Multi-Modality Cooperative Ranging

Mechanism: Fused Multilateration Over Heterogeneous Range Observations

The mechanism specifies that positioning is performed as a multilateration over a heterogeneous set of credentialed range observations rather than as a primary-modality estimate corrected by secondary modalities. Each ranging modality contributes observations into a uniform observation type whose payload carries the measured range, the modality identifier, the observing unit's identity and credential, the timestamp, the declared uncertainty distribution, and an environmental context tag. RF time-of-flight modalities (ultra-wideband, narrowband ranging, two-way ranging variants) contribute sub-decimeter to meter-scale ranges between cooperative units. Optical modalities (fiducial range against credentialed markers, structured-light range, lidar reflection against marked surfaces) contribute ranges with their own declared uncertainty profiles. Acoustic modalities (audible and ultrasonic echo, time-difference-of-arrival from acoustic emitters) contribute ranges robust to RF and optical denial. Additional modalities (radar, RFID proximity, NFC adjacency, BLE RSSI, magnetic dipole, GNSS pseudorange, inertial integration, visual SLAM correspondence) integrate as additional contributors under the same observation type.

Each contribution is governance-credentialed at the source. The contributing unit signs the observation with its credentialing chain, declares the range modality and the uncertainty model under which the range was produced, and records the observation into the substrate's lineage. The receiving unit — the unit performing the multilateration — evaluates each observation for admissibility before integrating it into the coordinate solution. Admissibility evaluation considers the contributor's credentialing chain, the freshness of the observation against the modality's declared decorrelation time, the consistency of the declared uncertainty with the modality's nominal envelope, and the absence of credentialing-chain anomalies that would indicate spoofing.

The fusion proceeds as a weighted multilateration in which each admissible observation contributes a residual whose weight is the inverse of the declared uncertainty's variance, modified by a per-modality confidence factor that captures the modality's expected reliability under the current environmental context. The solution minimizes the weighted sum of residuals subject to consistency constraints across modalities and produces both a point estimate and a covariance from which the per-modality contribution to the final uncertainty can be decomposed. The decomposition is itself recorded into the lineage, so that any downstream consumer can resolve which modalities most strongly influenced the final position.

Per-modality confidence is dynamic. A modality's confidence factor is updated continuously based on observed agreement and disagreement with other admissible modalities, on contextual signals indicating denial or degradation (RF interference signatures, optical occlusion, acoustic noise floors), and on credentialed broadcasts from oversight functions warning of regional adversarial conditions. A modality whose confidence falls below an admissibility floor is suspended from contributing to the solution, although its observations continue to be recorded into the lineage as a diagnostic record of why the suspension occurred.

Per-observation credential evaluation operates alongside per-modality confidence. Even within an admissible modality, an individual observation may be rejected if its credentialing chain fails to validate, if the contributing unit's credential has been suspended, or if the observation's uncertainty declaration falls outside the modality's nominal envelope in a way that is not consistent with the contributor's environmental context. Rejections are recorded with reasons; the lineage therefore reflects both the observations that contributed to the solution and the observations that were considered and rejected.

Operating Parameters: Modality Bounds, Confidence, and Cross-Checks

Modality-bound parameters specify, per modality, the nominal range envelope, the nominal uncertainty envelope as a function of range and environmental context, the decorrelation time beyond which an observation is considered stale, and the credentialing tier required of contributors. RF time-of-flight modalities operate within tens of meters at sub-decimeter uncertainty under nominal RF conditions; optical fiducial-range modalities operate within line-of-sight at uncertainties bounded by the fiducial's geometric resolution; acoustic modalities operate within the medium's effective propagation envelope at uncertainties bounded by the medium's temperature- and pressure-dependent sound speed.

Confidence-factor parameters specify how rapidly a modality's confidence factor responds to agreement and disagreement signals, the floor at which a modality is suspended, and the procedure by which a suspended modality is restored. Restoration may be automatic (after a sustained interval of consistency with admissible peers), credentialed (after an oversight broadcast clearing the modality's region), or contingent (only when an alternative diagnostic confirms the cause of the suspension has cleared).

Cross-modality consistency parameters specify the agreement thresholds between modalities under nominal conditions and the disagreement thresholds beyond which a credentialed disagreement event is emitted into the lineage. A disagreement event is itself a structured observation that downstream consumers — diagnostic functions, security functions, oversight bodies — may use to differentiate among sensor failure, environmental anomaly, and adversarial interference. The architecture does not silently absorb disagreement; it surfaces disagreement as evidence.

Solution-update parameters specify the rate at which the multilateration is recomputed, the policy for incorporating delayed observations from slow modalities (acoustic at long range, for example), and the smoothing applied across consecutive solutions. A high-update-rate consumer (a vehicle controller) receives solutions at the rate of the fastest contributing modality with declared uncertainty inflated to reflect the freshness of each contribution; a low-update-rate consumer (a logistics tracker) receives solutions at a slower rate with all admissible modalities fully integrated.

Credentialing-tier parameters specify the standing required to contribute observations within a deployment. Civilian deployments may admit commodity contributors at a baseline tier; defense deployments may require contributors at a defense-credentialing tier with hardware-rooted attestation. Mixed deployments admit both, but the multilateration weights each contribution according to its tier so that high-tier contributors dominate the solution within their range envelope.

Alternative Embodiments

The fusion mechanism admits embodiments at the estimator layer ranging from batch weighted-least-squares (suitable for offline post-processing of credentialed logs) through extended Kalman filter formulations (suitable for online vehicle-class deployments) to factor-graph formulations with marginalization (suitable for long-running cooperative-mesh deployments where historical observations remain relevant). Each embodiment preserves the structural commitments to per-observation credentialing, per-modality confidence, and lineage decomposition.

Modality-set embodiments range from minimal sets (two RF modalities and one optical, suitable for indoor cooperative robotics) through standard sets (RF time-of-flight, optical fiducial, acoustic, and inertial, suitable for outdoor mixed-environment deployments) to extended sets (the fifteen-plus modalities including radar, RFID, NFC, BLE, magnetic dipole, GNSS, visual SLAM, and emerging visible-light positioning, suitable for resilient defense and contested-environment operations). The architecture is open to additional modalities; new modalities integrate by declaring an uncertainty model and a credentialing-tier mapping.

Cooperative-topology embodiments include peer-to-peer (each unit ranges to and from its peers and computes its own position), anchored (a set of credentialed anchor units provides ranges to mobile units that compute their own positions), and infrastructure-mediated (a credentialed infrastructure node aggregates ranges and produces solutions for subscribing units). Mixed topologies are common; a defense deployment may use peer-to-peer ranging within a unit and infrastructure-mediated ranging at the formation level.

Confidence-update embodiments range from local (each unit updates per-modality confidence based on its own observations) through coalition (units within a credentialed coalition share confidence updates), to broadcast (oversight functions broadcast credentialed regional confidence updates that all units in the region honor). Broadcast embodiments are particularly relevant where regional adversarial conditions — GNSS jamming, UWB interference, optical denial, lidar blinding — affect many units and where coordinated suspension is more reliable than independent local detection.

Disagreement-handling embodiments range from passive (disagreement is recorded but does not change the solution) through reweighted (disagreement reduces the contributing modalities' weights), to escalated (disagreement above a threshold triggers a credentialed diagnostic event that may suspend a contributor or invoke oversight). Escalated embodiments are appropriate where adversarial interference is anticipated; passive embodiments are appropriate where disagreement is expected to reflect benign environmental anomalies and the cost of suspension is high.

Composition With the Wider Mesh-Coordinates Architecture

Multi-modality cooperative ranging composes with the credentialed-anchor layer of the mesh-coordinates architecture. Anchor units broadcast their identity and position with credentialing chains; mobile units range to anchors and to each other; the multilateration consumes both anchor-mediated ranges and peer-mediated ranges under the same observation type. The architecture does not privilege anchor-mediated ranges structurally; it weights them by their declared uncertainty and credentialing tier in the same way as any other contribution.

Composition with the position-lineage layer is direct. Each computed position is recorded into the lineage with its supporting observations, the per-modality decomposition of its uncertainty, and the credentialing chains of all contributors. Downstream consumers — collision-avoidance systems, allocation systems, audit functions — resolve a position's lineage to evaluate whether the position meets their reliability requirements before relying on it. A position whose lineage shows dominant contribution from a modality the consumer does not trust is rejected at the consumer rather than relied on silently.

Composition with the operator-intent layer governs how positions are used in adverse actions. A position used as evidence in an adverse classification (geofencing, restraint, restriction-of-movement) propagates through the operator-intent due-process credentialing chain; the position's lineage is part of the evidentiary record exposed to the subject under the right-of-review channel. Positions that depend on uncredentialed modalities, or whose per-modality confidence is below an admissibility floor for the proposed action's gravity, are inadmissible at the operator-intent gate.

Composition with the semantic-discovery substrate is achieved through positional observations as first-class memory records. A discovery object investigating a movement pattern, a sensor anomaly, or a denial event consumes positional observations and their lineages directly; the cognitive fields that depend on positional evidence inherit the credentialing constraints of the underlying ranges and the per-modality confidence at the moments of measurement.

Prior-Art Landscape

Single-modality positioning systems — GNSS-only, UWB-only, visual-SLAM-only, lidar-only — face structural failure modes when their primary modality is denied or degraded. GNSS jamming and spoofing produce regional outages and adversarial position injection; UWB interference disables short-range cooperative localization; optical denial through smoke, fog, or active blinding disables visual and lidar contributions; acoustic denial through sustained noise disables echo-based ranging. Single-modality hardening through anti-jamming antennas, anti-spoofing receivers, or environmentally robust optics raises the cost of denial but does not eliminate the structural single-point-of-failure.

Sensor-fusion systems combining a primary modality with secondary corrections — typically GNSS plus inertial, or visual SLAM plus inertial — partially address single-modality failure but treat the secondary modality as a corrector rather than as a peer contributor. When the primary modality is denied entirely, the corrector cannot sustain the solution; the system degrades to dead reckoning whose error grows unbounded. Multi-sensor Kalman-filter formulations admit multiple sensors as peers but typically lack per-observation credentialing, per-modality confidence broadcasts, and lineage decomposition; they treat sensor identity as a static assumption rather than as a credentialed property subject to admissibility evaluation.

Cooperative localization systems in robotics and ad-hoc networks have long combined ranges among peers but operate within trust assumptions that do not hold under adversarial conditions. A cooperative position derived from peers whose credentialing cannot be evaluated is vulnerable to a spoofing peer that contributes consistent but adversarial ranges. The architecture described here addresses this directly by making credentialing a first-class admissibility property of every observation and by supporting credentialed-disagreement events as evidence rather than as anomalies to be smoothed.

Defense PNT (positioning, navigation, and timing) initiatives explore alternative modalities — terrestrial broadcast PNT, satellite L-band PNT, magnetic-anomaly navigation, celestial navigation — as backups to GNSS. The architecture described here is compatible with these initiatives: each alternative modality integrates as an additional credentialed contributor under the same observation type, and the multilateration produces a composite position that draws from whichever modalities remain admissible under current conditions.

Disclosure Scope

The disclosure scope, supported by U.S. Provisional Application 64/049,409, covers fused multilateration over heterogeneous credentialed range observations spanning RF time-of-flight (UWB, narrowband, two-way ranging), optical (fiducial, structured light, lidar reflection), acoustic (audible and ultrasonic echo, TDoA), radar, RFID, NFC, BLE RSSI, magnetic dipole, GNSS pseudorange, inertial integration, visual SLAM correspondence, and emerging modalities including visible-light positioning, satellite L-band PNT, and terrestrial broadcast positioning; per-observation credentialing with admissibility evaluation against contributor standing, freshness, uncertainty consistency, and credentialing-chain integrity; per-modality dynamic confidence with local, coalition, and broadcast update embodiments; cross-modality consistency evaluation with credentialed disagreement events surfaced into the lineage; estimator embodiments spanning batch weighted-least-squares, extended Kalman filter, and factor-graph formulations; cooperative-topology embodiments spanning peer-to-peer, anchored, and infrastructure-mediated structures; and composition with credentialed anchors, position lineage, operator-intent due-process credentialing, and semantic-discovery substrates.

Defense and contested-environment operations gain structural resilience that single-modality hardening cannot match: loss of any single modality reduces position confidence but does not eliminate it, and credentialed disagreement among modalities surfaces as diagnostic evidence rather than as a silent source of error. Civilian deployments in challenging environments — urban canyons, dense indoor settings, mining operations, port and intermodal logistics — gain the same resilience under the same architecture. The architecture supports gradual modality adoption: emerging modalities integrate as additional credentialed contributors without architectural rebuild, allowing deployments to extend their resilience as new sensing technologies become available.