0
[ INFO] [1599484203.436513289]: index : 0
[ INFO] [1599484203.436609923]: index : 0
[ INFO] [1599484203.436734070]: index : 0
[ INFO] [1599484203.437791021]: index : 0
[ INFO] [1599484203.437896247]: index : 0
[ INFO] [1599484203.437998189]: index : 0
[ INFO] [1599484203.438090848]: index : 0
[ INFO] [1599484203.438179245]: index : 0
[ INFO] [1599484203.438321847]: index : 0
[ INFO] [1599484203.438404599]: index : 0
[ INFO] [1599484203.438573570]: index : 0
[ INFO] [1599484203.438938008]: index : 0
[ INFO] [1599484203.439214576]: index : 0
[ INFO] [1599484203.439501216]: index : 0
[ INFO] [1599484203.439761873]: index : 0
[ INFO] [1599484203.440019382]: index : 0
======= Backtrace: =========
/lib/x86_64-linux-gnu/libc.so.6(+0x777f5)[0x7f7cd1f157f5]
[ INFO] [1599484203.440351471]: index : 0
/lib/x86_64-linux-gnu/libc.so.6(+0x8038a)[0x7f7cd1f1e38a]
/lib/x86_64-linux-gnu/libc.so.6(cfree+0x4c)[0x7f7cd1f2258c]
[ INFO] [1599484203.440622768]: index : 0
/opt/ros/kinetic/lib/libstage.so.4.1.1(_ZN3Stg4Cell8AddBlockEPNS_5BlockEj+0x139)[0x7f7cd2e744d9]
/opt/ros/kinetic/lib/libstage.so.4.1.1(_ZN3Stg5World7MapPolyERKSt6vectorINS_11point_int_tESaIS2_EEPNS_5BlockEj+0x38e)[0x7f7cd2e7b54e]
/opt/ros/kinetic/lib/libstage.so.4.1.1(_ZN3Stg5Block3MapEj+0x65)[0x7f7cd2e54675[ INFO] [1599484203.440876951]: index : 0
/opt/ros/kinetic/lib/libstage.so.4.1.1(_ZN3Stg10BlockGroup3MapEj+0x27)[0x7f7cd2e55187]
/opt/ros/kinetic/lib/libstage.so.4.1.1(_ZN3Stg5Model3MapEj+0x20)[0x7f7cd2e5c220]
/opt/ros/kinetic/lib/libstage.so.4.1.1(_ZN3Stg5Model15MapWithChildrenEj+0x18)[0x7f7cd2e5c278]
/opt/ros/kinetic/lib/libstage.so.4.1.1(_ZN3Stg5Model15MapWithChildrenEj+0x36)[0x7f7cd2e5c296]
[ INFO] [1599484203.441218399]: index : 0
/opt/ros/kinetic/lib/libstage.so.4.1.1(_ZN3Stg13ModelPosition4MoveEv+0x245)[0x7f7cd2e6c115]
/opt/ros/kinetic/lib/libstage.so.4.1.1(_ZN3Stg5World6UpdateEv+0x25d)[0x7f7cd2e7ac8d]
/opt/ros/kinetic/lib/libstage.so.4.1.1(_ZN3Stg8WorldGui6UpdateEv+0x74)[0x7f7cd2e92484]
[ INFO] [1599484203.441502237]: index : 0
/usr/lib/x86_64-linux-gnu/libfltk.so.1.3(_ZN2Fl4waitEd+0x7d)[0x7f7cd2b4bbcd]
[ INFO] [1599484203.441743887]: index : 0
/home/geonhee-ml/rl_ws/devel/lib/stage_ros_add_pose_and_crash/stageros(main+0x2bb)[0x472d5d]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf0)[0x7f7cd1ebe840]
[ INFO] [1599484203.442030529]: index : 0
/home/geonhee-ml/rl_ws/devel/lib/stage_ros_add_pose_and_crash/stageros(_start+0x29)[0x46c1c9]
======= Memory map: ========
00400000-004d4000 r-xp 00000000 08:03 2102607                            /home/geonhee-ml/rl_ws/devel/lib/stage_ros_add_pose_and_crash/stageros
006d3000-006d4000 r--p 000d3000 08:03 2102607                            /home/geonhee-ml/rl_ws/devel/lib/stage_ros_add_pose_and_crash/stageros
006d4000-006d5000 rw-p 000d4000 08:03 2102607                            /home/geonhee-ml/rl_ws/devel/lib/stage_ros_add_pose_and_crash/stageros
006d5000-006d6000 rw-p 00000000 00:00 0 
026c3000-03503000 rw-p 00000000 00:00 0                                  [heap]
405cc000-405ce000 r-xs 00000000 08:03 4340609                            /tmp/.glRAEld9 (deleted)
4189c000-4194e000 rw-p 00000000 00:00 0 
7f7c84000000-7f7c84021000 rw-p 00000000 00:00 0 
7f7c84021000-7f7c88000000 ---p 00000000 00:00 0 
7f7c8bc00000-7f7c8c000000 rw-s 00000000 00:06 485                        /dev/nvidiactl
7f7c8c000000-7f7c8c022000 rw-p 00000000 00:00 0 
7f7c8c022000-7f7c90000000 ---p 00000000 00:00 0 
7f7c90000000-7f7c90022000 rw-p 00000000 00:00 0 [ INFO] [1599484203.442276950]: index : 0

7f7c90022000-7f7c94000000 ---p 00000000 00:00 0 
7f7c94000000-7f7c94022000 rw-p 00000000 00:00 0 
7f7c94022000-7f7c98000000 ---p 00000000 00:00 0 
7f7c98000000-7f7c98022000 rw-p 00000000 00:00 0 
7f7c98022000-7f7c9c000000 ---p 00000000 00:00 0 
7f7c9c000000-7f7c9c022000 rw-p 00000000 00:00 0 
7f7c9c022000-7f7ca0000000 ---p 00000000 00:00 0 
7f7ca0000000-7f7ca0022000 rw-p 00000000 00:00 0 
7f7ca0022000-7f7ca4000000 ---p 00000000 00:00 0 
7f7ca4000000-7f7ca4021000 rw-p 00000000 00:00 0 
7f7ca4021000-7f7ca8000000 ---p 00000000 00:00 0 
7f7ca8000000-7f7ca8022000 rw-p 00000000 00:00 0 
7f7ca8022000-7f7cac000000 ---p 00000000 00:00 0 
7f7cac000000-7f7cac021000 rw-p 00000000 00:00 0 
7f7cac021000-7f7cb0000000 ---p 00000000 00:00 0 
7f7cb03f1000-7f7cb07f1000 rw-s 00000000 00:06 485                        /dev/nvidiactl
7f7cb07f1000-7f7cb07f2000 ---p 00000000 00:00 0 
7f7cb07f2000-7f7cb0ff2000 rw-p 00000000 00:00 0 
7f7cb0ff2000-7f7cb0ff3000 ---p 00000000 00:00 0 
7f7cb0ff3000-7f7cb17f3000 rw-p 00000000 00:00 0 
7f7cb17f3000-7f7cb17f4000 ---p 00000000 00:00 0 
7f7cb17f4000-7f7cb1ff4000 rw-p 00000000 00:00 0 
7f7cb1ff4000-7f7cb1ff5000 ---p 00000000 00:00 0 
7f7cb1ff5000-7f7cb27f5000 rw-p 00000000 00:00 0 
7f7cb27f5000-7f7cb27f6000 ---p 00000000 00:00 0 
7f7cb27f6000-7f7cb2ff6000 rw-p 00000000 00:00 0 
7f7cb2ff6000-7f7cb2ff7000 ---p 00000000 00:00 0 
7f7cb2ff7000-7f7cb37f7000 rw-p 00000000 00:00 0 
7f7cb37f7000-7f7cb37f8000 ---p 00000000 00:00 0 
7f7cb37f8000-7f7cb3ff8000 rw-p 00000000 00:00 0 
7f7cb3ff8000-7f7cb3ff9000 ---p 00000000 00:00 0 
7f7cb3ff9000-7f7cb47f9000 rw-p 00000000 00:00 0 
7f7cb47f9000-7f7cb47fa000 ---p 00000000 00:00 0 
7f7cb47fa000-7f7cb4ffa000 rw-p 00000000 00:00 0 
7f7cb4ffa000-7f7cb4ffb000 ---p 00000000 00:00 0 
7f7cb4ffb000-7f7cb57fb000 rw-p 00000000 00:00 0 
7f7cb57fb000-7f7cb57fc000 ---p 00000000 00:00 0 
7f7cb57fc000-7f7cb5ffc000 rw-p 00000000 00:00 0 
7f7cb5ffc000-7f7cb5ffd000 ---p 00000000 00:00 0 
7f7cb5ffd000-7f7cb67fd000 rw-p 00000000 00:00 0 
7f7cb67fd000-7f7cb67fe000 ---p 00000000 00:00 0 [ INFO] [1599484203.442600838]: index : 0

7f7cb67fe000-7f7cb6ffe000 rw-p 00000000 00:00 0 
7f7cb6ffe000-7f7cb6fff000 ---p 00000000 00:00 0 
7f7cb6fff000-7f7cb77ff000 rw-p 00000000 00:00 0 
7f7cb77ff000-7f7cb7800000 ---p 00000000 00:00 0 
7f7cb7800000-7f7cb8000000 rw-p 00000000 00:00 0 
7f7cb8000000-7f7cb8021000 rw-p 00000000 00:00 0 
7f7cb8021000-7f7cbc000000 ---p 00000000 00:00 0 
7f7cbc000000-7f7cbc021000 rw-p 00000000 00:00 0 
7f7cbc021000-7f7cc0000000 ---p 00000000 00:00 0 
7f7cc0000000-7f7cc0044000 rw-p 00000000 00:00 0 
7f7cc0044000-7f7cc4000000 ---p 00000000 00:00 0 
7f7cc40a9000-7f7cc42a9000 rw-s 00000000 00:06 485                        /dev/nvidiactl
7f7cc42a9000-7f7cc42aa000 ---p 00000000 00:00 0 
7f7cc42aa000-7f7cc4aaa000 rw-p 00000000 00:00 0 
7f7cc4aaa000-7f7cc4aab000 ---p 00000000 00:00 0 
7f7cc4aab000-7f7cc52ab000 rw-p 00000000 00:00 0 
7f7cc54ab000-7f7cc54b4000 r-xp 00000000 08:03 7344320                    /usr/lib/x86_64-linux-gnu/libXcursor.so.1.0.2
7f7cc54b4000-7f7cc56b3000 ---p 00009000 08:03 7344320                    /usr/lib[ INFO] [1599484203.442860637]: index : 0
Aborted (core dumped)
geonhee-ml@geonhee:~$

@CzJaewan
Copy link
Collaborator Author

CzJaewan commented Sep 14, 2020 •
edited
Loading

stageros 수정

  • stageros -> stageros_w
    • stageros_w 내에서 wander control
    • -n 명령어 추가
    • ex) rosrun stage_ros_add_crash_pose stageros_w -u -n 2 /home/world/single_robot.world : wander 2개

Gazebo wander env test

  • wander 6, agent 1 env 학습, epi 100000
    • 물체 앞에 멈췄다 움직이는 경향이 생기나 충돌 또한 많음 7/10 주행 성공률
    • 원인
      • reward
      • stage env noise
      • lidar observation data 끼리의 time이 너무 짧음. [t_(-2), t_(-1), t]
        -> gazebo test 시 : step 간의 시간은 0.01~ 0.1 사이 평균적으로 한 step에 0.05 sec 걸린다고 판단됨.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@CzJaewan CzJaewan

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@CzJaewan