Skip to content

feat(server): radio-tomographic link-line estimator (library only) - #1809

Open
clonea1 wants to merge 2 commits into
ruvnet:mainfrom
clonea1:contrib/server-rti
Open

clonea1 wants to merge 2 commits into
ruvnet:mainfrom
clonea1:contrib/server-rti

Conversation

@clonea1

@clonea1 clonea1 commented Sep 4, 2026

Copy link
Copy Markdown
Contributor

Adds rti.rs: a grid solver that treats each transmitter-receiver link as an
ellipse and accumulates per-cell evidence from links whose response deviates
from their own resting level. 826 lines, 22 tests.

Pure functions over coordinates and floats -- it takes observations and a
config and returns an estimate, with no dependency on any other module here.
The adapter that feeds it live link metrics is deliberately NOT included,
because it needs both the per-link table and the room footprint, neither of
which is upstream yet. This lands the estimator and its tests; the consumer
follows.

HONEST LIMITATION, stated because it would be easy to omit. The elliptical
link-line kernel is the standard formulation, and on this deployment it did not
hold: measured 2026-08-28, per-link response was INVERSELY related to link-line
distance -- the opposite of what the model predicts. That measurement was taken
in a single room with three co-located nodes and does not necessarily
generalise to a distributed fleet, so it falsifies the model in that
configuration rather than in general.

The code is therefore offered as a correct and tested implementation of a
published technique, not as a demonstration that the technique works in a home.
Anyone adopting it should validate the kernel against their own geometry before
trusting an estimate, and the normalise_response/temporal_authority helpers are
useful independently of whether the ellipse model holds.


Rebased onto current main before opening: staged before today's seven merges, so it needed replaying to avoid reading as a revert of them. Clean rebase, no files deleted.

Joe and others added 2 commits September 4, 2026 15:02
Adds rti.rs: a grid solver that treats each transmitter-receiver link as an
ellipse and accumulates per-cell evidence from links whose response deviates
from their own resting level. 826 lines, 22 tests.

Pure functions over coordinates and floats -- it takes observations and a
config and returns an estimate, with no dependency on any other module here.
The adapter that feeds it live link metrics is deliberately NOT included,
because it needs both the per-link table and the room footprint, neither of
which is upstream yet. This lands the estimator and its tests; the consumer
follows.

HONEST LIMITATION, stated because it would be easy to omit. The elliptical
link-line kernel is the standard formulation, and on this deployment it did not
hold: measured 2026-08-28, per-link response was INVERSELY related to link-line
distance -- the opposite of what the model predicts. That measurement was taken
in a single room with three co-located nodes and does not necessarily
generalise to a distributed fleet, so it falsifies the model in that
configuration rather than in general.

The code is therefore offered as a correct and tested implementation of a
published technique, not as a demonstration that the technique works in a home.
Anyone adopting it should validate the kernel against their own geometry before
trusting an estimate, and the normalise_response/temporal_authority helpers are
useful independently of whether the ellipse model holds.

Co-Authored-By: claude-flow <ruv@ruv.net>

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant