跳到主要内容
ArcBlock Community

InfoPoint Whitepaper - A DID-Native Spatial Discovery and AI Retrieval Infrastructure

shenxiuqiang
支持

Version 0.9 Draft

Executive Summary

InfoPoint is a decentralized spatial discovery and semantic retrieval infrastructure designed for the AI-native internet.

The system transforms physical spaces into machine-readable semantic entities through:

  • BLE nearby discovery
  • LoRa spatial synchronization
  • DID-based spatial identity
  • Blocklet spatial runtimes
  • AI-native semantic retrieval

Unlike traditional IoT systems, InfoPoint does not attempt to convert edge hardware into miniature web servers. Instead, InfoPoint adopts a stateless spatial beacon architecture:

  • edge nodes broadcast spatial identity
  • nearby nodes synchronize lightweight spatial indexes
  • cloud runtimes host applications and services
  • mobile applications maintain local spatial graphs for AI retrieval

This architecture is optimized for:

  • scalability
  • low hardware cost
  • realtime contextual discovery
  • semantic interoperability
  • AI-native spatial search

The system is applicable across:

  • retail
  • restaurants
  • hospitality
  • exhibitions
  • tourism
  • creator economy
  • smart venues
  • urban semantic infrastructure

1. Problem

Modern internet infrastructure indexes webpages rather than physical space.

Discovery of nearby information still depends on:

  • centralized map providers
  • manual search
  • QR codes
  • closed ecosystems
  • platform-controlled ranking systems

Physical spaces lack:

  • globally addressable identity
  • semantic interoperability
  • realtime spatial state
  • AI-readable metadata
  • decentralized discovery mechanisms

Traditional beacon systems failed because they only broadcast URLs and lacked:

  • identity layer
  • semantic layer
  • trust layer
  • runtime layer
  • AI retrieval capability

The result is a fragmented and largely non-programmable physical environment.

2. Design Principles

InfoPoint follows several architectural principles:

2.1 Minimal Edge Hardware

Edge devices should not host complex applications, AI models, or databases.

The node should only:

  • broadcast spatial identity
  • synchronize lightweight nearby indexes
  • participate in spatial awareness

2.2 Decentralized Spatial Identity

Each physical space should possess:

  • a DID
  • ownership
  • verifiable identity
  • semantic metadata
  • globally resolvable runtime endpoints

2.3 AI-Native Retrieval

The system is designed for semantic spatial retrieval rather than keyword search.

Spatial context becomes directly consumable by AI systems.

2.4 Local Spatial Intelligence

Spatial graphs should primarily exist on user devices rather than centralized cloud services.

This improves:

  • privacy
  • latency
  • contextual awareness
  • offline capability

3. System Architecture

InfoPoint consists of six layers.

LayerResponsibility
BLE Discovery LayerNearby discovery
LoRa Spatial Sync LayerDistributed nearby index synchronization
DID Identity LayerGlobal spatial identity
Spatial Resolver LayerDID → Runtime resolution
Spatial Runtime LayerBlocklet application runtime
AI Semantic LayerRetrieval and reasoning

4. Hardware Architecture

InfoPoint hardware is intentionally lightweight.

The node is not a web server.

The node does not:

  • host AI
  • store media
  • run databases
  • serve rich applications

The node only:

  • broadcasts BLE discovery packets
  • synchronizes spatial indexes via LoRa
  • optionally provisions network configuration

BLE + LoRa Integrated Devices

Recommended development boards:

  • Heltec WiFi LoRa 32 V4
  • ESP32-C3 + SX1262
  • Nordic nRF52 + LoRa module

4.2 Required Capabilities

CapabilityRequired
BLE AdvertisingYes
LoRa CommunicationYes
Minimal Flash StorageYes
WiFi ProvisioningOptional
GPSOptional

5. Spatial Discovery

5.1 BLE Discovery

BLE is used exclusively for nearby discovery.

BLE advertisements contain:

  • InfoPoint UUID
  • node identifier
  • protocol version
  • compressed semantic metadata

Example:

javascript
{
  "v":1,
  "id":"cafe001",
  "t":"cafe"
}

The app scans nearby BLE packets and discovers nearby InfoPoint nodes.

5.2 QR Code Discovery

InfoPoint also supports QR-based entry.

QR codes provide:

  • compatibility with existing user behavior
  • instant onboarding
  • explicit user interaction
  • table-level or object-level addressing

Example restaurant deployment:

ObjectBinding
RestaurantMain Space DID
TableSub-space QR
MenuRuntime Plugin
PaymentCommerce Plugin

A restaurant may deploy:

javascript
Restaurant DID
↓
Table QR Code
↓
Specific Runtime Context
↓
Order Session

This enables:

  • table-specific ordering
  • localized interactions
  • multi-session runtime contexts

QR and BLE are complementary:

BLEQR
passive discoveryexplicit entry
ambient awarenessdirect interaction
nearby contextprecise targeting

6. LoRa Spatial Synchronization

LoRa is not used for content delivery.

LoRa is used for:

  • distributed spatial index synchronization
  • nearby awareness propagation
  • semantic metadata gossip

Example synchronized index:

javascript
{
  "did":"did:abt:cafe001",
  "type":"cafe",
  "lat":40.712,
  "lng":-73.99,
  "tags":["quiet","wifi","workspace"]
}

This allows a user to discover many nearby spaces through a single reachable node.

7. DID Identity and Resolution

Each space owns a DID.

Example:

javascript
did:abt:cafe001

The DID represents:

  • ownership
  • trust
  • identity
  • routing anchor

InfoPoint uses ArcBlock DID infrastructure.

7.1 DID Resolution Flow

javascript
BLE / QR
↓
DID
↓
InfoPoint Resolver
↓
Forge Account Metadata
↓
Runtime Endpoint
↓
Spatial Runtime

7.2 DID Metadata Structure

Example metadata:

javascript
{
  "infopoint": {
    "runtime":"https://space.infopoint.ai/cafe001",
    "resolver":"https://resolver.infopoint.ai",
    "spaceType":"cafe",
    "services":[
      "menu",
      "booking",
      "event",
      "ai-guide"
    ]
  }
}

7.3 Resolver Design

The resolver layer is intentionally separated from DID ownership.

Blockchain provides:

  • identity
  • ownership
  • authenticity

The resolver provides:

  • dynamic runtime routing
  • cache optimization
  • geo-aware routing
  • runtime migration

This separation enables runtime flexibility without modifying chain state.

8. Spatial Runtime

InfoPoint runtimes are implemented as Blocklet applications.

Each physical space corresponds to a programmable runtime.

8.1 Runtime Capabilities

Core runtime functions:

  • profile
  • events
  • booking
  • commerce
  • messaging
  • realtime state
  • analytics
  • AI interaction

8.2 Plugin System

The runtime uses a plugin architecture rather than fixed industry templates.

Examples:

IndustryPlugin
RestaurantMenu
HotelRoom
ExhibitionBooth
TourismGuide
RetailCoupon
PhotographyBooking
LivehouseEvent

This enables a single runtime framework to support multiple industries.

9. AI-Native Spatial Retrieval

The spatial graph is primarily maintained on the mobile application side.

The app gradually accumulates:

  • nearby spatial graph
  • semantic embeddings
  • nearby memory cache
  • user preference graph
  • realtime spatial state

This architecture aligns with recent Spatial-RAG research.

9.1 Retrieval Pipeline

javascript
User Query
↓
Semantic Parsing
↓
Spatial Filtering
↓
Local Spatial Graph
↓
Semantic Ranking
↓
AI Response

9.2 Example Query

User asks:

javascript
Nearby quiet cafes suitable for work

The system combines:

  • distance
  • occupancy
  • semantic tags
  • WiFi availability
  • realtime noise level
  • user preference history

This produces contextual semantic recommendations rather than static map results.

10. Local Spatial Intelligence

InfoPoint prioritizes local spatial intelligence over centralized cloud indexing.

Advantages:

  • lower latency
  • privacy preservation
  • offline capability
  • personalized semantic memory
  • reduced infrastructure cost

The app becomes a local AI-native spatial retrieval engine.

11. Security and Trust

InfoPoint uses DID-based ownership and identity verification.

The system separates:

LayerResponsibility
DIDownership
Resolverrouting
Runtimeservices
AI Layerretrieval

This minimizes coupling between trust, routing, and application execution.

12. Commercial Architecture

InfoPoint is infrastructure rather than a single-purpose application.

Potential commercial layers include:

  • hardware sales
  • hosted spatial runtimes
  • plugin marketplace
  • enterprise analytics
  • AI services
  • spatial search APIs
  • DID registration services

The primary long-term value is:

  • realtime spatial semantic data
  • AI retrieval infrastructure
  • urban spatial intelligence

13. Comparison

SystemNearby DiscoveryIdentitySemantic LayerAI Retrieval
QR CodeManualNoNoNo
BLE BeaconPartialWeakWeakNo
Google MapsCloud-onlyCentralizedPartialLimited
InfoPointNativeDIDNativeNative

14. Conclusion

InfoPoint proposes a practical architecture for transforming physical spaces into decentralized semantic infrastructure.

The system combines:

  • BLE discovery
  • QR entry
  • LoRa synchronization
  • DID identity
  • Blocklet runtimes
  • AI-native retrieval

to create a new category of infrastructure:

AI-native spatial discovery and semantic retrieval.

Rather than connecting devices, InfoPoint semanticizes physical space.

回复