Vendor and Product Reality

Geek+ (Beijing Geekplus Technology), founded in 2015, is among the global leaders in AMR-driven warehouse fulfillment, with a product portfolio spanning the P-Series goods-to-person shelf-lift robots, the S-Series sortation robots, the M-Series moving-rack platforms, and the RoboShuttle case-handling robots. Customers include Nike, Walmart, DHL, Decathlon, Conforama, and Geodis, with deployments serving e-commerce fulfillment, third-party logistics, apparel distribution, and increasingly micro-fulfillment for grocery. The company's fleet management software, Geek+ RMS, orchestrates thousands of robots within a single facility under a centralized planner.

Architecturally, Geek+ robots localize primarily by reading QR-code floor markers laid out on a regular grid, with onboard inertial estimation between markers and vision or LiDAR for obstacle detection. The fleet manager assigns missions, plans paths through a graph of marker-defined waypoints, and arbitrates intersections through a centralized traffic controller. The system performs reliably at scale because the planner has full observability of every robot in its fleet and full authority over every lane segment in the warehouse.

The operational reality of contemporary fulfillment, however, is that warehouses no longer host a single fleet. A modern facility may run Geek+ goods-to-person robots in one zone, AutoStore cube-storage shuttles in another, conveyors and sorters from Vanderlande or Dematic in a third, and forklifts driven by humans throughout. Each subsystem has its own controller, its own marker scheme, QR codes, AprilTags, magnetic strips, RFID floor tags, and its own definition of who is allowed to traverse a given lane. Geek+ RMS governs Geek+ robots; it cannot govern, or even reliably predict, the trajectory of an AutoStore shuttle or a human-driven pallet jack.

The Architectural Gap

The structural property Geek+ does not provide is a substrate in which lane authority is credentialed and decentralized, allowing multiple fleet managers and human-supervised classes to share a common floor without any single party owning all lanes. Geek+ RMS treats the warehouse floor as the authority scope of the Geek+ planner. When another vendor's fleet enters that scope, the only available coordination mechanisms are physical separation (caged-off zones), schedule-based deconfliction (time windows), or coarse handoffs at zone boundaries, all of which sacrifice throughput and require manual layout engineering.

Equally, Geek+ does not natively fuse multiple marker classes under a single localization frame. A QR-code grid is the assumed fabric; AprilTags used by another vendor's robots, magnetic strips left over from a legacy AGV deployment, or UWB anchors used for high-value-asset tracking are not first-class participants in the Geek+ localization pipeline. The gap is the absence of a marker-track substrate that admits multiple marker classes, credentials lane segments to authorities other than the Geek+ planner, and routes traffic according to credentialed admissibility rather than centralized assignment.

What The Credentialed Marker and Track Provides

The Marker and Track substrate disclosed in the provisional rests on one primitive: an infrastructure-resident marker or sentinel emits a self-describing, authority-credentialed observation rather than a bare identifier. As disclosed, each marker's stored data carries an authority credential identifying the authority that installed or maintains the marker, a temporal-scope or time-to-live field establishing freshness, a spatial reference and local-geometry parameters, governance-policy-defined advisory parameters, a dynamic device hash encoding identity continuity, and a cryptographic attestation binding the data to the authority. A receiving unit resolves position and route by reading the observation and evaluating it through a governance policy against a published authority taxonomy, an outcome recorded in a lineage field. On top of that primitive the substrate provides three coupled behaviors.

The first, credentialed routing, treats every traversal as an admissible event under a credentialing authority that may differ for different lane segments. A traversal is permitted not because a central planner assigned it but because the credential presented by the moving entity satisfies the admissibility rule of the segment's authority, evaluated through the composite admissibility evaluator disclosed in the specification. Geek+ robots present their own credentials; another vendor's shuttles present theirs; humans present role-based credentials issued by the warehouse operator, which the specification lists as an explicit warehouse-domain authority taxonomy spanning facility-operations, zone-supervisor, shift-lead, and individual-operator authorities.

The second behavior, lane authority, partitions the floor into credentialed lane segments whose authority is explicitly named and may be delegated. A picking aisle may be authority-delegated to the Geek+ planner for the duration of a wave; a cross-dock corridor may be authority-shared between Geek+ and the human shift supervisor; a charging lane may remain under the warehouse operator's direct authority at all times. Authority transitions are themselves admissible events, recorded in the lineage layer so that any incident can be reconstructed in terms of who held authority where and when.

The third behavior, multi-class marker fusion, follows from the specification's signaling-mechanism-independent marker primitive and its sensor-fusion configuration. QR codes, AprilTags, magnetic strips, RFID floor tags, UWB anchors, and any other physically realizable marker class are treated as observation sources whose credentialed readings are fused into a single localization frame under the composite admissibility evaluator. Each marker class carries its own observer credential and evidential weight, and readings are fused into a coherent pose estimate that any participating fleet can consume. This is what would allow a Geek+ robot to localize against an AprilTag left by a maintenance contractor or a UWB anchor installed for a high-value-asset audit, without modification to the Geek+ vision stack.

The self-describing observation is also what makes the substrate resistant to spoofing and replay in a way a bare QR grid is not. Because each observation carries an issuing identity, a dynamic device hash tied to identity continuity, and a freshness bound, a receiving unit can reject a forged or stale marker reading rather than routing on it. Coverage is progressive: the substrate operates on sparse credentialed markers and degrades gracefully where density is low, so an operator can deploy credentialed markers incrementally in the lanes that matter first.

Composition Pathway

Integration with Geek+ proceeds through a substrate adapter running alongside Geek+ RMS, exposing the fleet's planned trajectories, occupied lanes, and current credentialing posture into the marker-track substrate. The adapter consumes RMS telemetry through Geek+'s integration APIs and republishes lane-occupancy and routing-intent observations under the warehouse operator's credentialing authority. Geek+ RMS continues to plan and dispatch its own robots; the substrate is informed of those decisions and admits or refuses them at the lane-authority layer.

Cross-fleet coordination is achieved by deploying analogous adapters in front of AutoStore controllers, Vanderlande WCS systems, and human-supervised equipment such as forklift telematics and warehouse worker badge systems. Each adapter republishes its native intent into the substrate under its own credential, and the substrate arbitrates lane authority across fleets according to admissibility rules the warehouse operator defines. A picking aisle's authority can be handed from Geek+ to a maintenance crew for a battery swap and back to Geek+ when the swap is complete, with every transition admissibly logged.

Multi-class marker fusion is enabled by installing additional marker classes in lanes where they are needed, AprilTags for vendor-neutral re-localization, UWB anchors for high-value-zone audit, magnetic strips preserved from legacy deployments, without disturbing the QR-code grid that Geek+ already relies on. The substrate fuses readings from all classes into a single frame, and Geek+ robots benefit transparently because the substrate publishes a fused pose that the Geek+ adapter can feed back into RMS as augmented localization input.

Commercial and Licensing Implication

A warehouse deployment that fields Geek+ AMRs alongside other vendors' fleets or human-supervised classes and that achieves credentialed routing, lane authority, and multi-class marker fusion implicates the subject matter disclosed in the provisional. The Geek+ robots and RMS are themselves unencumbered; the credentialed substrate that lets them coexist with other authorities on a shared floor is what the disclosure covers. Logistics integrators planning multi-vendor fulfillment and third-party logistics operators planning shared-tenant warehouses can evaluate the AQ portfolio before architecting such a substrate.

Adaptive Query offers field-of-use licensing keyed to warehouse and fulfillment operations, with terms calibrated to the number of credentialed fleets sharing the floor, the marker-class breadth, and the governed-event volume. Geek+ and peer fleet vendors retain full control of their robots and planners; the substrate is licensed to the integrator or end-user warehouse operator. Licensing is structured to permit pilot evaluation and to scale with production deployment, and is compatible with co-marketing arrangements between Adaptive Query and AMR vendors operating in governed logistics environments.

Enablement and Embodiment Scope

A skilled implementer can build the substrate from components the specification enumerates. A credentialed marker is any passive or active device whose stored data carries the authority credential, temporal scope, spatial reference, local-geometry parameters, dynamic device hash, and cryptographic attestation described above, realized through any of the disclosed signaling modalities: radio-frequency backscatter, optical or photonic patterns, acoustic resonators, magnetic-signature material, or chemical and spectroscopic signatures. Markers install as road studs, floor markers, wall markers, threshold strips, ceiling markers, platform-edge strips, or warehouse-aisle elements, among other form factors, and may combine a human-perceptible delineation function with the machine-readable observation in a single body. A receiving unit is any operating unit carrying a governance policy, a composite admissibility evaluator, and a lineage recorder; the attestation may use digital-signature, threshold-signature, zero-knowledge, or post-quantum mechanisms.

The architecture is deployable across topologies. It runs distributed, with each unit evaluating credentialed observations locally, or through a governance-credentialed central aggregator, or in a hybrid, with identical credentialing, admissibility, and lineage semantics in each. It is domain-general: the same primitive applies to roadway, warehouse, port, airfield, mining, retail, and campus environments, and it spans a fidelity spectrum from a floor equipped only with passive markers, to markers plus active sentinels, to a fully instrumented multi-tier mesh, degrading gracefully as marker density falls. This article is a dated public disclosure of that subject matter tied to the filing below.

Disclosure Scope

The credentialed marker and track substrate, the self-describing authority-credentialed observation, the composite admissibility evaluation, the lineage recording, the dynamic-device-hash identity continuity, and the progressive-density deployment described in this article are disclosed in U.S. Provisional Application No. 64/049,409. All statements about Geek+, its P-Series, S-Series, M-Series, and RoboShuttle robot families, its RMS fleet manager, its QR-code marker localization, and about other named vendors and the multi-vendor warehouse market are external context describing third-party products and the industry landscape. They are provided for accurate comparison only and are not claims of U.S. Provisional Application No. 64/049,409. Named products belong to their respective owners, and their mention does not imply endorsement or affiliation.