← Back to papers

Semi-Autonomous Reconnaissance Satellite: A Laboratory Demonstration of End-to-End On-Board Mission Autonomy

Dr. Eran Shtoyerman — Algorithms Group Leader
ISI
Engineering Technology AI/ML in Satellite Data Missions Automation Satellite Manufacturing Systems Engineering & Integration

Schedule

Poster Wednesday, May 27, 2026 · 2:30 PM · Posters Area – Kiosk 1

Abstract

Real-time satellite responsiveness to user requests from the field remains largely unavailable in current space systems. Such capabilities are typically restricted to low-altitude airborne platforms, primarily due to limited satellite communication bandwidth, intermittent ground contact, and the reliance on ground-based mission planning. Enabling true responsiveness requires a paradigm shift toward highly autonomous satellites, in which target selection, task execution, and mission planning are performed on board.
For example, a user may request a satellite to perform the following complex task in real-time: detect a moving object exhibiting camouflage characteristics, estimate its velocity from video data, and autonomously track the target over time. Such “chat-like” tasking commands can be conveyed over narrow-band communication links, potentially from handheld user terminals, while the satellite performs the entire chain of perception, reasoning, and control functions autonomously.
In the current work, we demonstrate, in a laboratory environment, an end-to-end processing chain for semi-autonomous microsatellites capable of executing such complex, high-level user requests.
The demonstrated processing chain includes: 1) User request sent to receiver 2) Compilation of an action list from the user request 3) Acquisition of multi-spectral video data 4) Motion detection and velocity estimation 5) AI-based object identification 6) Real-time computation of maneuver commands for target tracking 5) Transmission of target geolocation data along with a set of highly compressed image “stamps” over a low-bandwidth downlink. All processing stages were implemented on an edge-computing architecture representative of future ISI satellites payload computers. The experiment is planned to be extended to an on-orbit validation using an operational satellite.
The results validate the feasibility of executing complex, closed-loop reconnaissance tasks entirely on-board advanced high-resolution satellites. This capability enables a new class of responsive space systems, supporting real-time interaction on future small satellite constellations.

Authors

  • Dr. Eran Shtoyerman — Algorithms Group Leader
    ISI
  • Ms. Gal Pinchas — Development leader
    ISI
  • Dr. Doron Shterman — CTO
    ISI