Rssi to distance conversion

[simonasaitta]

As continuation of issue #33 we need to do some tuning to this conversion. The position formula seems to correctly work now (99f1655) but not the conversion one.
We could try to:

• keep changing the N and 1 meter value until we do get feasible results
• find a new formula
• experiment with the rssi values returned by the board to check out if they are reliable by being constant at same distances (or with a small error window)

[marcomicera]

I was taking care of this but I found out a bug #56. I'll take care of that first, so we will have way more distance detections to work on.

[marcomicera]

In the meantime, I've added some more useful printouts.

[marcomicera]

#56 has been fixed. I'm back on this.

[marcomicera]

Here are my thoughts about the current formula.

My configuration

"boards": [
    {
        "mac": "...:68",
        "x": 0,
        "y": 50
    },
    {
        "mac": "...:c0",
        "x": 50,
        "y": 50
    }
],
"room_coordinates":[
    {"x": 0, "y": 0},
    {"x": 0, "y": 100},
    {"x": 100, "y": 0},
    {"x": 100, "y": 100}
]

My device "...:d1" is at x: 50, y: 0.

Some data

                   Board @(0,50)        Board @(50,50)
                      +---+                +---+
                      |:68|+--------------+|:c0|
                      +---+      50cm      +---+
                        +                    +
                        |                    |
                        |                    | 50cm
                70.71cm |                    | RSSI: -49/-41db
        RSSI: -68/-53db |                    | Avg. RSSI = -45db
      Avg. RSSI = -61db |                    |           +
                +       |                    +           |
                |       |                  +---+         |
                |       +-----------------+|:d1|         |
                |                          +---+         |
                |                       Device @(50,0)   |
                |                                        |
                |                                        |
                |                                        |
                v                                        v
      ONE_METER_RSSI should be -64db       ONE_METER_RSSI should be -51db

(Desmos example)
So basically the formula is not enough. I'll measure the average RSSI of various distances in order to derive some kind of relation between these two measures.

[simonasaitta]

Checkout this post

[marcomicera]

Here's our current guess compared to real measurements.

             RSSI (at 1 meter)  -  RSSI (measured)       
             -------------------------------------       
                             10 * n
distance = 10                                     * 100

With RSSI (at 1 meter) = -52dBm and n = 2.

[suidpit]

The idea to definitely test this:

• Let's measure, for each board, RSSI to one meter.
• Set that to be the constant we use on the server for each board.
• Change the position of the device (0.5 mts, 2.0 mts) and see if results on server are consistent.