MiRAGE

Objectives of the Product

The human-centric approach to satellite operations gives rise to significant inefficiencies. Latency, limited communication windows, and expensive downlink represent bottlenecks that are detrimental to the effectiveness of the mission.

To address these problems, AIKO has developed MiRAGE, an onboard automation software package that provides the spacecraft with complete independence from ground control.

MiRAGE analyzes data from the satellite and its operational environment to devise and execute a dynamic schedule of tasks. The enabled autonomous capabilities therefore augment satellites with the ability to perceive and react to unexpected events, lowering operational costs and improving the quality of delivered services.

MiRAGE consists of three key pillars: (i) onboard data processing, (ii) operations planning, and (iii) autonomous control. Collectively, these pillars enable autonomous sensing, mission planning, satellite control, and cooperative systems in vast scenarios.

---

Customers and their Needs

The customers of MiRAGE are spacecraft manufacturers and operators. Their needs include:

• Reducing operating costs.
• Overcoming bottlenecks from human in the loop approaches.
• Increasing the satellite lifespan.
• Responding rapidly to unexpected events.
• Optimizing use of in-space and on-ground resources.
• Increasing the quality of services or products to become more competitive.

---

Targeted customer/users countries

Space is a global business, therefore AIKO targets satellite manufacturers and operators across the world.

---

Product description

MiRAGE is a software package that enables in-space mission autonomy for satellites.

MiRAGE boasts event detection and pattern recognition capabilities obtained through payload and telemetry data processing.

The software abstracts a simple cognitive architecture from complex space systems, providing satellites with the ability to:

• Sense the environment and its status (through onboard data processing).
• Plan tasks according to acquired or inferred knowledge (operations planning).
• Self-maneuver in the orbital environment (autonomous control).
• Operate in coordination with other constellation nodes to accomplish complex and distributed tasks.

The MiRAGE core libraries have been successfully tested for x86-64 and ARM computing architectures. Moreover, processing modules are compatible with a wide range of hardware accelerators (including Intel Myriad, Google Coral, Nvidia Jetson), resulting in several key advantages:

• Reduced inference time on the deep learning model.
• Reduced workload on the CPU.
• Optimized power consumption.

MiRAGE White Paper

MiRAGE Brochure

---

Added Value

MiRAGE enables autonomous operations on satellites, overcoming the limitations of human centric spacecraft operations. The independence from ground control paves the way for benefits such as a reduction of mission operations costs, increased activity lifespan, and optimized use of resources.

MiRAGE will also become the pillar for logistical scalability of novel constellation architectures, in which hundreds or even thousands of satellites will operate collaboratively to reach mission goals.

---

Current Status

Officially started in May 2021, the InCubed program is set to last two years. In this timeframe, AIKO intends to improve the technical soundness of MiRAGE by testing it in real operational scenarios, gaining flight hours in the process.

For this endeavor, AIKO is running the Early Adopters Program (EAP) to grant selected partners with early technological access to MiRAGE, ensuring compatibility of the product with the needs of potential customers.

Thus, AIKO is leveraging the InCubed development resources:

• To complete the product development roadmap and prepare it for commercial exploitation, bringing MiRAGE from a TRL (Technology Readiness Level) of 6 to a TRL of 9 by 2022.
• To conduct required qualification and testing campaigns.
• To acquire flight hours and training data in diverse scenarios to optimize the machine learning models.
• To verify and validate the product on the use cases identified during the EAP.