My research focuses on verifiable geospatial technologies and their relevance to the spatial governance of intelligent machines.
My thesis is that credible location-contingent commitments -- commitments to follow location-contingent policies, credible because compliance is mechanically verifiable rather than merely asserted -- are both useful and technically feasible.
These technologies are useful for:
The most effective credible commitments will require every input to be verifiable all the way back to the sensor or the issuing institution, and that requirement gives the work below its shape. Read the longer framing
High level writing that articulates the case for spatial governance technologies and defines key terms.
This research note will define the term this page is built on -- what makes a location-contingent commitment credible, and what it takes to construct one.
Astral: a spatial extension for the decentralized web — the FOSS4G NA 2025 talk introducing the Astral toolkit
Talk2025Some earlier artifacts here carry a "web3" or "decentralized web" framing because they were built to integrate with smart contract applications; much of the logic carries through to multi-agent systems. The longer framing explains the through-line; see also Web3 is a set of design principles.
The key enabling technologies involve reliably verifying the timing and location of devices, people and events.
Location cannot be "proven" in a mathematical sense; instead, verification requires the collection and evaluation of hard-to-forge location evidence that holds up in adversarial conditions.
This framework composes evidence from independent location verification systems into credibility assessments, so verifiers can quantify confidence in claims about where things happened.
Verified location is only one part of location-contingent policy evaluation. The geographic policy zones an evaluation pulls in must be trustworthy too: resolvable, efficiently fetchable, and cryptographically controlled by the appropriate parties.
The Location Protocol specifies cryptographically verifiable location records -- coordinates, extents, and proofs -- that independent systems can produce and consume interoperably.
The Astral documentation describes location records -- signed, verifiable location data artifacts that add cryptographic attribution and integrity checking to raw geospatial data.
Working notes toward a paper on verifiable spatial data registries -- systems of signed, identified spatial records whose integrity, write authority, and history can be checked without trusting the operator or the writers.
GeoDIDs — an early DID wrapper for spatial data assets; groundwork for the Location Protocol
Spec2021HistoricalIPLD-Encoded GeoTIFFs — content-addressed raster tiles -- effectively a registry scheme for raster data
Spec2021HistoricalState of Provenance Verification for Remote Sensing Imagery — a survey of provenance techniques for satellite and aerial imagery
Report2025This work looks at how verified locations can be evaluated against policy zones in trustworthy workflows. Incorporating a range of algorithms expands the policymaker's toolkit beyond naive geofencing.
Astral Compute evaluates spatial predicates -- distance, containment, intersection -- inside a trusted execution environment and returns cryptographically signed results that a third party can verify.
This forthcoming note will set out verifiable geocomputation -- performing spatial computation of any kind, from simple predicates to full analyses, so that the result carries evidence a third party can check rather than requiring trust in whoever ran it.
The pattern generalizes -- location-contingent mechanisms apply wherever machines act in governed physical space. These sketches show the breadth of what can be built, while my focus stays on the governance of AI compute.
Tying the authorization certificate a FlexHEG device requires to verified location, so an issuer can grant or withhold it according to where the hardware is.
Attenuating a system's capabilities as a function of verified location evidence.