PORRITTINC
( LAYTON, UT )
POSITION PAPER · NEUROMORPHIC R&D

A reflex
layer for
the plant.

Biology separates reflex from deliberation. We think a plasma-era balance of plant needs the same split: a fast protective tier beneath a slower supervisory layer, with human judgment above both.

← Back to Reflex Layer for BoP

POSITION PAPER · TECHNICAL OPINION

This is a technical position paper stating Porritt Inc.'s engineering view. It is not a peer-reviewed publication and does not report experimental results. Where it forecasts capability, it is stating design intent rather than achieved performance.

Position
GAP

Between relay and software

Machine protection relays are fast but blunt. Supervisory software is capable but slow. There is a band of events that need both speed and context, and nothing occupies it well.

PROPOSAL

A bounded reflex tier

A narrow, deterministic, independently verifiable tier that acts in microseconds on a small set of well-characterised conditions — and does nothing else.

DISCIPLINE

Narrow by design

The value of a reflex tier comes from being small enough to verify exhaustively. Scope creep would destroy the property that makes it trustworthy.

HONEST LIMIT

Verification is the hard part

Deterministic verification of a learned function is an unsolved problem. A reflex tier may need to be conventionally programmed and merely informed by learning.

Objectives
μs
Reflex objective
BOUNDED
Scope
HUMAN
Above both tiers
R&D
Not deployed
Open questions

Whether a learned classifier can ever be certified for a protective function, or whether the reflex tier must remain conventionally programmed, is the central open question. Our current view is the latter, with learning confined to advisory roles.

Discuss the research.

We welcome technical challenge on these positions. Collaboration enquiries welcome.

CONTACT TIMOTHY → Back to Reflex Layer for BoP →