search

robotika / subt-artf

Dataset of Artifacts for SubT Challenge
MIT License
1 stars 0 forks source link

P1 - ver80rc1p1 - cb38d467-8aa8-483b-99ce-9e537a6e20fb #22

Open m3d opened 4 years ago

m3d commented 4 years ago

2x drona, 2x Freyja - ale zatim to vypada, ze Freyja zase nevyhazuje breadcrumbs

m3d commented 4 years ago 
- event:
  id: 38
  type: breadcrumb_deploy
  time_sec: 344
  robot: D40W600L

tj. minimalne druha Freyja vyhazovala (a prvni tam opravdu nema zadny zaznam)

m3d commented 4 years ago

C600L:

31.964000000 INFO /cloudsim2osgar [cloudsim2osgar.py:117(main.__init__)] [topics: /clock, /rosout] freya 2 (with comms beacons)

a pozdeji

64.920000000 DEBUG /cloudsim2osgar [cloudsim2osgar.py:138(main.__init__)] [topics: /clock, /C600L/imu/data, /rosout, /statistics, /C600L/breadcrumb/deploy] receiving: deploy []
64.920000000 DEBUG /C600L_proxy [/tmp/binarydeb/ros-melodic-roscpp-1.14.9/src/libros/transport/transport_tcp.cpp:142(TransportTCP::initializeSocket)] [topics: /rosout, /C600L/cmd_vel, /statistics] Adding tcp socket [48] to pollset
64.920000000 INFO /C600L_proxy [/osgar-ws/src/osgar/subt/ros/proxy/ros_proxy_node.cc:588(receiveZmqThread)] [topics: /rosout, /C600L/cmd_vel, /statistics] MD BROADCAST SUCCESS

64.920000000 INFO /C600L/ros_ign_bridge_breadcrumbs [/tmp/ros-ign-bridge-release/include/ros_ign_bridge/factory.hpp:137(Factory<ROS_T, IGN_T>::ros_callback)] [topics: /rosout] Passing message from ROS std_msgs/Empty to Ignition ignition.msgs.Empty (showing msg only once per type
64.928000000 DEBUG /C600L_proxy [/tmp/binarydeb/ros-melodic-roscpp-1.14.9/src/libros/transport/transport_tcp.cpp:142(TransportTCP::initializeSocket)] [topics: /rosout, /C600L/cmd_vel, /statistics] Adding tcp socket [48] to pollset
64.928000000 INFO /C600L_proxy [/osgar-ws/src/osgar/subt/ros/proxy/ros_proxy_node.cc:328(CommClientCallback)] [topics: /rosout, /C600L/cmd_vel, /statistics] Message Contents[{'breadcrumb': [29.78708638973244, 3.5457925476668537, 0.6808749500822268]}
]
65.000000000 DEBUG /cloudsim2osgar [cloudsim2osgar.py:138(main.__init__)] [topics: /clock, /C600L/imu/data, /rosout, /statistics, /C600L/breadcrumb/deploy] receiving: sim_time_sec 65
65.000000000 INFO /cloudsim2osgar [cloudsim2osgar.py:143(main.__init__)] [topics: /clock, /C600L/imu/data, /rosout, /statistics, /C600L/breadcrumb/deploy] ignoring: sim_time_sec 65
zbynekwinkler commented 4 years ago

Logy stahuji.

zbynekwinkler commented 4 years ago

Zatím mám teambase a koukám do rosmsg.radio a vypadá to, že C600L breadcrumbs vyhazovala - celkem 3:

$ ogger aws-ver80rc1p1-T1400.log --stream rosmsg.radio | grep breadcrumb
0:07:37.885019 24 [b'C600L', b"{'breadcrumb': [29.78708638973244, 3.5457925476668537, 0.6808749500822268]}\n"]
0:11:56.266991 24 [b'C600L', b"{'breadcrumb': [59.68973193655027, 1.0588105675836137, 0.3770711204725444]}\n"]
0:16:45.653213 24 [b'C600L', b"{'breadcrumb': [89.67643669060436, 2.2321477469346567, 0.508753192596617]}\n"]

ale nikdo jiný.

zbynekwinkler commented 4 years ago

Hmm. Možná bychom do ground truth měli kreslit i pozice breadcrumbs...

state tlog

m3d commented 4 years ago

Zatím mám teambase a koukám do rosmsg.radio a vypadá to, že C600L breadcrumbs vyhazovala - celkem 3:

$ ogger aws-ver80rc1p1-T1400.log --stream rosmsg.radio | grep breadcrumb 0:07:37.885019 24 [b'C600L', b"{'breadcrumb': [29.78708638973244, 3.5457925476668537, 0.6808749500822268]}\n"] 0:11:56.266991 24 [b'C600L', b"{'breadcrumb': [59.68973193655027, 1.0588105675836137, 0.3770711204725444]}\n"] 0:16:45.653213 24 [b'C600L', b"{'breadcrumb': [89.67643669060436, 2.2321477469346567, 0.508753192596617]}\n"] ale nikdo jiný.

no z pohledu Teambase to i muze znamenat (a myslim, ze je to tento pripad), ze proste neni spojeni, tj. ze uz se tam ty pozice vyhazovanych, ale nevyhozenych, breadcrumbs nedostanou ... melo by to byt videt na grafu z Teambase (tam bych asi take pridal ty breadcrumbs a detekovane artefakty - resp. reportovane pres radio)

zbynekwinkler commented 4 years ago

ack:

$ ogger aws-ver80rc1p1-D40W600L.log | grep bread
15      radio.breadcrumb        336 |     12 |   0.0Hz
35    breadcrumbs.deploy          6 |      6 |   0.0Hz
36  breadcrumbs.location        168 |      6 |   0.0Hz

$ ogger aws-ver80rc1p1-C600L.log | grep bread
32      radio.breadcrumb       196 |     7 |   0.0Hz
40    breadcrumbs.deploy         6 |     6 |   0.0Hz
41  breadcrumbs.location       168 |     6 |   0.0Hz

Tj. obe se snažily o deploy.

$ ogger aws-ver80rc1p1-D40W600L.log --stream breadcrumbs.deploy
0:59:44.302989 35 []
1:05:57.023198 35 []
1:10:52.313041 35 []
1:16:47.566555 35 []
1:22:05.117877 35 []
1:26:49.276785 35 []

$ ogger aws-ver80rc1p1-D40W600L.log --stream breadcrumbs.location
0:59:44.303450 36 [139.7781499693492, -25.01200823272591, -4.997655865500959]
1:05:57.023465 36 [156.1349034760204, -49.5805002068171, -10.421414484607894]
1:10:52.313325 36 [171.19091371817763, -74.97570965800693, -15.76873426297611]
1:16:47.566798 36 [191.89569024228544, -95.94610763728842, -21.441185253967433]
1:22:05.118143 36 [221.79751980145738, -98.43909269927924, -22.22005121034919]
1:26:49.277042 36 [251.40516762748885, -103.28535471387356, -22.401608820602146]

$ ogger aws-ver80rc1p1-C600L.log --stream breadcrumbs.deploy breadcrumbs.location
0:07:29.882067 40 []
0:07:29.882457 41 [29.78708638973244, 3.5457925476668537, 0.6808749500822268]
0:11:47.939784 40 []
0:11:47.940203 41 [59.68973193655027, 1.0588105675836137, 0.3770711204725444]
0:16:37.606875 40 []
0:16:37.607143 41 [89.67643669060436, 2.2321477469346567, 0.508753192596617]
0:21:32.624387 40 []
0:21:32.624666 41 [119.60683875197853, 3.073322563629506, -1.5280489279319678]
0:26:29.561629 40 []
0:26:29.561940 41 [149.51111319369795, 3.3674539869917473, -4.045188868417916]
0:31:29.773588 40 []
0:31:29.773856 41 [179.30387847026407, 1.6460692566140354, -0.8811029583104538]

Tj. podle očekávání se druhá Freya snažila vyhazovat o dost dále (a později) než ta první.

zbynekwinkler commented 4 years ago

Je možné, že se jenom prostě nepropojili až na teambase? Možná bychom neměli vyhazovat po vzdálenosti, ale po ztrátě komunikace - respektive se vrátit na poslední místo, kde komunikace fungovala a tam to vyhodit. Ostatní týmy se snaží vyhazovat, když nějak klesá RSSI, ale co jsem pochopil, tak s tím jsou nějaké problémy. Je mi ale jasné, že tohle by byl větší zásah :worried:

zbynekwinkler commented 4 years ago

Obě drony se vydržely správně lokalizovat až zhruba do doby, kdy jim měly dojít baterky:

$ python -m subt.tools.validator
DatabaseError: database disk image is malformed
min x: -47.29 y: -379.23
max x: 511.06 y: 26.44
width: 5603px height: 4076px
Ground truth count: 151025
Processing logfile: aws-ver80rc1p1-T1400.log
  Autodetected name: T1400
  skiping teambase
Processing logfile: aws-ver80rc1p1-B10W1300R.log
  Autodetected name: B10W1300R
  Autodetect pose3d stream: offseter.pose3d
  Trace reduced count: 1403
  Ground truth seconds: 1409
  sim_time_sec:   222, error:  1.04m, distance:  153.79m from origin, path:  186.21m
  sim_time_sec:   347, error:  2.03m, distance:  135.75m from origin, path:  305.60m
  sim_time_sec:   752, error:  3.02m, distance:  327.25m from origin, path:  711.88m
  sim_time_sec:   972, error:  4.09m, distance:  227.28m from origin, path:  944.40m
  sim_time_sec:  1033, error:  5.02m, distance:  169.90m from origin, path: 1014.44m
  Maximum error: 1410.35m -> FAIL
Processing logfile: aws-ver80rc1p1-A1300L.log
  Autodetected name: A1300L
  Autodetect pose3d stream: offseter.pose3d
  Trace reduced count: 1402
  Ground truth seconds: 1409
  sim_time_sec:   382, error:  1.01m, distance:  246.60m from origin, path:  353.67m
  sim_time_sec:   567, error:  2.07m, distance:  197.88m from origin, path:  547.53m
  sim_time_sec:   584, error:  3.03m, distance:  183.66m from origin, path:  564.02m
  sim_time_sec:   645, error:  4.00m, distance:  138.27m from origin, path:  624.63m
  sim_time_sec:  1279, error:  5.16m, distance:  239.84m from origin, path: 1265.04m
  Maximum error: 848.27m -> FAIL
Processing logfile: aws-ver80rc1p1-D40W600L.log
  Autodetected name: D40W600L
  Autodetect pose3d stream: localization.pose3d
  Trace reduced count: 1403
  Ground truth seconds: 1409
  sim_time_sec:   243, error:  1.01m, distance:  185.89m from origin, path:  210.03m
  sim_time_sec:   264, error:  2.26m, distance:  162.06m from origin, path:  235.28m
  sim_time_sec:   400, error:  3.04m, distance:  185.12m from origin, path:  395.52m
  sim_time_sec:   417, error:  4.05m, distance:  200.36m from origin, path:  413.71m
  sim_time_sec:   542, error:  5.20m, distance:  295.07m from origin, path:  522.63m
  Maximum error: 40.92m -> FAIL
Processing logfile: aws-ver80rc1p1-C600L.log
  Autodetected name: C600L
  Autodetect pose3d stream: localization.pose3d
  Trace reduced count: 410
  Ground truth seconds: 1409
  sim_time_sec:   242, error:  1.04m, distance:  191.52m from origin, path:  249.97m
  sim_time_sec:   261, error:  2.04m, distance:  180.54m from origin, path:  264.04m
  Maximum error: 2.34m -> OK
zbynekwinkler commented 4 years ago

Ale zpět k breadcrumbs. Nejprve C, která jela jako první:

$ ogger aws-ver80rc1p1-T1400.log --stream rosmsg.radio | grep bread
0:07:37.885019 24 [b'C600L', b"{'breadcrumb': [29.78708638973244, 3.5457925476668537, 0.6808749500822268]}\n"]
0:11:56.266991 24 [b'C600L', b"{'breadcrumb': [59.68973193655027, 1.0588105675836137, 0.3770711204725444]}\n"]
0:16:45.653213 24 [b'C600L', b"{'breadcrumb': [89.67643669060436, 2.2321477469346567, 0.508753192596617]}\n"]

versus

$ ogger aws-ver80rc1p1-C600L.log --stream breadcrumbs.location
0:07:29.882457 41 [29.78708638973244, 3.5457925476668537, 0.6808749500822268]
0:11:47.940203 41 [59.68973193655027, 1.0588105675836137, 0.3770711204725444]
0:16:37.607143 41 [89.67643669060436, 2.2321477469346567, 0.508753192596617]
0:21:32.624666 41 [119.60683875197853, 3.073322563629506, -1.5280489279319678]
0:26:29.561940 41 [149.51111319369795, 3.3674539869917473, -4.045188868417916]
0:31:29.773856 41 [179.30387847026407, 1.6460692566140354, -0.8811029583104538]

Ten čtrvrtý byl asi vyhozen moc pozdě a nepropojil se s těmi prvními třemi. Díky tomu, že algoritmus vyhazování druhé Freyi byl stejný, tak tohle přešla a svoje breadcrumbs naházela ještě dál.

Čas 16:37: image a 21:32: image

Není tam nic zásadního vidět. Lokalizace ještě fungovala ok. Tj. zdá se, že vše běželo, jak mělo. Jen to neseje.

zbynekwinkler commented 4 years ago

Ok, zpět na stromy. Jediné pozice breadcrumbs přítomné v state.tlog jsou tyto:

{'D40W600L__breadcrumb___0': x: 139.4285030910548
 y: -23.26669406789718
 z: -6.334343784307123
,
 'D40W600L__breadcrumb___1': x: 156.92357474968108
 y: -47.599128489355834
 z: -11.977990347449404
,
 'D40W600L__breadcrumb___2': x: 173.01352986728364
 y: -72.84212488852077
 z: -17.498251104885945
,
 'D40W600L__breadcrumb___3': x: 194.8781903527646
 y: -93.56383345451263
 z: -23.479635602783528
,
 'D40W600L__breadcrumb___4': x: 224.8326211749733
 y: -96.23738560725971
 z: -24.33324580311369
,
 'D40W600L__breadcrumb___5': x: 254.36062051317333
 y: -100.4500402092267
 z: -24.673276134265688
}

Tj. C600L opravdu nevyhazovala.

zbynekwinkler commented 4 years ago

rosout.log obsahuje všech 6 řádek obsahujících "receiving: deploy []". Tj. vyhnilo to někde za cloudsim2osgar.

Také jsem našel 

[ INFO] [1602013990.957267910, 64.920000000]: Passing message from ROS std_msgs/Empty to Ignition ignition.msgs.Empty (showing msg only once per type

takže minimálně jedenkrát se to dostalo z ROSu do ign světa - minimálně 1 breadcrumb by tam měl být.

Asi je to ready na issue - udělám ho.

zbynekwinkler commented 4 years ago

Je to 655 na https://github.com/osrf/subt/issues/, ale nechtěl jsem sem dávat přímo link, ať to nemate na druhé straně, kde se objeví link zpátky.