TECHNOLOGY / QxAMP
The fusion layer.
Quantum Edge-AI Encoding & Multiplexing links photon sensing, hybrid computation and distributed analysis in a four-layer platform architecture.
EXPLORE THE PLATFORM
Four layers. A continuous workflow.
01 / Edge AI
The computational workflow
Photon preprocessing, quantum feature mapping, variational transformation, latent-feature extraction and synthetic data generation. The manuscript describes these as cooperating software families.
Source: Volume 1, §4.2, page 12.
02 / Hardware
From photons to processable events
An asynchronous free-space detector feeds timing and processing functions. The manuscript pairs the sensor with Jetson AGX Orin compute. Optical sensing specifications do not by themselves establish radio-frequency reception.
Source: Volume 1, §4.3, page 15.
03 / Solvers
Represent. Optimize. Multiplex.
The ground solver family includes Qx65Qe stream VQE, Qx10Fm feature mapping and Qx65Qa multiplexing VQA. These are the named algorithmic modules in the research architecture.
Source: Volume 1, §4.5, page 19.
04 / Backend
Extend the computational envelope
The paper proposes ground GPU simulation and quantum backends for larger circuits, noise studies and solver development. Topological hardware capabilities are described as an architecture direction, not a flight-verified service.
Source: Volume 1, §4.6, page 20.
AT THE SENSOR
Every event
carries information.
The asynchronous free-space single-photon detector provides timestamped optical events for downstream feature extraction. The following values are supplied specifications reproduced in Table I, not flight measurements.
| Spectral range | 700–1500 nm |
|---|---|
| Calibration wavelength | 1325 nm |
| Typical timing jitter | <161 ps at 28% efficiency |
| Maximum timing jitter | <195 ps at 28% efficiency |
| Adjustable dead time | 100 ns–80 μs |
| Free-running count rate | Approximately 1 Mcps |
Source: Volume 1, Table I, page 16.
Manuscript v3.9.1 · page 15 · Original figure available in the private research workspace.
Manuscript v3.9.1 · page 12 · Original figure available in the private research workspace.
HYBRID COMPUTATION
A coordinated
software stack.
The manuscript connects photon processing and generative models through Qx-Core, Qx-Sense, Q-Sim and Q-Tensor. Its software discussion includes CUDA-Q, JetPack, TensorRT and PyTorch as technology references.
GPU-based quantum simulation and execution on physical quantum hardware are distinct modes. The platform’s research direction includes both.
Explore algorithm familiesSource: Volume 1, §4.2.3–4.2.4, page 14.
FROM RESEARCH TO ENGINEERING
Explore the architecture.
Inspect the decisions.
Open the Q-LEO engineering workspace for spacecraft layers, interactive CAD and transparent simulation studies.