A1102 utility of LoraWan?

Products & Technology SenseCAP
1

Hi, I’m using sencecap A1102 with person detection and connected to LoRaWan and receive some data like

    measurementId:"4165"
    measurementValue:0.65
    type:"person Conf"

measurementValue:"Person Detection--Swift YOLO"
measurementId:"4174"
type:"Model Type"

With Person count and connected to LoRaWan I receive some data like

valid:true

        measurementValue:"Person Detection"
        type:"Model Type"
        measurementId:"4174"

err:0

I send data every 5mn to LoRaWan.

What is the advantage to send data to LoRaWan ?
There is no information received on TTN for person detect and no information of person count …

Or maybe I do somethong wrong …

My purpose is to collect data with lorawan to count people.
A1102 needs to save data add people count send total very x minutes put count to zero and loop.

2

Hi Cobalt
I just tried to set up my own A1102…Not succesful to join the TTG platform
I entered Join EUI/APP KEY bit device is not registered to TTN. May I ask if you faced problems to connect to The Things TTN. OR did you use SensCAP Platform? I want to use the device for People Counting the same as you. If I can get my device working I will pass my ideas to you.

3

I now have my own A1102 reporting to The Things LORAWAN,
I get similar results and it does not recognise 2 persons!
My payload raw is
“valid”: true
301205FFFFFC18FFFFFC183234FFFFFC18FFFFFC183256FFFFFC18FFFFFC183278FFFFFC18FFFFFC18339A05FFFFFC183F3202C0

The decode is
measurementId": “4174”,
“measurementValue”: “Person Detection”,
“type”: “Model Type”
}

Absolutely useless!

4

Look at how the sensor formats the payload and check TTN’s payload decoder. You may need to modify the decoder to correctly interpret the data, especially for person count and detection.

5

I did analyse the payload format. This are only 3 types :

model_table = {
        '60086': {
            'modelName': 'Person Detection--Swift YOLO',
            'task': 'Detection',
            'classes': {
                "0": "person"
            }
        },
        '60113': {
            'modelName': 'Digital Meter Electricity',
            'task': 'Detection',
            'classes': {
                "0": "zero", "1": "one", "2": "two", "3": "three", "4": "four",
                "5": "five", "6": "six", "7": "seven", "8": "eight", "9": "nine"
            }
        },
        '60242': {
            'modelName': 'Person Detection',
            'task': 'Detection',
            'classes': {
                "0": "Person"
            }
        }

The payload detect which model are in the camera if none of the 3, the model is “unkown” like face recognition, bike, …

For exemple in CAmerica the result :

[{'dataId': '30', 'dataValue': '1205FFFFFC18FFFFFC18'}, {'dataId': '32', 'dataValue': '34FFFFFC18FFFFFC18'}, {'dataId': '32', 'dataValue': '56FFFFFC18FFFFFC18'}, {'dataId': '32', 'dataValue': '78FFFFFC18FFFFFC18'}, {'dataId': '33', 'dataValue': '9A05FFFFFC183F3202C0'}]
============================
[[{'measurementValue': 'Person Detection', 'measurementId': '4174', 'type': 'Model Type'}]]

No more information.

I asked to Seeep SenseCap people more information on S2100 … No answer

It seems that they put on a market something look nice on paper but nothing more and no help !

6

Hi there,

OK, so slow down with that…
I have been following this thread and I don’t think it’s been sufficiently diagnosed and tested to make such a stark determination.
“pretty but not supported”

You are NOT the first customer of this SO understand 2 things. It’s new tech and NOT all the functions are understood, the info available needs updating, more changes are in the process as well as some WiKi updates. Be patient and give them time to deduce the solution for your issue. Meanwhile can you Say it detects people accurately ? and can it decerne with HIGH confidence from a bicycle or a dog ? How did you train the model ? Did you train it?

I can tell you from experience’s with both , the Unknown category needs 3 times as many samples, Boosts the confidence of known people’s.

Can I ask , Does the model detection work? before you add in the LORA portion and the whole TTN business ( I frankly don’t see the point of it? ) Probably because I’m not understanding the Whole purpose.

So there are a lot of smart folks on here willing to help be patient and they will chime in more. AFAIK there are almost NO complaints about the quality of the tech and operation so , hang in there. You’ll get it going.

Can you post your question here so we can analyze please?

HTH
GL :slight_smile: PJ :v:

IMO, it’s a good platform and great hardware , but there is a learning curve with all it can do, it crosses over into many areas of tech. :crossed_fingers:

7

I only used the available models. For example, counting people. We could expect there to be feedback on the number of people seen, sent with LoRa.
There is no information on the number of people. Just information on the model used.
We are waiting for support to be able to retrieve information via LoRa. Our goal is to count the number of bikes on a cycle path, retrieve the partial total every hour in order to make a count throughout the year with an autonomy of the whole of several months as indicated on the site.

The best is to put on the web side “Product in beta version” as the result is not what people expect.
We are using other product like LoRa gateway and not problem with it.

I am keeping hope for the A1102 product which can a great product as the wiki explain

:wink: