Sense selectively
Use opt-in or otherwise authorized motion signals and permitted facility records. Device participation and signal availability are never assumed universal.
MINIMUM NECESSARY INPUT A CONCEPT FOR BUILT ENVIRONMENTS / OCT 2026
Know the building.
Not every movement.
A different approach to facility awareness: fuse device motion, spatial context, and existing operational data into periodic, confidence-aware snapshots of who may be on-site, where capacity is strained, and what remains unknown.
Independent research concept · synthetic demo · no live tracking or government affiliation
SPATIAL INTELLIGENCE / CONCEPT 01● MODEL ONLY
DATA CAPTURESensors + mapsConstrained inference
OUTPUTZone censusAggregated · time-bounded
INFERRED, NOT ASSERTEDv0.1 / THESIS
THE THESIS
Turn uncertain spatial signals into operational evidence—without making individualized surveillance the default.
SCROLL TO EXPLORE ↓01 / THE CONCEPT
From a location trail
to a useful operational snapshot.
Independently Derived Positioning System (IDPS) takes inspiration from inertial navigation and map-constrained positioning. The proposed building adaptation turns imperfect, sometimes intermittent signals into bounded, auditable estimates at the resolution an operator actually needs.
Constrain by space
Use floor plans, walkable pathways, doors, floors, and work-center boundaries to improve inference where satellite positioning is unreliable.
MAP-AWARE REASONINGObserve in snapshots
Generate time-boxed, zone-level census estimates rather than preserving a constant feed of every person's movements.
PERIODIC BY DESIGNShow what is unknown
Preserve ranges, missing coverage, stale observations, and reconciliation gaps. Unobserved never automatically means absent.
EVIDENCE WITH LIMITSCORE DESIGN RULE
The appropriate output is not necessarily an individual's coordinates. For many operational decisions, a reliable zone-level count with a visible uncertainty band is more defensible and more useful.
See the trust model02 / INTERACTIVE LABORATORY
A building that
answers back.
Explore how operating conditions, participation, uncertainty, and snapshot timing affect the workforce picture. Every number below is generated from a synthetic scenario.
SIMULATION MODENo live devices · no personal data
NORMAL OPERATIONS
Illustrative weekday ebb and flow with routine clinic, inpatient, and support staffing.
Designed to test visibility—not individual productivity.MODELED ON-SITE PRESENCE ⓘ—Illustrative uncertainty range
SCHEDULED REFERENCE—Illustrative roster baseline, not ground truth
STRONG OBSERVATIONS—Signal-supported, not identity tracking
ZONES NEEDING REVIEW—Demand / uncertainty tests
BUILDING VIEW / SYNTHETIC ZONES
Main medical center
Spatially aggregated model, not a live floor plan.
SELECT ANY ZONE
SCHEMATIC / NOT TO SCALE09:30 SNAPSHOT
Lower load Moderate ElevatedColor = modeled presence / nominal capacity
TIME SERIES / 30-MINUTE STEPS
One shift. Many possible pictures.
ASSUMPTIONS / ADJUSTABLE
Change the evidence quality.
82%
Share of synthetic personnel with usable authorized signals.18%
Drift, floor ambiguity, and weak map/anchor matches.Models capture timing; this demo interpolates 30-minute scenario steps.
ASSUMED EFFECTIVE CAPTURE—Input-derived visibility proxy, not proven positional accuracy.
CENSUS RECONCILIATION / AGGREGATE ONLY
At a glance, with uncertainty intact.
| ZONE / WORK CENTER | DIRECT SIGNAL | MODELED | ESTIMATE RANGE | TARGET | ASSESSMENT |
|---|
✳
Important: this is an interactive hypothesis tester, not a functioning indoor-positioning product. The engine creates synthetic staffing truth, simulated detection losses, and a simplified estimator. It does not connect to sensors, identity systems, DMHRSi, or government networks; it is not evidence of measured performance.
03 / REFERENCE ARCHITECTURE
The intelligence is in
the fusion and restraint.
The original IDPS inspiration—derive position from onboard motion and a constrained map—becomes a proposed building-specific, multi-source evidence pipeline. Each handoff is inspectable and governed.
01
Permitted inputs
Phone or wearable IMU, optional BLE / Wi-Fi anchors, access events, shift roster
AUTHORIZATION + SOURCE TAGS→
02
Spatial model
Floor plans, zones, passages, floors, facility geofence, work-center mapping
VERSIONED FACILITY TOPOLOGY→
03
Inference
Pedestrian dead reckoning, map matching, optional anchor correction, uncertainty
NO RAW TRAILS BY DEFAULT→
04
Snapshot engine
Time window, deduplication, zone aggregation, minimum cohort size, bounds
EPHEMERAL PROCESSING→
05
Operational view
Zone census, stale/missing flags, roster reconciliation, demand gaps
AUDITABLE DERIVED OUTPUTSOURCE REALITY
Not every signal means the same thing.
01 / Motion sensors
Relative movement; susceptible to drift and device carriage differences
NO ABSOLUTE POSITION02 / Map constraints
Physically plausible paths; depends on accurate, updated floor plans
MAP FRESHNESS03 / Optional anchors
Correct accumulated error; may add infrastructure and privacy complexity
COST / COVERAGE04 / Operational records
Rostered and checked in are not the same as present or available
SEMANTIC MISMATCHOUTPUT CONTRACT
Every snapshot should answer four questions.
01
What is known?Observed evidence with time and provenance
02
What is estimated?Model version, assumptions, and uncertainty bounds
03
What is missing?Coverage, staleness, and unresolved people or areas
04
Who may act?Authorized role, allowable purpose, audit record
OBSERVABILITY ≠ OMNISCIENCE
04 / WHERE IT MAY MATTER
A facility problem.
A cross-sector capability.
The first compelling wedge is military medical workforce visibility. The underlying capability—bounded, privacy-aware facility census—could also support other buildings where staffing, access, and operational demand shift.
01 / PRIMARY PILOT↗
Military treatment facilities
Periodic staff footprint by work center; visible gaps between planned, detected, and uncertain presence during ordinary operations, pandemics, deployments, or reassignment.
- Work-center workforce census
- Surge and continuity planning
- Roster cross-checks—not timesheets
PRIORITY OPPORTUNITY →
02 / ADJACENT↗
Civilian hospital campuses
Staffing footprint and zone pressure across wards, emergency departments, diagnostic spaces, and nonclinical services.
WORKFORCE VISIBILITY
03 / ADJACENT↗
Critical facilities
Privacy-respecting shift footprint and contingency awareness for public facilities, labs, and essential workplaces.
OPERATIONAL CONTINUITY
04 / ADJACENT↗
Emergency coordination
Time-bounded information to support drills, accountability reconciliation, and resource staging—with dedicated emergency safeguards.
RESPONSE SUPPORT
05 / THE MHS USE CASE
Not a replacement for everything.
A better question for census.
DMHRSi is a broad MHS human-resources platform—not simply a facility-tracking application. IDPS would first have to demonstrate value as a narrowly scoped personnel-presence evidence layer, complementing or potentially replacing parts of current census compilation. It does not independently establish duty hours, tasks performed, labor costs, or clinical readiness.
▤ SYSTEM OF RECORD / ESTABLISHED
DMHRSi
Defense Medical Human Resources System–internet
- Manpower and personnel management
- Time and attendance / labor accounting
- Training and readiness records
- Work-center and HR reporting
Official DHA fact sheet describes broader functions; these are not replicated by the proposed IDPS model.
◉ PROPOSED / TO BE VALIDATED
IDPS census layer
Derived facility-presence evidence
- Periodic zone-level on-site estimates
- Observed / modeled / unresolved distinctions
- Uncertainty and data-quality indicators
- Roster reconciliation for authorized operators
Potentially augments or displaces a specific census workflow only after controlled performance and governance testing.
THE ACTUAL PROBLEM
A 2025 GAO review identified inaccurate and incomplete medical personnel timecard data and distinctions missing from workforce monitoring. Location estimates would not repair timecard activity codes, prove time worked, or resolve command-assignment rules. The opportunity is a different kind of evidence: a better bounded picture of physical facility presence.
06 / PRIVACY + GOVERNANCE
Make restraint
an architectural property.
Privacy should not live in a policy footnote. Make it visible in the data contract, the default interface, the retention schedule, and the authority to access exceptional detail.
01COLLECT LESS
Purpose-bound collection
Facility operations only; authorized devices, covered hours, and geofenced spaces. No ambient audio or covert person-monitoring.
02EXPOSE LESS
Aggregate-first output
Suppress small cohorts and unnecessary role breakdowns; do not display individual paths in the routine dashboard.
03AUTHORIZE
Role and purpose controls
Separate facility operations from personnel administration; individual access requires an explicit legal basis, authorization, and audit.
04EXPIRE
Bounded retention
Favor short-lived raw readings, defined snapshot retention, encryption, and deletion verification.
05VERIFY
Confidence and fairness testing
Measure error by device, job role, shift, floor, and facility; do not silently treat missing signals as noncompliance.
06GOVERN
Documented labor and legal review
Resolve workforce notice, consent where applicable, federal privacy obligations, bargaining implications, security authorization, and incident response before a pilot.
AGGREGATION CONTROL / PRINCIPLE
QUERY: ON-SITE STAFF BY ZONE
Emergency Department34–41 EST.
Imaging19–26 EST.
Small specialty teamSUPPRESSED
⌑
Default disclosure: aggregate count, bounded estimate, freshness, and confidence. Small-cohort suppression shown here is illustrative. A production threshold would be approved by the privacy authority.
NO ALWAYS-ON TRAILS
NO PERSON-LEVEL SCORECARDS
NO ABSENCE INFERENCE FROM SILENCE
NO PERSON-LEVEL SCORECARDS
NO ABSENCE INFERENCE FROM SILENCE
07 / VALIDATION PATH
Don't sell certainty.
Run the experiment.
The defensible next step is an independently evaluated, synthetic-first pilot with a small approved real-world cohort only after operational, privacy, security, labor, and records-management review.
PHASE 00 WEEKS 1–4
Define the truth set
Choose one facility and limited work centers. Establish approved use, baseline process, representative shifts, floor-plan fidelity, and independent manual/reference census methods.
DELIVERABLE → STUDY PROTOCOL + RISK REVIEW
PHASE 01 WEEKS 5–8
Test the positioning hypothesis
Bench-test inertial drift, map constraints, floor transitions, device carriage, sparse anchors, nonparticipation, and cohort suppression with synthetic and consented test data.
DELIVERABLE → ERROR + COVERAGE REPORT
PHASE 02 WEEKS 9–12
Shadow the census workflow
Generate periodic snapshots beside the approved existing process. Compare zone counts, missing coverage, operator workload, and decision usefulness; prohibit operational reliance until validated.
DELIVERABLE → GO / REFINE / STOP DECISION
MEASURE
What would prove it?
Zone count error vs. independent referenceCompleteness & missingnessFalse absence and drift eventsPerformance by floor / device / shiftTime-to-snapshot & analyst burdenPrivacy-control failures / exceptions
Pilot acceptance thresholds must be pre-registered by accountable stakeholders. No accuracy, savings, safety, or regulatory-compliance results are claimed by this prototype.
08 / NEW QUESTIONS WORTH ASKING
The most interesting shift
is not the positioning.
The larger opportunity may be a new type of operational evidence—designed around what decisions require, what inference can support, and what uncertainty must remain visible.
Unknown is a first-class state.
A scheduled employee with no sensor signal should not become a falsely declared absence. Make unresolved coverage explicit rather than silently completing the picture.
Workforce presence is not workforce readiness.
Physical presence, assigned work center, licensure, privileges, duty status, and availability are different questions. Useful operational systems join them without confusing them.
Snapshots can be a design advantage.
A lower-frequency question—how many people are plausibly here now?—may require less invasive processing than persistent trajectories, while still supporting periodic planning.
Evidence quality belongs in the product.
Operators should see freshness, source mix, uncertainty, and a path to challenge a count. Trust is earned through inspectable limitations, not a beautiful single number.
09 / EVIDENCE + BOUNDARIES
Grounded in research.
Open about what is unproven.
These sources inform the problem statement and technical feasibility. They do not validate this particular IDPS adaptation, simulator, or any claim of performance in military treatment facilities.
OFFICIAL SYSTEM CONTEXT / DHA 2025
DMHRSi fact sheet
Defines DMHRSi's broader HR, manpower, readiness, time/attendance, and labor accounting roles.
READ SOURCE ↗GOVERNMENT AUDIT / GAO 2025Defense health workforce staffing
Documents incomplete timecard inputs and challenges monitoring military personnel availability.
READ SOURCE ↗TECHNICAL REVIEW / IEEE 2024Indoor pedestrian dead reckoning
Reviews sensor fusion, infrastructure dependencies, and navigation challenges for smartphones.
READ SOURCE ↗PRIVACY / NISTPrivacy Framework
Risk-management framework informing responsible collection, governance, and system design.
READ SOURCE ↗ESTABLISHED
Indoor positioning can combine inertial sensors and map constraints. DMHRSi is a broad MHS HR platform. Staffing data quality issues are documented.
PROPOSED
Adapt those methods to produce periodically refreshed, aggregate facility-staff census estimates with explicit uncertainty and governance.
NOT YET PROVEN
Accuracy under real building conditions, opt-in coverage, comparative costs, clinical decision benefits, legal approvals, and viability as a replacement census workflow.
THE NEXT STEP
Build the evidence
before the system.
Start with a narrow workforce-census question. Establish the truth set. Test the difficult assumptions. Then decide what deserves to be built.
Return to simulation