Turning orbital data into decisions — on board.
An Apohele R&D program developing intelligent payloads that process sensor data directly on the spacecraft, so missions downlink insight instead of raw data. Currently in active research and development.
This is an active R&D program, not a finished or operational product. The capabilities described on this page are development targets and research objectives — not delivered or flight-proven specifications.
Downlink is the bottleneck, not the sensor
A small spacecraft can capture far more data than it can ever send to the ground. Contact windows are short and downlink is scarce — and much of what's captured, like cloud-covered scenes, is never useful.
This program's goal is on-board intelligence that decides what matters before anything is transmitted — filtering, prioritizing, and tagging data at the edge so every downlink carries useful information rather than raw, unsorted captures.
Intelligence at the edge of the mission
The program's research is organized around three directions, each an active area of work rather than a delivered capability.
On-board inference
Running trained models directly on the payload to interpret sensor data in real time, within the size, weight, and power limits of a small spacecraft.
Intelligent data reduction
Filtering, prioritizing, and summarizing data on orbit so a mission's limited downlink is spent only on what actually matters.
Reconfigurable on orbit
Designed so new models can be deployed to hardware already in flight — letting one payload be re-tasked for new missions after launch.
Where on-board intelligence could matter
These are target application areas the program is exploring — each a research direction rather than a delivered capability.
Cloud & quality screening
Identifying and setting aside low-value frames — such as cloud-covered scenes — so limited downlink is not spent on data the mission can't use.
Event & object detection
Flagging features of interest, such as wildfires, floods, or vessels, and prioritizing them for rapid downlink.
Signal interpretation
On-board interpretation of RF and other sensor streams, so missions downlink detections and summaries rather than raw captures.
Mission autonomy
Supporting on-board decision-making and health monitoring to reduce a mission's dependence on ground contact.
A research roadmap
- Now
Research & architecture
Defining the on-board processing architecture and evaluating model and hardware approaches against small-spacecraft constraints.
- Next
Prototyping & validation
Building and characterizing prototypes against representative data to validate performance and the path to flight.
- Future
Flight demonstration
Maturing the payload toward an in-orbit technology demonstration aboard a Cubestat platform.
Let's talk about your next mission
Whether you're planning a payload, evaluating platforms, or exploring quantum navigation research, our team would like to hear from you.