DVP Status: PLANNED — This profile is currently under development. Contributions and feedback welcome.
VAP PROFILE: AUTONOMOUS MOBILITY

DVP - Driving Vehicle Protocol

AI Flight Recorder for Autonomous Driving

"Every Decision, Cryptographically Recorded"

— Record every decision with cryptographic proof

" Not asking "why" after the accident — leaving "evidence" before the accident. "

2018, Uber autonomous vehicle fatality. 2016, Tesla Autopilot fatality.

Investigators faced the same question:

"What did the AI see, what did it decide, and why did it choose that action?"

Traditional EDRs (Event Data Recorders) capture vehicle physical state.

But the AI's internal decision process goes unrecorded.

DVP fills this gap — recording the complete causal chain of AI decisions in a tamper-proof format.

View Architecture Regulatory Compliance Join Development

Why DVP is Needed

Structural challenges in current autonomous driving systems

AI Decision Black Box

How inputs from LiDAR, camera, and radar are processed, and why decisions like "go straight," "stop," or "avoid" are made — current EDRs don't record this.

Ambiguous Liability

At Level 3 and above, driving responsibility shifts to the system. In an accident, it's impossible to prove whether it was "human error" or "system defect."

Regulatory Requirements

UNECE WP.29 R157, EU AI Act Annex III require autonomous driving AI recording and accountability. No corresponding technical standard exists.

Real Accident Investigation Scenario (Anonymized)

Accident Occurs

Investigation Starts

Log Retrieval

Issues Found

"Unknown if AI recognized the pedestrian"

"Cannot identify brake decision trigger"

"Timestamp integrity unverifiable"

DVP Application Scope

Target systems and recording requirements

Target Systems

🚗

Autonomous Vehicles

SAE Level 3-5

Required
🛡️

ADAS (Advanced Driver Assistance)

Level 2+

Recommended
🚆

Railway Operations AI

Autonomous Operations

Required
🚁

Drone Autonomous Control

Autonomous Flight

Recommended
🚌

Autonomous Bus/Shuttle

Level 4

Required

Recording Target Events

[Sensor Input] → [Recognition] → [Decision] → [Control Output] ↓ ↓ ↓ ↓ LiDAR point cloud Object detection Path planning Steering Camera images Classification Speed decision Accelerator Radar data Tracking ID Priority Brake GPS/IMU Trajectory pred Risk assessment Signal

Sensor Layer

  • • Raw sensor data hashes
  • • Frame synchronization
  • • Calibration status

Perception Layer

  • • Detection results
  • • Confidence scores
  • • Object tracking IDs

Planning Layer

  • • Path candidates
  • • Decision rationale
  • • Risk scores

Control Layer

  • • Command outputs
  • • Actuator responses
  • • Safety interventions

DVP Architecture

Data flow and system integration

System Data Flow

┌─────────────────────────────────────────────────────────────────────────┐ │ Vehicle System │ │ │ │ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ │ │ │ Sensors │───▶│ AI Stack │───▶│ Actuators │ │ │ │ │ │ │ │ │ │ │ │ LiDAR/Cam/ │ │ Perception │ │ Steering │ │ │ │ Radar/GPS │ │ Planning │ │ Throttle │ │ │ └──────────────┘ └──────┬───────┘ │ Brake │ │ │ │ └──────────────┘ │ │ │ │ │ ┌────────▼────────┐ │ │ │ DVP Logger │ ← Intercept all decision events │ │ │ (Sidecar) │ │ │ └────────┬────────┘ │ │ │ │ └──────────────────────────────┼──────────────────────────────────────────┘ │ ┌──────────▼──────────┐ │ Hash Chain │ ← Cryptographically linked │ + Signature │ ← Digital signatures │ + Merkle Root │ ← Periodic anchoring └──────────┬──────────┘ │ ┌──────────▼──────────┐ │ Secure Storage │ ← Crash/fire resistant │ (EDR Integrated) │ └─────────────────────┘

DVP Event Structure (Conceptual) 

{
  "event_id": "019234ab-7c8d-7def-8123-456789abcdef",
  "timestamp_ns": 1734567890123456789,
  "event_type": "PERCEPTION_DECISION",
  "vehicle_id": "VIN_XXXXXXXXXXXX",
  "provenance": {
    "actor": {
      "type": "AI_MODEL",
      "identifier": "perception_v3.2.1",
      "model_hash": "sha256:abc123..."
    },
    "input": {
      "lidar_frame_id": "frame_12345",
      "camera_frame_ids": ["cam_front_12345", "cam_left_12345"],
      "input_hash": "sha256:def456..."
    },
    "context": {
      "speed_kmh": 45.2,
      "weather": "RAIN_LIGHT",
      "visibility_m": 120,
      "active_mode": "AUTONOMOUS_L4"
    },
    "action": {
      "decision": "EMERGENCY_BRAKE",
      "confidence": 0.94,
      "trigger": "PEDESTRIAN_DETECTED",
      "predicted_collision_ms": 1200
    }
  },
  "prev_hash": "sha256:789xyz...",
  "signature": "ed25519:..."
}

Regulatory Compliance Mapping

International regulatory alignment

Regulation Jurisdiction Requirements DVP Support
UNECE WP.29 R157 UN (Global) ALKS data recording obligations ✅ Full Support
EU AI Act Annex III EU Transportation AI high-risk classification, logging requirements ✅ Full Support
ISO 26262 International Functional safety, traceability requirements ✅ Complementary
ISO/PAS 21448 (SOTIF) International Safety of the Intended Functionality ✅ Evidence Support
NHTSA AV Policy USA Autonomous vehicle data recording guidance ✅ Compliant Design
Japan Road Vehicle Act Revision Japan Level 3+ data recording requirements ✅ Planned Support

UNECE R157 DSSAD Requirements Mapping

Alignment with Data Storage System for Automated Driving (DSSAD) requirements

✅

5-second pre-accident data retention

All events preserved in hash chain

✅

Timestamp accuracy

UUID v7 + PTP synchronization

✅

Data integrity guarantee

Cryptographic hash chain

✅

Authority access requirements

Standard format export capability

Physical EDR Integration

Role separation between existing EDR and DVP

Existing EDR

Event Data Recorder

  • Speed, acceleration, brake operation
  • Seatbelt status, airbag deployment
  • Steering angle, throttle position

→ Records "what the vehicle did"

DVP Layer

AI Decision Recording

  • Sensor input hashes
  • AI recognition results (object detection, classification)
  • Decision logic (path planning, risk assessment)
  • Control command rationale

→ Records "why the AI decided that"

Physical Data
AI Decisions
Complete Record

Integrated Flight Recorder

Physical State + AI Decisions

Use Cases

How DVP transforms accident investigation and liability determination

Scenario: Autonomous Vehicle and Pedestrian Contact Accident

Phase EDR Only With DVP Integration
Accident Occurs Records collision speed and deceleration Same + AI recognition/decision history
Investigation Starts "Brake was activated" "Pedestrian detected 2.3 seconds prior, misclassified as bicycle with 0.67 confidence"
Cause Identification "Unknown if system or human error" "Recognition model v3.2 low-light classification accuracy issue"
Improvement Measures Speculation-based Specific model improvement points identified
Litigation Response Insufficient evidence Cryptographically verifiable evidence trail

Insurance & Liability Determination Scenario

Accident Occurs

DVP Log Retrieval

Cryptographic Verification

  • • Hash chain integrity ✓
  • • Timestamp verification ✓
  • • Signature verification ✓

Liability Clarification

Liability Determination

AI system decision error → OEM/Developer liability
Human override → Driver liability
Sensor failure → Component manufacturer liability

Technical Specifications Summary

DVP core technical requirements

Timestamp Accuracy
Nanoseconds
IEEE 1588 PTP synchronized
Event Recording Frequency
Up to 1,000/sec
Maximum event throughput
Hash Algorithm
SHA-256
Future SHA-3 support
Signature Algorithm
Ed25519
Future Dilithium (PQC) support
Storage Requirements
~10MB/hour
After compression
Physical Requirements
EDR Compliant
Crash/fire resistant

Roadmap

DVP development and standardization timeline

2025 Q4

DVP Draft Specification v0.1 Release

Initial draft specification for public review and feedback

2026 Q1

OEM & Tier 1 Technical Validation

Begin proof-of-concept testing with automotive industry partners

2026 Q2

UNECE WP.29 GRVA Proposal

Submit to UNECE Working Party 29 GRVA (Autonomous Driving)

2026 Q3

DVP v1.0 Official Release

Stable specification release with reference implementations

2027

ISO/SAE Collaboration

International standardization activities with ISO and SAE

Relationship with VAP/VSO

DVP's position in the framework hierarchy

┌────────────────────────────────────────────────────────────────────────────┐ │ │ │ VAP (Verifiable AI Provenance) │ │ Cross-domain upper framework for all domains │ │ │ └────────────────────────────────────┬───────────────────────────────────────┘ │ │ defines & maintains │ ┌────────────────────────────────────▼───────────────────────────────────────┐ │ │ │ VSO (VeritasChain Standards Organization) │ │ Standards body that develops and maintains VAP │ │ │ └────────────────────────────────────┬───────────────────────────────────────┘ │ │ publishes profiles │ ┌──────────┬───────────────┼───────────────┬──────────┐ │ │ │ │ │ ▼ ▼ ▼ ▼ ▼ ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐ ┌─────────┐ │ VCP │ │ DVP │ │ MAP │ │ EIP │ │ PAP │ │ Finance │ │Automotive│ │ Medical │ │ Energy │ │ Public │ │ Profile │ │ Profile │ │ Profile │ │ Profile │ │ Profile │ └────┬────┘ └────┬─────┘ └─────────┘ └─────────┘ └─────────┘ │ │ ▼ ▼ v1.0 Planned Released Domain-specific "concrete protocol implementations"

Get Involved

Join the development of DVP and shape the future of autonomous vehicle safety

View VAP Framework

Understand the upper framework that DVP extends

VCP Specification (Financial)

See the released VCP v1.0 as reference implementation

Participate in DVP Development

For OEM, Tier 1, and regulatory authority representatives
"Aircraft have physical black boxes. Autonomous vehicles need AI black boxes too."

— VeritasChain Standards Organization

"The question is not whether autonomous vehicles will have accidents.
The question is whether we can prove what happened when they do."

This work is licensed under CC BY 4.0 International