IFEVS is pleased to announce the publication of a comprehensive research paper that translates insect neurobiology into practical edge AI architectures for safety-critical systems.

"Edge AI in nature: insect-inspired neuromorphic Reflex Islands for safety-critical edge systems" presents a complete hardware-software blueprint demonstrating how insects achieve millisecond sensor-motor loops with minimal energy consumption—principles that can revolutionize autonomous vehicles, medical implants, cargo e-bikes and propulsion systems.

The paper introduces a latency-first control hierarchy that partitions tasks between a fast, dedicated Reflex Tier and a slower, robust Policy Tier, realized through neuromorphic hardware utilizing MRAM synapses and spin-torque nano-oscillator (STNO) reservoirs. Beyond the theoretical framework, the research includes experimental validation of the IFEVS hybrid thruster, achieving >40% thermal efficiency with flat performance across load ranges—directly inspired by insect flight muscle operation.

Key Contributions

  • Neuromorphic Reflex Islands: Instant-on, memory-centric control using spintronic primitives
  • Thermal Governance Framework: Bio-inspired burst budgeting with closed-form solutions for small-scale propulsion
  • Multi-Domain Applications: Validated use cases in fuel-based propulsion, autonomous cargo e-bikes, medical implants, and industrial robotics
  • Safety Certification Pathway: Explicit WCET envelopes and freedom-from-interference boundaries for ISO 26262/ASIL compliance

The research was conducted by IFEVS team members Pietro Perlo, Marco Dalmasso, Marco Biasiotto and Davide Penserini, building on the company's experience in hybrid propulsion and edge AI integration.

Download the full paper: www.preprints.org/manuscript/202511.1387

This work is supported by the European Commission through the EIC Pathfinder MultiSpin.AI, HORIZON NeAIxt and HORIZON EdgeAI-Trust projects, positioning IFEVS at the forefront of neuromorphic computing for mobility and autonomous systems.