Cost-Optimized Autonomous Flight:
NexusV2 Flight Controller

Integrated SoC Architecture

The NexusV2 FC is a cost- and power-optimized autonomous flight platform designed for production-ready UAV programs. It consolidates flight control, mission autonomy, and onboard AI perception into a single heterogeneous compute SoC, eliminating the need for external companion computers while achieving 40–60% lower BOM cost compared to FPGA-centric designs.

AI-Powered Navigation & Perception

• Visual Localization: Offline map-based localization using cameras, visual feature extraction, and map tile matching with 10–15ms end-to-end latency.
• Object Detection: Integrated Neural Processing Engine (NPU) running INT8 quantized models at 30–40 FPS for target detection and obstacle recognition.
• Multi-Object Tracking: Concurrent tracking of 4–6 objects using Kalman filters and optical flow assistance.

Cost & Power Advantages

• 40–60% Lower BOM Cost: Heterogeneous SoC architecture eliminates expensive FPGA components.
• Reduced Power Consumption: Integrated NPU consumes significantly less power per inference than FPGA-based solutions.
• Simplified Manufacturing: Lower PCB complexity and reduced thermal requirements.
• Faster Development: Software-driven AI model iteration without hardware redesign.

Safety & Reliability

• Sensor Redundancy: Dual or triple IMU support with multiple camera inputs.
• Fault Handling: Automatic transition from GNSS to vision-inertial navigation with graceful AI degradation.
• Certification Ready: Architecture aligns with DO-178C and DO-254 principles for aerospace standards.