Help With Hal (our space activity monitor)

6 views
Skip to first unread message

Trent

unread,
Dec 30, 2017, 2:21:16 AM12/30/17
to MAG Lab
Hi Folks,

It seems we are having some issues with HAL and our reading of HAL in order to report an accurate "open" or "closed" on the website.  Hal monitors its different input (motion and reed switches for doors, as well as temperature), hal also writes to the access controll board (custom shield that reads the rfid scanner, and opens/locks the door automatically) lists of rfid numbers that can open the door and allows remote opening of the space for certain situations.

We have been having issues with false positives and negatives for the space status.  We need some help getting hal and the website to read status of the sensors and report a true status of the space as open or closed.  This involves correct collection and reporting of the sensors from hal to the website and writing code on the website to interpret the data and report based off of some logic.

can anyone help with this?

Jacob Christ

unread,
Dec 30, 2017, 10:25:18 AM12/30/17
to maglab...@googlegroups.com
Is Hal's code / schematics posted somewhere? If not, maybe the maglab github account would be a good place.  Not volunteering (yet), but without this information anyone willing to work would probably be at a loss.

Jacob

Jacob Christ
ProLinear/PONTECH, Inc.
+1 (909) 652-0670 Phone
http://www.pontech.com

--
You received this message because you are subscribed to the Google Groups "MAG Lab" group.
To unsubscribe from this group and stop receiving emails from it, send an email to maglaboratory+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Trenton Wilson

unread,
Dec 30, 2017, 9:48:17 PM12/30/17
to maglaboratory
i was thinking we would start from scratch.  there is the code to manage the access control system which is available (http://www.accxproducts.com/wiki/index.php?title=File:Open_Access_Control_v4_std_136.zip) there. 

Kia Kroas

unread,
Dec 31, 2017, 3:10:52 AM12/31/17
to MAG Lab
Schematics for the Open Access Board are available from the wiki that Trent linked.

Code for Hal (the sensors) are on GitHub, although I think they may be private repositories.

Code for halley (connected to the open access board to control the doors and access list) should also be on GitHub in our organization account.

Sorry I haven't had a chance to keep a handle on Hal lately, have my focus on work and it's going way waayyy past deadline so don't quite have the mental bandwidth to look into it now. Should cc our new it/web person (Brandon) as he took over in October---and I think recently changed what counts as "open" vs closed.

Brian Seo

unread,
Dec 31, 2017, 3:14:47 AM12/31/17
to maglab...@googlegroups.com, Brandon Lu
Thanks for the feedback. I have included Brandon in the cc, in case he is not in the Google group.

Brandon Lu

unread,
Dec 31, 2017, 7:06:06 AM12/31/17
to MAG Lab
Hello all,

It could be possible that a human did not turn the Open Switch to the off position and the broken Pod Bay Door which consistently read open generated a condition where the space was falsely open.

The current logic which determines the open status of the space is as follows:

Set Open Status to open if Open Switch is on and (Pod Bay Door is open or Front Door is open) otherwise set to closed.

Based on discussions that we've had about the space and the broken open status, I am strongly leaning towards writing a disclaimer on the "Are we open?" page and changing the logic so that Open Status is determined solely on the status of the Open Switch.  This logic change is based on the fact that Winter is cold, and open doors make the space colder.  Additionally, this change would coincide with the history and intent of naming our monitoring system HAL.

In the movie 2001: A Space Odyssey, a computerized AI controls the Discovery One's systems on its trip to Jupiter.  The AI is named HAL 9000 and insists that it is "foolproof and incapable of error."  During the course of the movie, Hal reports the status of an antenna control on the ship as faulty.  The astronauts perform extra-vehicular activity (EVA) to retrieve the supposedly faulty antenna control, but find nothing wrong with it.  Hal insists that this is due to human error and that the antenna control should be reinstalled to observe its behavior until it fails.  Ground control's twin HAL 9000 disagrees with Hal's observations.  The astronauts have a discussion about Hal's behavior in supposed secret in an EVA pod where they assume that Hal cannot understand their conversation.  Hal is able to read their lips and determine their conversation.  The astronauts agree to disconnect Hal if he is proven wrong.  On the EVA performed to replace the antenna control, Hal overrides the astronaut's control over his EVA pod and severs the umbilical line which kills the astronaut and sends him adrift.  Another astronaut goes to retrieve the now dead astronaut.  Hal shuts off life support systems, killing all remaining astronauts aboard the ship.  The astronaut who went to retrieve the dead astronaut comes back and finds that Hal refuses to open the Pod Bay Doors under the observation that the astronauts would disable (kill) Hal if it did open the Pod Bay Doors.

I should mention that our HAL can not open the Pod Bay Door.  In short, our HAL is a satirical impression of HAL 9000.  It is purposely non-sentient, unimportant, and human-reliant.  And it would be reasonable to change the Open Status of the space to be controlled solely by humans.

Did somebody leave the cat in?  The Shop Motion sensor is still picking up something...
To unsubscribe from this group and stop receiving emails from it, send an email to maglaborator...@googlegroups.com.

Jacob Christ

unread,
Dec 31, 2017, 1:54:12 PM12/31/17
to maglab...@googlegroups.com
Brandon,

Thanks for the discourse. I propose making HAL sentient but leaving the unimportant and human-reliant traits.  Based on HAL's history its clear to me that the errors do not need to be fixed and should be embraced, but we should create a second HAL so that the they (the two HAL's) have an opportunity to disagree with each other.

All seriousness aside, I have put together a repo to document HAL here: https://github.com/MAGLaboratory/hal

Jacob


Jacob Christ
ProLinear/PONTECH, Inc.
+1 (909) 652-0670 Phone
http://www.pontech.com

To unsubscribe from this group and stop receiving emails from it, send an email to maglaboratory+unsubscribe@googlegroups.com.

Trenton Wilson

unread,
Dec 31, 2017, 4:15:15 PM12/31/17
to maglaboratory
my opinion on how the space should display an open and closed state is as follows:

the space has two states: Active, and Inactive

the space becomes inactive after a certain period of time that both the doors have been closed and there has been no motion in the space (regardless of switch position).  While inactive the website displays closed.  the state of the switch(open/closed) is noted for use in reporting decision while space is active.

the space becomes active as soon as a door is opened, after being inactive.  As soon as a door open activates the space, the website should display open.  The space remains so long as a door is open, or there is motion activity in the space, unless there is a change of state on the door switch in which case the space is still active, but the website displays closed.  This will require the logic to be looking for the switch to be switched to the opposite state it was in while inactive to display a closed condition while the space is active.

I suggest the time interval for de-activating the space be 30 mins

Jacob Christ

unread,
Jan 1, 2018, 1:26:44 AM1/1/18
to maglab...@googlegroups.com
If the lights (or a really annoying beeber) were tied to motion sensors then there would be motivation to make sure there is motion that will be sensed.  The beeper can go for like one or two minutes when no motion is sensed so that it doesn't just beep forever.

Jacob
Reply all
Reply to author
Forward
0 new messages