Watchdog_pi - AIS Guard Zone GUID

[rgleason]

Jon Thanks for clarifying the data is NOT filtered by o's AIS Operating controls or the Options Ship AIS settings! This statement should be in the documentation!

[PierreFDK]

Hi Jon,

I concur using the AIS "Mark targets lost after (min)" setting to determine if the target is still valid. This existing parameter should set the timeout of AIS targets for all AIS related OCPN processes. Will test watchdog_pi-1.9.054-win32.exe
Pierre

[PierreFDK]

Hi Jon,

I have tested the latest WD 1.9.54 patch. The repeating alarm issue is solved indeed which is really useful to keep the number of alarms acceptable.
My testing road map is as per "pi_Watchdog feature request". C1 feature is implemented with Time stamp of the alarm in the message box now available.

I have focus on "A1. WD alarm configuration to set Boundary State "Any", "Active" and "Inactive" that triggers a WD alarm". The WD "Edit Alarm" window has a new layout with "Boundary Type" and "Boundary State" settings now applying to "AIS Guard Zone" alarm. But when tested those settings are not respected. The WD alarm is triggered whenever an AIS target is within a Boundary. See test 02 report.

Please confirm this patch is intended to address the A1 feature request.
Pierre

[jongough]

Pierre,
I have just tried on my Linux system your tests and I get the expected results, i.e. it is working. With the latest patch I added and 'About' page. Can you go into OCPN/Settings/Plugins select Watchdog and then click on Preferences. This should show the version of the plugin being run. Can you check that it shows you the patch level 054.

The more complicated behaviour, i.e. crossing from outside to inside has not been implemented. I am still trying to work out a way of doing this in a reasonable way. So only a few of your requirements are currently implemented.

I will do some further testing on a windows machine to see if there is different behaviour in that environment.

Jon

[PierreFDK]

Hi Jon,

I confirm the WD patch is 1.9.054 and OD 1.3.16
I have uninstalled previous version before installing the new patch.
Would be good to know if this is a WinOS issue or simply me!
Cheers
Pierre

[jongough]

Pierre,
I am setting up a test with Win 7 at the moment, when that is done I will do it with Win 10 and see if they react differently (not unknown) to linux. I just need to get all the versions up to date and built, then I can do the testing.

Jon

[PierreFDK]

Hi Jon,
all my testing is Win7 environment.
Pierre

[jongough]

Hi Pierre,
I think I have found it, or at least a similar issue. I have put up on dropbox a new version of WD, 1.9.055, and OD, 1.3.20. You will find them here:
https://www.dropbox.com/s/5igui6huxy...win32.exe?dl=0
https://www.dropbox.com/s/r2m2vn3vji...win32.exe?dl=0

Jon

[PierreFDK]

Hi Jon,

I have tested feature A1 successfully with WD 1.9.55, OD 1.3.20
Only one comment when same AIS target penetrate/exit/penetrate same boundary the Guard Zone AIS alarm is only triggered for the first penetration. This behavior may be rectified with implementation of feature request A2 I guess.
Can you confirm the same AIS target will not trigger the same GZ alarm as long as it is considered still valid (valid = the AIS target lost after (min) threshold not being exceeded)
Pierre

[jongough]

Pierre,
I have changed it so that when a target leaves a GZ and then re-enters it it will trigger the alarm again. This is WD patch 1.9.056, you will find a windows installable here:
https://www.dropbox.com/s/muirrqp0lo...win32.exe?dl=0

Jon

[PierreFDK]

Hi Jon,
I have tested the WD patch 1.9.056 and this is working fine, target triggers GZ alarm on re-entry. Very good!
Pierre

[rgleason]

Jon & Pierre, I downloaded what is now the previous version ran it and was impressed at the pil tool. Then looked at Pierre's test reports and step by step testing. This is and outstanding model for the testing process I think. After looking at Pierre's report I felt I knew where in the process we were, and what needed to be tested. That is ideal.
I've had to recoop two computers from 2004 that went down due to auto updates and have an arch job, but will catch up to this later. Thankyou both.
ODraw is getting some really nice tools.

[rgleason]

Same with WD!

[jongough]

Pierre,
Thinking about the cases which are now not covered I have the following questions:

AIS GZ with boundary "exclusion" - hatched pattern inside the boundary contour:
AIS targets located inside the boundary and staying inside the boundary do not trigger WD alarm

What happens with popups i.e. if a target that is not seen before suddenly appears inside the boundary? The target may have moved from outside to the inside, but we did not see the AIS target before it appeared in the boundary zone. This can happen when class B transmitters are being blanketed by class A, or the may not be a line of site between you and the target (small low headland which big targets appear over, but small ones don't).

AIS GZ with boundary "inclusion" - hatched pattern outside of the boundary contour:
AIS targets located outside and staying outside the boundary do not trigger WD alarm

For the inclusion alarm you would really need to be able to specify which targets are of interest. Currently all targets are of interest and as such the alarm will trigger for any target outside of the boundary. Does this mean that you want to be able to specify the MMSI for each target to monitor, or each target to ignore. This issue is at what point in time do we set the start time for targets inside and outside the boundary?

AIS targets located outside and moving inside the boundary do not trigger WD alarm

I think this is just a special case of the previous alarm, am I correct?

AIS GZ with boundary "neither" - no hatch (just a line):

This works the same way as an exclusion boundary, i.e. target outside no alarm, target inside raise alarm.

I think you want a new alarm that is boundary crossing alarm such that crossing the boundary in any direction will raise the alarm, but being inside or outside is of no importance. Is this correct

Jon

[PierreFDK]

Originally Posted by jongough

Pierre,
Thinking about the cases which are now not covered I have the following questions:

AIS GZ with boundary "exclusion" - hatched pattern inside the boundary contour:
AIS targets located inside the boundary and staying inside the boundary do not trigger WD alarm

What happens with popups i.e. if a target that is not seen before suddenly appears inside the boundary? The target may have moved from outside to the inside, but we did not see the AIS target before it appeared in the boundary zone. This can happen when class B transmitters are being blanketed by class A, or the may not be a line of site between you and the target (small low headland which big targets appear over, but small ones don't).

AIS GZ with boundary "inclusion" - hatched pattern outside of the boundary contour:
AIS targets located outside and staying outside the boundary do not trigger WD alarm

For the inclusion alarm you would really need to be able to specify which targets are of interest. Currently all targets are of interest and as such the alarm will trigger for any target outside of the boundary. Does this mean that you want to be able to specify the MMSI for each target to monitor, or each target to ignore. This issue is at what point in time do we set the start time for targets inside and outside the boundary?

AIS targets located outside and moving inside the boundary do not trigger WD alarm

I think this is just a special case of the previous alarm, am I correct?

AIS GZ with boundary "neither" - no hatch (just a line):

This works the same way as an exclusion boundary, i.e. target outside no alarm, target inside raise alarm.

I think you want a new alarm that is boundary crossing alarm such that crossing the boundary in any direction will raise the alarm, but being inside or outside is of no importance. Is this correct

Jon

Hi Jon,
I discover I am way behind, I have been to careful on what was going on on the forum and it seems that no thread info were send to my email.
Anyway, here are my thoughts regarding your questions. Basically, I believe we shall remain pragmatic and considering that the alarm is designed to monitor the vicinity of our vessel we should not miss to many receptions of AIS messages. So

AIS GZ with boundary "exclusion" - hatched pattern inside the boundary contour:
AIS targets located inside the boundary and staying inside the boundary do not trigger WD alarm
What happens with popups i.e. if a target that is not seen before suddenly appears inside the boundary?
-> No Alarm. Note that this case could occur when creating the GZ alarm too (defining an OD boundary that already include AIS targets)

AIS GZ with boundary "inclusion" - hatched pattern outside of the boundary contour:
AIS targets located outside and staying outside the boundary do not trigger WD alarm
For the inclusion alarm you would really need to be able to specify which targets are of interest.
-> only targets that were inside then crossing outside should trigger this alarm.
=> If I may, I will make an attempt to describe a logical process to treat the AIS target vs WD alarm:
First, from the time of creation of the WD alarm all AIS messages received can be attributed a location against each OD boundaries (i.e [MMSI123456789]POS(n) : 0=outside, 1=inside). No alarm can be treated with only one received message. Then, for each following AIS message received, the POS parameter is updated and checked against the previous one (POS(n)=0, POS(n+1)=0 or 1). When POS(n+1) differ from POS(n), the movement of the vessel (inside to outside or outside to inside) is checked against the WD settings in order to trigger or not an alarm (alarm triggered only when crossing to the hatched side of the boundary). When a target is not received for the set time (general AIS lost target threshold) the POS parameter is reset for this target.

AIS GZ with boundary "neither" - no hatch (just a line):
This works the same way as an exclusion boundary, i.e. target outside no alarm, target inside raise alarm.

I think you want a new alarm that is boundary crossing alarm such that crossing the boundary in any direction will raise the alarm, but being inside or outside is of no importance. Is this correct
-> Correct

cheers
pierre