ability to read values of loops

[emiltin]

Ability to subscribe to actual value of a detector. Perhaps also get min/max value within the last period. Ability to read values would make it easier to adjust the trigger level.

[otterdahl]

We can make a suggestion of how this status can be designed, but we to sync this with the suppliers.

[otterdahl]

I'm unsure how to design this status in a good way. I suppose the following is needed:

• current value
• max value within the last period
• min value within the last period

The current value can be requested (StatusRequest) but not subscribed to with update on change (because it will change all the time)

Max and min value can be subscribed to with a given interval, but not be requested or subscribed with update on change - since it is not clear which period it refers to.

ObjectType	Status	Description	Name	Type	Value	Comment
Traffic Controller	S00XX	Current value of detector	value	real	[value]

ObjectType	Status	Description	Name	Type	Value	Comment	Name	Type	Value	Comment
Traffic Controller	S00XX	Max/min value of detector	max	real	[value]	Max value of detector since last period	min	real	[value]	Min value of detector since last period

[emiltin]

Is is a problem to allow rate=0? If it really changes all the time, then maybe you want to get updates all the time. If you don't want that, use rate=1, rate=60 or some other interval.

Regarding min and max, conceptually it culd make sense to not allow rate=0, since you need sme time interval to aggregate across. But it reality, even when rate=0, updates will be send with (possibly irregular) interval. Min and max could then simply be the min and max value seen since last update was send.

Can value and mix/max be combined into a single StatusUpdate?

[emiltin]

It's a bit unclear what we really want here

[emiltin]

If the controller keeps track of when a car/bicycle is detected, and this is based on where tops in the input value are, then maybe we could subscribe to these tops. Ie. get a list of tops since last status update, with the level of each top. This could then be logged and used for checking if the detector sensitivity is set correcty.

[emiltin]

Range should be defined eg. frm 0-255. Integers are enough. 255 should be the maximum value the sensor can output.

[emiltin]

Looking at equipment, there's a few ways to look at the current loop values:

• current value
• difference
• max difference

But we need input from suppliers

[emiltin]

If the controller keeps track of when a car/bicycle is detected, and this is based on where tops in the input value are, then maybe we could subscribe to these tops. Ie. get a list of tops since last status update, with the level of each top. This could then be logged and used for checking if the detector sensitivity is set correcty.

Can we find a clear definition of what "tops" are that works across suppliers? Keeping in mind that it's not simply a treshold, since for adjusting loops we also want to see tops below the treshold

[emiltin]

max value since last update?

[rafaelcf2]

Some TLCs can provide other kind of information like Frequency, flux or indutance like below

Dynniq TLC (very user-friendly interface) Loop	Detector	Freq [kHz]	Ind [uH]	Signal	Sens [dL]	Hold [min]	PH	Reset
1-1-1	D1_A1	56	155	0	50	90	0	#
1-1-2	D2_A1	62	124	0	50	90	0	#
1-1-3	D3_A1	47	223	0	50	90	0	#
1-1-4	D4_A1	56	150	0	10	90	0	#

We need also to address the detector to the signal group like D1_A1 or D1_G01

Other way to read the values is just specify a message like:

timestamp, frequency rate in hz, data stream in hex 8 or 16bit with 10 to 2000 interpolated values

/Rafael

[SwarcoPalm]

The configuration parts sound fairly useful to me (like which group does it belong to, what is the sensitivity, maybe the hold or balancing timer). I'm not sure about frequency and inductance? What would you do with these values? Are they of any real use?

[rafaelcf2]

Only if the loop detectors are in bad shape, very high impedance or very low inductance are sideefects of a bad loop

[emiltin]

The simplest starting point might be to work just with a "value", and to send just the min and max, rather than the full data set.

[emiltin]

If we want to make this technology agnostic, so that it's useful for any detector, like camera, loop and radar, then perhaps we should provide a way to read what technology is used, to help you make sense of the data.

[emiltin]

Loop: signal strength Radar: signal strength Camera: confidence?

What kind of statistics can tell you what you need, without looking at all the raw data? min max median spread