1. Analyze and Adopt

Domain: Aerospace Engineering / Satellite Communications / Open-Source Space Research Expert Persona: Senior Space Systems Architect & Digital Signal Processing (DSP) Specialist

2. Peer Review Group

This material is most relevant to the Satellite Communications and Space Systems Engineering community, specifically groups involved in open-source hardware/software for SmallSats and deep-space communication. Recommended Reviewers:

• AMSAT (Radio Amateur Satellite Corporation) Engineering Team: To evaluate the feasibility of the proposed signal identification and on-orbit upgrade protocols.
• IEEE Aerospace and Electronic Systems Society (AESS): To review the integration of low-power Neural Processing Units (NPUs) and radiation-resilient memory architectures.
• Deep Space Network (DSN) Amateur Affiliates: To assess the "Earth-Venus-Earth" (EVE) link budget and ground station precision timing requirements.

3. Summary

Abstract: This transcript documents a technical working session of the Open Research Institute (ORI) regarding the development of open-source satellite subsystems and ground station infrastructure. Key technical discussions focus on the implementation of machine learning (ML) for real-time signal identification on constrained hardware, targeting the AMSAT-UK mission. The engineers address the trade-offs between power consumption (targeting <0.5W) and the use of dedicated NPUs for feature extraction. The session also details ground station advancements at the Haswell site, including the calibration of Rubidium-based precision timing systems and the mechanical integration of multi-band feeds for the 60-foot dish. Finally, the group evaluates the link budget and regulatory challenges for the Earth-Venus-Earth (EVE) phase 2 mission, emphasizing the need for high-power, long-duration transmissions at 2.4 GHz to overcome path loss and limited reflectivity.

Technical Summary and Key Takeaways:

• 0:46 Neural Processing for On-Orbit ML: Discussion on utilizing Neural Processing Units (NPUs) capable of 26 TOPS at 2W for real-time data analysis. While 2W is efficient for terrestrial use, it exceeds the 0.5W total power budget for a 2U LEO spacecraft, requiring a shift toward simpler heuristic designs for signal identification.
• 3:31 Signal Identification Strategy: For the AMSAT-UK implementation, the system will focus on a finite set of known signals. The workflow involves feature extraction (bandwidth, power, modulation index) followed by a table lookup for identification (e.g., BPSK), rather than a power-hungry "unknown signal" classifier.
• 6:45 On-Orbit Software Upgrades: Emphasis on the necessity of in-flight reprogrammability to iterate on ML models and maximize processor efficiency as mission priorities evolve.
• 9:04 Resilience and Watchdogs: Implementation of dual-layer "Watchdog" systems (software and hardware) to recover from system hangs or "bricked" states during remote upgrades, ensuring the system reverts to a stable original software state if a new deployment fails.
• 10:53 Radiation Mitigation in LEO: Addressing SEU (Single Event Upset) risks from gamma rays. Baseline designs for the mission utilize triple-redundancy memory and radiation-resilient FPGA architectures (Lattice ICE40) to maintain data integrity.
• 15:54 Spectral Constraints: The subsystem is being designed for 30 kHz of spectrum at a 10.7 MHz IF, with the final channelization handled via polyphase filters.
• 23:12 Fishbowl SDR/GPSDO Development: Technical update on the "Fishbowl" (Pluto SDR clone) daughterboard. Success was reported in synchronizing the 40 MHz VCO with a 1 PPS GPS signal on the FPGA to create a high-precision reference clock.
• 29:46 Precision Timing Infrastructure: The Haswell ground station has integrated a Rubidium frequency standard (HP 5065A) monitored against GPS to achieve frequency stability of 1 part in 10^-12. The system is intentionally not "steered" by GPS during measurements to avoid ionospheric artifacts.
• 34:58 Ground Station Command and Control: The 60-foot dish utilizes a fiber optic link between the feed-mounted V200 SDR/down-converters and the control room to minimize signal loss and interference.
• 52:11 Earth-Venus-Earth (EVE) Mission Parameters: Phase 2 (October 2026) faces a significant challenge: Venus will be within 6–8 degrees of the sun, posing thermal risks to the dish feed.
• 55:32 EVE Link Budget and Power: To achieve a successful link at 2.4 GHz, the station plans to combine four to six 250W amplifiers to reach 1500W. Long integration times (weeks of continuous transmission) are required due to the extremely low bit rate of the Venusian reflection channel.
• 1:02:13 Regulatory Strategy: The project aims to operate under Part 97 (Amateur Radio) regulations to demonstrate that deep-space communication is achievable with high-end amateur equipment, though Part 5 (Experimental) licensing remains a fallback for higher power needs.