Pedal by joeljacobs · Pull Request #251 · commaai/openpilot

joeljacobs · 2018-05-22T19:12:16Z

Added pedal fingerprints, and _comma.dbc to Odyssey, Ridgeline, Pilot and RDX.
Altered _comma.dbc and panda/board/safety/safety_honda.h to account for variance in Pedal threshold.
Tested with 4 Pedal boards and 2 cars (RDX and Odyssey). These threshold changes account for variance in Pedal a-to-d and possibly the physical gas pedals.

jfrux · 2018-05-22T19:16:00Z

Yay! Thanks Joel!

rbiasini · 2018-05-22T19:21:14Z

  }],
  HONDA.ODYSSEY: [{
-    57L: 3, 148L: 8, 228L: 5, 229L: 4, 316L: 8, 342L: 6, 344L: 8, 380L: 8, 399L: 7, 411L: 5, 419L: 8, 420L: 8, 427L: 3, 432L: 7, 450L: 8, 463L: 8, 464L: 8, 476L: 4, 490L: 8, 506L: 8, 542L: 7, 545L: 6, 597L: 8, 662L: 4, 773L: 7, 777L: 8, 780L: 8, 795L: 8, 800L: 8, 804L: 8, 806L: 8, 808L: 8, 817L: 4, 819L: 7, 821L: 5, 825L: 4, 829L: 5, 837L: 5, 856L: 7, 862L: 8, 871L: 8, 881L: 8, 882L: 4, 884L: 8, 891L: 8, 892L: 8, 905L: 8, 923L: 2, 927L: 8, 929L: 8, 963L: 8, 965L: 8, 966L: 8, 967L: 8, 983L: 8, 985L: 3, 1029L: 8, 1036L: 8, 1052L: 8, 1064L: 7, 1088L: 8, 1089L: 8, 1092L: 1, 1108L: 8, 1110L: 8, 1125L: 8, 1296L: 8, 1302L: 8, 1600L: 5, 1601L: 8, 1612L: 5, 1613L: 5, 1614L: 5, 1615L: 8, 1616L: 5, 1619L: 5, 1623L: 5, 1668L: 5
+      57L: 3, 148L: 8, 228L: 5, 229L: 4, 316L: 8, 342L: 6, 344L: 8, 380L: 8, 399L: 7, 411L: 5, 419L: 8, 420L: 8, 427L: 3, 432L: 7, 450L: 8, 463L: 8, 464L: 8, 476L: 4, 490L: 8, 506L: 8, 542L: 7, 545L: 6, 597L: 8, 662L: 4, 773L: 7, 777L: 8, 780L: 8, 795L: 8, 800L: 8, 804L: 8, 806L: 8, 808L: 8, 817L: 4, 819L: 7, 821L: 5, 825L: 4, 829L: 5, 837L: 5, 856L: 7, 862L: 8, 871L: 8, 881L: 8, 882L: 4, 884L: 8, 891L: 8, 892L: 8, 905L: 8, 923L: 2, 927L: 8, 929L: 8, 963L: 8, 965L: 8, 966L: 8, 967L: 8, 983L: 8, 985L: 3, 1029L: 8, 1036L: 8, 1052L: 8, 1064L: 7, 1088L: 8, 1089L: 8, 1092L: 1, 1108L: 8, 1110L: 8, 1125L: 8, 1296L: 8, 1302L: 8, 1600L: 5, 1601L: 8, 1612L: 5, 1613L: 5, 1614L: 5, 1615L: 8, 1616L: 5, 1619L: 5, 1623L: 5, 1668L: 5, 513L: 6,


minor thing: extra spaces

rbiasini · 2018-05-22T19:23:51Z


 // these are set in the Honda safety hooks...this is the wrong place
-const int gas_interceptor_threshold = 328;
+const int gas_interceptor_threshold = 800;


why 800? shouldn't this be 120/0.2539?

I made the determination experimentally - If it was lower than that, the RDX and sometimes the Odyssey would occasionally have analog pedal noise that would cause OP to disengage, when the pedal was at "zero".
I tried setting it to 600 and it disengaged on the RDX. For some reason, leaving the dbc at 120 wasn't an issue.

[deleted my prev message]
if the dbc is at 120, with 0.2539 factor, then the threshold here should be 120/0.2539 (add 1 to be conservative). If, as you say, setting it at 600 causes sporadic cancellations, then we should figure out why :)

rbiasini · 2018-05-22T19:25:01Z

Nice @joeljacobs . Would be better to make independent PRs for opendbc and panda repos.

joeljacobs · 2018-05-22T19:31:12Z

I wondered about that. I'll remove the panda from the PR.

joeljacobs · 2018-05-22T19:40:47Z

Done. Added relevant pull-request for Panda

rbiasini · 2018-05-22T21:37:08Z

@joeljacobs , great. Same thing for opendbc too :).

joeljacobs · 2018-05-29T19:58:00Z

@rbiasini Do I need to update this PR for 0.4.6? I don't think there is any problem with this PR since it's just fingerprints.

rbiasini · 2018-05-29T22:17:24Z

-    57L: 3, 145L: 8, 228L: 5, 229L: 4, 308L: 5, 316L: 8, 339L: 7, 342L: 6, 344L: 8, 380L: 8, 392L: 6, 399L: 7, 419L: 8, 420L: 8, 422L: 8, 425L: 8, 426L: 8, 427L: 3, 432L: 7, 464L: 8, 471L: 3, 476L: 4, 490L: 8, 506L: 8, 545L: 5, 546L: 3, 597L: 8, 660L: 8, 773L: 7, 777L: 8, 780L: 8, 795L: 8, 800L: 8, 804L: 8, 808L: 8, 819L: 7, 821L: 5, 829L: 5, 871L: 8, 882L: 2, 884L: 7, 892L: 8, 923L: 2, 927L: 8, 929L: 8, 963L: 8, 965L: 8, 966L: 8, 967L: 8, 983L: 8, 985L: 3, 1027L: 5, 1029L: 8, 1036L: 8, 1039L: 8, 1064L: 7, 1088L: 8, 1089L: 8, 1108L: 8, 1125L: 8, 1296L: 8, 1365L: 5, 1424L: 5, 1600L: 5, 1601L: 8, 1613L: 5, 1616L: 5, 1618L: 5, 1668L: 5, 2015L: 3
+    57L: 3, 145L: 8, 228L: 5, 229L: 4, 308L: 5, 316L: 8, 339L: 7, 342L: 6, 344L: 8, 380L: 8, 392L: 6, 399L: 7, 419L: 8, 420L: 8, 422L: 8, 425L: 8, 426L: 8, 427L: 3, 432L: 7, 464L: 8, 471L: 3, 476L: 4, 490L: 8, 506L: 8, 545L: 5, 546L: 3, 597L: 8, 660L: 8, 773L: 7, 777L: 8, 780L: 8, 795L: 8, 800L: 8, 804L: 8, 808L: 8, 819L: 7, 821L: 5, 829L: 5, 871L: 8, 882L: 2, 884L: 7, 892L: 8, 923L: 2, 927L: 8, 929L: 8, 963L: 8, 965L: 8, 966L: 8, 967L: 8, 983L: 8, 985L: 3, 1027L: 5, 1029L: 8, 1036L: 8, 1039L: 8, 1064L: 7, 1088L: 8, 1089L: 8, 1108L: 8, 1125L: 8, 1296L: 8, 1365L: 5, 1424L: 5, 1600L: 5, 1601L: 8, 1613L: 5, 1616L: 5, 1618L: 5, 1668L: 5, 2015L: 3, 513L: 6,
+    # sent messages
+    0xe4: 5, 0x1fa: 8, 0x200: 6, 0x30c: 8, 0x33d: 5, 0x35e: 8, 0x39f: 8, 513L: 6,


some of the msg adresses are repeated in the fingerprint lists (the sent messages are in the lists already).

I recommend the following:

not separating anymore between "sent messages" and the rest of the fingerprint. We only did it for the Acura ILX but it should go away (feel free to change it in this PR)

In each fingerprint list, all the message addresses should be in decimal Long format (no hex)

In each fingerprint list, all the message addresses should be in monotonically increasing order (feel free to comment that 513L is the pedal)

I noticed a lot of ids have an L after the id (835L: 4,) But a lot (especially the HEX ones) don't have the L. Do I need to add it to the hex numbers I'm converting to Dec?
Well, I did. :) Hope it was ok.

rbiasini · 2018-05-29T22:20:10Z

@joeljacobs see my comment. Since some changes are needed, feel free to update to 0.4.6.

joeljacobs · 2018-05-31T23:04:32Z

@rbiasini how does this look?

… order.

joeljacobs · 2018-06-18T20:01:30Z

@rbiasini This is now updated for the new refactoring. I also did what you asked and sorted all of the IDs and removed the #sent messages lines.

joeljacobs · 2018-06-18T21:36:17Z

I also converted all hexadecimal entries for all fingerprints, like you suggested.

rbiasini · 2018-06-21T07:41:19Z

is this the same as #268? if yes, than we can close this one as I ended up verifying the correctness of PR 268. Thanks @joeljacobs

joeljacobs · 2018-06-21T09:36:49Z

Yes, I think he and I both changed my original PR for 0.4.7, so they should be more or less the same.

jfrux · 2018-06-26T20:38:08Z

Updated PR is now at #274

rbiasini · 2018-07-16T23:27:15Z

merged in another pr

256d274 Fix Mac installation instruction per: https://github.com/commaai/panda/pull/308/files bfd8ff1 Update cppcheck commit with more coverage b143a1c Fixed Misra complaint 606f1d9 Fixed RTC on non-uno boards, second try. Cannot work when there is no xtal. 933c757 Fix RTC on non-uno boards (#311) 48d0d0c VW button spam: fix safety and add tests (#306) 6cccf96 Fan and IR at 0 when in power savings mode (#309) 0537328 board get_sdk scripts were left on python2 de18a7e bump version after uno merge 1965817 Changed default values for testing a12a148 Uno (#274) 7d29dc5 bump panda version. We really need a better way 4007532 VW: stricter limits to comply with comma safety policy e2e2be9 add safety mode to health packet 101238c turned on VW ignition based CAN logic a0d8d5d fix misra 5.3: check_ignition is intended as check_started and can't be used twice ea636de made check_ignition function to both look at ignition_line and ignition_can 1102e69 make ignition logic common for all cars (#303) 3a110c6 bump version after CMSIS core upgrade 55dfa52 Update core to CMSIS 5.6 release (#251) ee86490 fix linter 2 f410b11 fix linter 55957d6 proper python3 exception inheritance 6ba0f47 fix linter errors 5c49fe0 Merge pull request #145 from gregjhogan/uds 0f36199 timeout is float 396d6aa safety_replay only installs few extra requirements 25af7d3 Misra also need python 3 env 7434c5c centralize requirements for tests a0c37c7 coverage not needed in linter reqs fce38a9 Linter python (#299) 62e2f5c update cppcheck commit 711810d more uds debug 4454e3a better CAN comm abstraction 6b1f28f fix more encoding and some bytes cleanup (#300) 43adad3 fix WARNING_INDICATOR_REQUESTED name 9c857da 0x b64d6fa typing 768fdf7 bytes() > chr().encode() 1be15ea custom errors from thread 68da831 more python3 eb358e8 uds lib example 4f28858 updates for python3 932745f support tx flow control for chunked messages b1c3712 add timeout param cdf2f62 bug fixes b1a3195 fix rx message filtering bug 80fb6a6 convert uds lib to class 59cd2b4 handle separation time in microseconds 4429600 fix separation time parsing c641e66 fix typo 48b8dcc fix flow control delay scale 78f413d flow control delay 33a5167 bug fixes 8ee89a0 multi-frame tx 5e89a9c clear rx buffer and numeric error ids 9662300 fix remaining size calculation 01ef1fa zero pad messages before sending 1ddc973 uds can communication dca176e syntax errors 95be481 SERVICE_TYPE enum 98e73b5 more UDS message type implementation c1c5b03 uds lib 162f485 fix chr to bytes conversions (#298) 4972376 Update VW regression test to follow Comma safety index refactoring (#296) f9053f5 more Python 3 fixes, attempting to fix jenkins wifi regresison test (#295) 2f9e076 Panda safety code for Volkswagen, Audi, SEAT, and Škoda (#293) git-subtree-dir: panda git-subtree-split: 256d274

* initial commit * release version

* 차선변경 차선판단 설정 추가 (LaneLineCheck) 색상 오인식 문제 대응: 0=색상+타입(기존), 1=타입만(점선/실선) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * LaneLineCheck=2 옵션 추가: 실선에서 토크 override 허용 점선은 바로 차선변경, 실선은 토크 넣으면 override Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * LaneLineCheck=2 실선 토크 override 버그 수정 solid_line_blocked일 때 토크+BSD 체크 후 바로 laneChangeStarting으로 전환하도록 수정. 기존에는 이후 분기의 lane_change_available 체크에 걸려서 실선 토크 override가 동작하지 않았음. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * HUD 차선 색상 표시를 LaneLineCheck 옵션에 연동 LaneLineCheck>=1일 때 HUD 차선 색상을 색상 raw값 대신 타입(점선/실선) 기준으로 판단하도록 변경하여 차선변경 로직과 HUD 표시 간 불일치 해소 Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

256d274 Fix Mac installation instruction per: https://github.com/commaai/panda/pull/308/files bfd8ff1 Update cppcheck commit with more coverage b143a1c Fixed Misra complaint 606f1d9 Fixed RTC on non-uno boards, second try. Cannot work when there is no xtal. 933c757 Fix RTC on non-uno boards (commaai#311) 48d0d0c VW button spam: fix safety and add tests (commaai#306) 6cccf96 Fan and IR at 0 when in power savings mode (commaai#309) 0537328 board get_sdk scripts were left on python2 de18a7e bump version after uno merge 1965817 Changed default values for testing a12a148 Uno (commaai#274) 7d29dc5 bump panda version. We really need a better way 4007532 VW: stricter limits to comply with comma safety policy e2e2be9 add safety mode to health packet 101238c turned on VW ignition based CAN logic a0d8d5d fix misra 5.3: check_ignition is intended as check_started and can't be used twice ea636de made check_ignition function to both look at ignition_line and ignition_can 1102e69 make ignition logic common for all cars (commaai#303) 3a110c6 bump version after CMSIS core upgrade 55dfa52 Update core to CMSIS 5.6 release (commaai#251) ee86490 fix linter 2 f410b11 fix linter 55957d6 proper python3 exception inheritance 6ba0f47 fix linter errors 5c49fe0 Merge pull request commaai#145 from gregjhogan/uds 0f36199 timeout is float 396d6aa safety_replay only installs few extra requirements 25af7d3 Misra also need python 3 env 7434c5c centralize requirements for tests a0c37c7 coverage not needed in linter reqs fce38a9 Linter python (commaai#299) 62e2f5c update cppcheck commit 711810d more uds debug 4454e3a better CAN comm abstraction 6b1f28f fix more encoding and some bytes cleanup (commaai#300) 43adad3 fix WARNING_INDICATOR_REQUESTED name 9c857da 0x b64d6fa typing 768fdf7 bytes() > chr().encode() 1be15ea custom errors from thread 68da831 more python3 eb358e8 uds lib example 4f28858 updates for python3 932745f support tx flow control for chunked messages b1c3712 add timeout param cdf2f62 bug fixes b1a3195 fix rx message filtering bug 80fb6a6 convert uds lib to class 59cd2b4 handle separation time in microseconds 4429600 fix separation time parsing c641e66 fix typo 48b8dcc fix flow control delay scale 78f413d flow control delay 33a5167 bug fixes 8ee89a0 multi-frame tx 5e89a9c clear rx buffer and numeric error ids 9662300 fix remaining size calculation 01ef1fa zero pad messages before sending 1ddc973 uds can communication dca176e syntax errors 95be481 SERVICE_TYPE enum 98e73b5 more UDS message type implementation c1c5b03 uds lib 162f485 fix chr to bytes conversions (commaai#298) 4972376 Update VW regression test to follow Comma safety index refactoring (commaai#296) f9053f5 more Python 3 fixes, attempting to fix jenkins wifi regresison test (commaai#295) 2f9e076 Panda safety code for Volkswagen, Audi, SEAT, and Škoda (commaai#293) git-subtree-dir: panda git-subtree-split: 256d274 old-commit-hash: 22023eb

Joel Jacobs added 14 commits May 7, 2018 18:39

Changed fingerprint stuff

79f59a3

No safety no preenable

d58eff6

Pedal works! - makes loud braking noises, but works!

6b289b2

Change dbc pedal scaling

bba17cf

reverted all but fingerprint and dbc - sorta works

996d392

set pedal threshold to 600

99a7527

used generator

8d6a5a4

add fingerprint

cdc635b

tested adjustments for both new boards

da51dd1

Merge branch 'pedal_fp_and_dbc' into pedal

c2b37b4

wanted changes from pedal_fp_and_dbc

b4355bf

add fingerprints

51d85f2

typo fix

1ac3e95

Added RDX to fingerprints.py

9c0d1e5

Fixed typo

ce22bc1

rbiasini reviewed May 22, 2018

View reviewed changes

minor whitespace change

c91e57a

removed panda changes

675505c

joeljacobs mentioned this pull request May 22, 2018

Update gas_interceptor_threshold due to experiments wth RDX and Odyssey commaai/panda#117

Closed

joeljacobs mentioned this pull request May 22, 2018

Changes for Pedal Threshold commaai/opendbc#96

Closed

Removed OpenDBC changes

8f76e23

rbiasini reviewed May 29, 2018

View reviewed changes

Joel Jacobs added 3 commits May 30, 2018 19:24

remove changes prior to merge

c012a79

Merge branch 'comma_devel' into pedal

39fbdf6

Cleaned up fingerprints.py

c5145e7

Joel Jacobs and others added 4 commits June 12, 2018 18:39

Merge branch 'devel' into pedal

b82c0bb

Merge branch 'comma_devel' into pedal

4e4dbf6

updated PR for 0.4.7. This includes cleaning up fingerprint numerical…

5daedb7

… order.

Merge branch 'pedal' of github.com:joeljacobs/openpilot into pedal

64ee308

jfrux mentioned this pull request Jun 22, 2018

Adds Pedal Fingerprints for Honda and small change to Pilot / Ridgeline steerKpV and steerKiV to fit 3.3333 ratio #268

Closed

7 tasks

rbiasini closed this Jul 16, 2018

joeljacobs deleted the pedal branch September 20, 2018 22:48

Vaggysag pushed a commit to Vaggysag/openpilot that referenced this pull request Apr 21, 2020

Update core to CMSIS 5.6 release (commaai#251)

55dfa52

* initial commit * release version

nworb-cire pushed a commit to opgm/openpilot that referenced this pull request Aug 2, 2024

GAC: remove legacy onroad button (commaai#251)

f5a518b

Uh oh!

Conversation

joeljacobs commented May 22, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jfrux commented May 22, 2018

Uh oh!

rbiasini May 22, 2018

Choose a reason for hiding this comment

Uh oh!

rbiasini May 22, 2018

Choose a reason for hiding this comment

Uh oh!

joeljacobs May 22, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rbiasini May 22, 2018

Choose a reason for hiding this comment

Uh oh!

rbiasini commented May 22, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

joeljacobs commented May 22, 2018

Uh oh!

joeljacobs commented May 22, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rbiasini commented May 22, 2018

Uh oh!

joeljacobs commented May 29, 2018

Uh oh!

rbiasini May 29, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

joeljacobs May 31, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rbiasini commented May 29, 2018

Uh oh!

joeljacobs commented May 31, 2018

Uh oh!

joeljacobs commented Jun 18, 2018

Uh oh!

joeljacobs commented Jun 18, 2018

Uh oh!

rbiasini commented Jun 21, 2018

Uh oh!

joeljacobs commented Jun 21, 2018

Uh oh!

jfrux commented Jun 26, 2018

Uh oh!

rbiasini commented Jul 16, 2018

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

joeljacobs commented May 22, 2018 •

edited

Loading

joeljacobs May 22, 2018 •

edited

Loading

rbiasini commented May 22, 2018 •

edited

Loading

joeljacobs commented May 22, 2018 •

edited

Loading

rbiasini May 29, 2018 •

edited

Loading

joeljacobs May 31, 2018 •

edited

Loading