search

robotika / subt-artf

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

FPR1 (Run 1) 859b7191-a38b-4477-8a68-4a70b2bc7355-c-1-fc #170

Open m3d opened 2 years ago

m3d commented 2 years ago

Prvni stripek z Finals Preliminary Round, svet 1, kdy to fak je super gulas tunnel-urban-cave. Videl jsem jen kus Freyja C a snazila se. Posledni reportovane radio bylo:

21:05:42.390016 24 [['TYPE_RESCUE_RANDY', [6576, 46364, 98], 'B60W1C2400LH3600W', True], ['TYPE_GAS', [118695, -69121, -91], 'A0W1C2400RH3600W', True], ['TYPE_GAS', [122325, -74672, -102], 'A0W1C2400RH3600W', False], ['TYPE_GAS', [126595, -74714, -110], 'A0W1C2400RH3600W', False], ['TYPE_ROPE', [93431, 51786, 803], 'B60W1C2400LH3600W', True], ['TYPE_GAS', [135385, -74696, -126], 'A0W1C2400RH3600W', False], ['TYPE_HELMET', [345, 1535, 103], 'H1980W600RH3600W', False], ['TYPE_HELMET', [205473, 45740, 889], 'B60W1C2400LH3600W', True], ['TYPE_HELMET', [340421, 62178, 12105], 'B60W1C2400LH3600W', True], ['TYPE_HELMET', [338649, 57200, 11162], 'B60W1C2400LH3600W', False], ['TYPE_ROPE', [325820, 137959, 23119], 'B60W1C2400LH3600W', True], ['TYPE_PHONE', [325297, 136454, 22862], 'B60W1C2400LH3600W', False], ['TYPE_GAS', [126370, -74672, 127], 'C300W1C1500RH3600W', False], ['TYPE_RESCUE_RANDY', [199828, -60666, -64], 'A0W1C2400RH3600W', True], ['TYPE_BACKPACK', [15257, -23317, 230], 'D360W1C1500LH3600W', True], ['TYPE_GAS', [135311, -74696, 127], 'C300W1C1500RH3600W', False], ['TYPE_EXTINGUISHER', [46682, -12169, 268], 'D360W1C1500LH3600W', True], ['TYPE_VENT', [-4494, -60351, 3628], 'F660W900RH3600W', True], ['TYPE_PHONE', [148763, 19098, 10], 'C300W1C1500RH3600W', True], ['TYPE_BACKPACK', [135807, -76802, 344], 'D360W1C1500LH3600W', False], ['TYPE_GAS', [135225, -74484, 2023], 'F660W900RH3600W', False], ['TYPE_PHONE', [199427, -93715, 2436], 'F660W900RH3600W', False], ['TYPE_PHONE', [199287, -98314, 2458], 'F660W900RH3600W', False], ['TYPE_PHONE', [199374, -118705, 2632], 'F660W900RH3600W', False]]
m3d commented 2 years ago

Teambase view (BTW je to tam pekne zasitovane, kdyz toto jsou online pozice)

3:29:05.424276 received: ['TYPE_RESCUE_RANDY', [6576, 46364, 98], 'B60W1C2400LH3600W', True]
5:06:25.213994 received: ['TYPE_GAS', [118695, -69121, -91], 'A0W1C2400RH3600W', True]
5:18:09.711571 received: ['TYPE_GAS', [122325, -74672, -102], 'A0W1C2400RH3600W', False]
5:23:54.756276 received: ['TYPE_GAS', [126595, -74714, -110], 'A0W1C2400RH3600W', False]
5:28:59.241795 received: ['TYPE_ROPE', [93431, 51786, 803], 'B60W1C2400LH3600W', True]
5:42:30.640997 received: ['TYPE_GAS', [135385, -74696, -126], 'A0W1C2400RH3600W', False]
6:07:51.810044 received: ['TYPE_HELMET', [345, 1535, 103], 'H1980W600RH3600W', False]
8:13:49.253338 received: ['TYPE_HELMET', [205473, 45740, 889], 'B60W1C2400LH3600W', True]
11:30:50.479536 received: ['TYPE_GAS', [126370, -74672, 127], 'C300W1C1500RH3600W', False]
11:35:33.077509 received: ['TYPE_RESCUE_RANDY', [199828, -60666, -64], 'A0W1C2400RH3600W', None]
11:35:33.419545 received: ['TYPE_RESCUE_RANDY', [199828, -60666, -64], 'A0W1C2400RH3600W', True]
11:45:17.983337 received: ['TYPE_BACKPACK', [15257, -23317, 230], 'D360W1C1500LH3600W', True]
11:49:05.552351 received: ['TYPE_GAS', [135311, -74696, 127], 'C300W1C1500RH3600W', False]
14:03:37.877634 received: ['TYPE_EXTINGUISHER', [46682, -12169, 268], 'D360W1C1500LH3600W', True]
14:32:40.280799 received: ['TYPE_VENT', [-4494, -60351, 3628], 'F660W900RH3600W', True]
15:37:56.750985 received: ['TYPE_PHONE', [148763, 19098, 10], 'C300W1C1500RH3600W', True]
16:29:08.957383 received: ['TYPE_HELMET', [340421, 62178, 12105], 'B60W1C2400LH3600W', True]
16:29:08.957383 received: ['TYPE_HELMET', [338649, 57200, 11162], 'B60W1C2400LH3600W', False]
16:29:08.957383 received: ['TYPE_ROPE', [325820, 137959, 23119], 'B60W1C2400LH3600W', True]
16:29:08.957383 received: ['TYPE_PHONE', [325297, 136454, 22862], 'B60W1C2400LH3600W', False]
17:08:56.455714 received: ['TYPE_BACKPACK', [135807, -76802, 344], 'D360W1C1500LH3600W', False]
17:09:01.339555 received: ['TYPE_GAS', [135225, -74484, 2023], 'F660W900RH3600W', False]
18:52:42.466800 received: ['TYPE_PHONE', [199427, -93715, 2436], 'F660W900RH3600W', False]
18:54:25.049070 received: ['TYPE_PHONE', [199287, -98314, 2458], 'F660W900RH3600W', False]
19:26:19.213663 received: ['TYPE_PHONE', [199374, -118705, 2632], 'F660W900RH3600W', False]
21:16:41.202328 received: ['TYPE_HELMET', [173188, -68677, 930], 'F660W900RH3600W', False]

FPR1-networking

m3d commented 2 years ago

potvrzene:

(osgar) md@md-ThinkPad-P50:~/git/osgar$ cat /data/aws/final_prelim/1/score.yml 
11

artefakty:

backpack        [15.1, -23, 0]
backpack        [114.8, -26.8, -10]
cube            [12.37, -55.5, 10]
cube            [249.32, 10.08, -5]
drill           [191.57, -110.71, 1.71]
drill           [206.13, -13.62, -5]
extinguisher    [46.17, -12.64, 0]
extinguisher    [311.28, -10.13, 5]
gas             [10.5, -100, 10]
gas             [118.5, -68.5, 0]
helmet          [205.71, 45.26, 0.75]
helmet          [340.4, 61.38, 12]
phone           [92.5, 11.72, 0]
phone           [148.51, 19.02, 0]
rescue_randy    [6.77, 46.46, -0.26]
rescue_randy    [199.84, -60.3, -0.01]
rope            [93.39, 51.67, 0.73]
rope            [325.73, 136.78, 23.14]
vent            [-4.5, -60.5, 4]
vent            [95.5, -62.83, 4.3]

state tlog

jisa commented 2 years ago

Vsichni roboti toho najezdili docela dost. A na to, ze jich vzdycky pulka dela to stejne, skoncily kazdy nekde uplne jinde. Napr. Freyja D vymetla levou cast prave (!) pulky sveta.

Jestli breadcrumbs opravdu vypadavali, tak jsme to slusne zasitovali.

jisa commented 2 years ago

Je tam neco divne s reportovanim plynu:

5:06:25.213994 received: ['TYPE_GAS', [118695, -69121, -91], 'A0W1C2400RH3600W', True] 5:18:09.711571 received: ['TYPE_GAS', [122325, -74672, -102], 'A0W1C2400RH3600W', False] 5:23:54.756276 received: ['TYPE_GAS', [126595, -74714, -110], 'A0W1C2400RH3600W', False] 5:42:30.640997 received: ['TYPE_GAS', [135385, -74696, -126], 'A0W1C2400RH3600W', False]

Meli bychom reportovat jenom prvni detekci a pak dat pokoj, ne?

Jeste k tomu 11:30:50.479536 received: ['TYPE_GAS', [126370, -74672, 127], 'C300W1C1500RH3600W', False] coz je skoro identicke se tretim reportem robota A.

A se ctvrtym: 11:49:05.552351 received: ['TYPE_GAS', [135311, -74696, 127], 'C300W1C1500RH3600W', False]

A jeste jednou: 17:09:01.339555 received: ['TYPE_GAS', [135225, -74484, 2023], 'F660W900RH3600W', False]

Hm, a pro plyn bychom u drony meli Z snizit o vysku nad podlahou? To muze byt peknych par metru.

m3d commented 2 years ago

Je tam neco divne s reportovanim plynu

Dobry postreh! Plyn jsme resili v Urban, kde drony nebyly, v cave nebyl plyn a ted kdyz se reportuje XYZ, tak se reportuje urcite spatne (resp. ani nevim jaka je definice - predpokladam dvere, pulka a zem). Drony se sice nelisi na 5m, ale ... zkratim to. ACK na prozkoumat!, dik

m3d commented 2 years ago

V pravidlech je: Localization Point: The center point of the entry door threshold (at floor level) of the room where the gas is detected. In rooms with multiple doorways, only one door will be open (i.e., any others will be closed or blocked) and will serve as the relevant localization point.

m3d commented 2 years ago 
5:06:25.213994 received: ['TYPE_GAS', [118695, -69121, -91], 'A0W1C2400RH3600W', True]
5:18:09.711571 received: ['TYPE_GAS', [122325, -74672, -102], 'A0W1C2400RH3600W', False]
5:23:54.756276 received: ['TYPE_GAS', [126595, -74714, -110], 'A0W1C2400RH3600W', False]
5:42:30.640997 received: ['TYPE_GAS', [135385, -74696, -126], 'A0W1C2400RH3600W', False]

... tak zatim potvrzuji, ze mezi 1. a 2. reportem je 6.6m a threshold na report mame 4m, tj. tam je ocekavane, ze neco rekne. Vlastne i 126m a 135m naznacuji velky krok ... obavam se, ze se tam Freyja v techto mistech ztratila??

jisa commented 2 years ago

Ocekavane to neni. V cloudsim2osgar se gas_detected posila jen a pouze, zmeni-li se hodnota ROSovske bool zpravy. Tj. pri spolehlivem senzoru bychom meli videt prichozi pozitivni zpravu pri vjezdu do mistnosti a negativni na vyjezdu. Mozna par prepnuti sem a tam v blizkem okoli, pokud prejizdime prah mistnosti tam a zpatky. Nemeli bychom ale videt spoustu zprav s velkym rozestupem. Co vidime tomuhle ocekavani neodpovida. Chova se to spis jako zasumeny senzor. To by vysvetlovalo i ten desetimetrovy skok. Stoji nas to pomerne hodne detekci. Sest. Je-li limit na reporty nejakych 25 nebo kolik, ctvrtinu jich vyplytvame jen na tohle.

m3d commented 2 years ago

log mam stazeny a po obede na nej kouknu - ty casy jsou zavadejici, protoze to jsou casy, kdy to vidi Teambase, ale Freyja po 5h je v nejake mistnosti se spousotu odbocek - mozna bug na strane modelu? Ma tam byt jenom jeden vstup.

m3d commented 2 years ago

z pohledu A-Freyji to vypada takto:

(osgar) md@md-ThinkPad-P50:~/git/osgar$ python -m osgar.logger /data/aws/final_prelim/1/a0w1c2400rh3600w.log --stream artifact_filter.artf_xyz
5:03:30.693209 35 [['TYPE_GAS', [118695, -69121, -91], 'A0W1C2400RH3600W', None]]
5:15:15.968431 35 [['TYPE_GAS', [122325, -74672, -102], 'A0W1C2400RH3600W', None]]
5:21:01.331224 35 [['TYPE_GAS', [126595, -74714, -110], 'A0W1C2400RH3600W', None]]
5:39:36.532380 35 [['TYPE_GAS', [135385, -74696, -126], 'A0W1C2400RH3600W', None]]
7:04:22.841586 35 [['TYPE_RESCUE_RANDY', [199828, -60666, -64], 'A0W1C2400RH3600W', None]]

a

(osgar) md@md-ThinkPad-P50:~/git/osgar$ python -m osgar.logger /data/aws/final_prelim/1/a0w1c2400rh3600w.log --stream fromrospy.gas_detected
0:00:04.958434 8 False
5:03:30.692272 8 True
5:12:27.582125 8 False
5:15:15.967591 8 True
5:17:52.421943 8 False
5:21:01.330345 8 True
5:26:28.784871 8 False
5:39:36.531474 8 True
5:48:46.957469 8 False
m3d commented 2 years ago

proc jsi to Jakube zavrel? (predpokladam omyl)

tajgr commented 2 years ago

Jo, sorry.. překlikl jsem se.

m3d commented 2 years ago

gas-problem-FPR1-freyjaA napisu organizatorum m. (prechody jsou u krabic a u druhych dveri)

m3d commented 2 years ago

Mail odeslan, pozadal jsem je o potvrzeni, ze je to problem na jejich strane. Ta mistnost je rozhodne pruchozi a dal pokracuje do jeskyne a pak na kolejiste, coz je proti pravidlum. Jirko, jeste jednou diky!

jisa commented 2 years ago

Nu, dobra. Ale ani v tomto pripade bychom nemeli reportovat po sesti metrech, ne? Leda by ten model byl opravdu hodne rozbity.

m3d commented 2 years ago

1) opravit dronu - to je urcite na nasi strane 2) reportovat jinak nez po 4m ... to znamena zmenu logiky. Oni maji rozhodne ten model spatne (mensi bounding box a dva vstupy) a nevim, zda si to my chceme na nasi strane nejak zeslozitovat??

jisa commented 2 years ago

Jaka zmena logiky? Nas kod v tuhle chvili rika "reportuj na vzestupne hrane signalu". To se bud nedelo, nebo tech vzestupnych hran bylo necekane hodne. Rad bych vedel, proc. Pokud tvrdis, ze maji uprostred te velke mistnosti maly bounding box pro pritomnost plynu, pak to chovani chapu a prestanu se do toho vrtat. Nicmene ze by to meli spatne napric pres nekolik svetu mi prijde necekane. Spis bych veril, ze v nejakou chvili zasumeli detektor plynu a my jsme si toho nevsimli. To je moznost, kterou bych rad vyloucil. Jeji reseni by bylo jednoduche. Neudelat ho a v kazdem behu prijit o nekolik reportu by mohlo dost bolet.

m3d commented 2 years ago

Tech vzestupnych hran tam bylo vice nez melo a dvere tam mely byt take jen jedny. Jeste aby nedoslo k omylu - FPR svety jsou fakticky jenom 3 a pousteli to v kazdem 3x, tj. ten problem s GAS mame ve svetech 1, 2 a 3. (mozna jsem to mel oznacit jinak). Uvidime, co odepisi, ale jsem presvedceny, ze to maji spatne ... a podle vseho ani vysledky neplanuji ukazovat??