Custom Anchor Self-Calibration Algorithm

[jonathanrjpereira]

@jopequ In your paper, you have mentioned that you have written your own custom anchor self-calibration firmware:

However, Decawave’s function to autoposition the anchors has not given good results in our experience. Moreover, the calculation takes around 40 s, which is unassumable in some mobile settings. In order to tackle this issue, we have written our own firmware for the anchors in order to recalculate their position if they move.

which is described as follows:

During the autopositioning process, the first anchor to become an initiator, which is activated via a start command through the UART interface, is considered the origin of coordinates. Then, we assume that some information about the position of the other anchors exists. The minimum information required is to know the order of the anchors over the boundary of their convex envelope in a counter-clockwise direction. We also predefine the x axis to follow the direction of the vector that is defined from the first to the second anchor following the aforementioned order.

Do you have any additional documentation/example code related to your custom anchor self-calibration algorithm?

[jonathanrjpereira]

@jopequ @kotovuori Does Convex Envelope refer to the term Convex Hull? If yes, then I have a few additional questions:

• What exactly do you mean by the term autopositioning? The anchor_positioning_firmware folder describes the Single-Sided Two Way Ranging Anchors functionality. Is forming an inter-anchor Euclidean distance matrix considered autopositioning?
• How do you use a Convex Hull to determine the positions of the anchors given that the input required to calculate a convex hull are the locations of the points itself?
• To find the order of the anchors over the boundary of their convex envelope in a counter-clockwise direction, wouldn't you need to know the locations of the anchors itself?