What this portfolio covers

Two granted US patents, US 12,001,207 (issued June 2024) and US 12,530,030 (issued January 2026), with 33 claims between them. The patent family also includes EP3786756B1, a granted European patent. All share a priority date of August 9, 2019, and expire March 5, 2041.

US 12,001,207 describes a dual-module safety system for autonomous vehicles and drones. A safety-determining module compares live camera images against stored reference images and generates a safety value. A separate control module only executes navigation instructions if that safety value meets a predetermined threshold. If confidence drops below the threshold, the system hands off to a remote human operator.

US 12,530,030 (continuation) adds a clear-passage-determining module. It evaluates whether a navigation instruction can be executed based on current traffic, e.g., waiting for oncoming vehicles to clear before turning left. The continuation has 20 claims across method, computer program product, and system categories.

Claim 13 of US 12,530,030 covers "deep learning via network topology for converting navigation instructions into directional and acceleration values." The patent abstract describes the same pipeline: converting camera images and navigation instructions into control values. The claims are implementation-agnostic, covering this functional pipeline whether it uses CNNs, transformers, or end-to-end learned models. The specification names Nvidia's DAVE-2 network as a suitable implementation.

In short: Dual-module safety architecture, clear-passage determination, and end-to-end deep learning coverage (Claim 13). Camera-based navigation with safety confidence scoring, behavioral cloning, and remote operator handoff. Two US patents, 33 claims, protection through March 5, 2041.

Coverage areas

  • Dual-module camera-based navigation safety systems
  • Clear-passage determination for traffic-aware navigation
  • End-to-end deep learning for converting navigation instructions into vehicle control values (Claim 13)
  • Behavioral cloning from human operator demonstrations
  • Camera-based guidance systems for autonomous vehicles
  • Commercial and consumer drone navigation
  • Machine learning navigation algorithms
  • Confidence-based safety assessment with operator handoff

Timing matters

Tesla's unsupervised FSD rollout and the broader shift toward end-to-end neural networks have accelerated the camera-first AV space. Regulators are paying closer attention to safety systems, and patent filings in this area are increasing. Licensing now, before the space gets more crowded, gives you a cleaner position.

A few things happening in the market right now:

  • Tesla is scaling unsupervised FSD, and competitors need differentiated IP
  • End-to-end neural network navigation is replacing modular pipelines, and this portfolio explicitly covers end-to-end deep learning (Claim 13 of US 12,530,030)
  • NHTSA and international regulators are tightening autonomous safety requirements
  • Camera-based perception is overtaking lidar in cost-sensitive applications

Patent details

  • US 12,001,207 — Issued June 4, 2024 — 13 claims (dual-module safety architecture)
  • US 12,530,030 — Issued January 20, 2026 — 20 claims (clear-passage-determining module)
  • EP3786756B1 — European patent (same patent family)
  • Expiration: March 5, 2041 (both US patents, terminal disclaimer)
  • Applications: Autonomous vehicles, drones, AI navigation

View full portfolio details · US 12,001,207 on Google Patents · US 12,530,030 on Google Patents

Licensing

This patent portfolio is available for licensing. If you're building camera-based autonomous navigation for vehicles, drones, or other platforms, get in touch to discuss terms.

Next Page →