Visualization of the Intuition-1 satellite.
Set to launch into orbit in just one month (November 2023), we at KP Labs have embarked on the ambitious Intuition-1 satellite project. With its compact CubeSat 6U design (30x20x10 cm), this venture isn’t just another satellite.
What exactly is Intuition-1?
Intuition-1 is a satellite mission designed to observe the Earth using a hyperspectral instrument and an on-board computing unit capable of processing data using neural networks (artificial intelligence) in orbit. By segmentation, we mean the ability to automatically determine the features of images – that is, to search for similar patterns.
Building upon this innovation, the Genesis project (developed with ESA Φ-lab for European Space Agency), created by KP Labs, takes a giant leap forward in soil health analysis. Hyperspectral imaging in Genesis captures light across a broad spectrum, allowing for the identification of soil elements such as K2O, P2O5, Mg, and pH level. Instead of sending large datasets to Earth, the onboard AI carries out immediate analysis in orbit. The combination of these techniques provides Genesis with an efficient approach to soil health analysis.
Intuition-1 is a technology demonstrator that aims to prove that the use of artificial intelligence to process hyperspectral data already in orbit positively affects the efficiency of the remote sensing process and that hyperspectral instruments based on KP Labs optics can observe phenomena so far impossible to detect by standard panchromatic, multispectral or radar satellites.
Core Components of Intuition-1
Data Processing Unit (Leopard DPU):
- Size and Weight: The unit boasts a compact design, fitting within a 0.8U volume and weighing just 1.3 kg. This lightweight, compact configuration includes redundant processing nodes, ensuring that the unit remains efficient without compromising on space.
- Processing Prower: Leopard showcases computing capacity, achieving up to 3 TOPS during neural network inferences. It houses a robust 16 GB DDR4 RAM with Error Detection and Correction (EDAC), ensuring uninterrupted operations and data integrity. To support its system and applications, a resilient 4 GB SLC flash memory with EDAC is incorporated, while a substantial 4 x 256 GB SLC flash storage retains processed data. Recognizing space’s demanding environment, a radiation-hardened controller oversees its operations, further fortified by redundant processing nodes to ensure mission reliability. At its core, a sophisticated System-on-a-Chip pairs with a Field-Programmable Gate Array (FPGA) and ARM processing cores, magnifying its efficiency when combined with machine learning libraries like Caffe and TensorFlow.
Data Processing Unit (Leopard DPU).
- Satellite-to-Earth Data Dynamics: Leopard DPU fundamentally shifts the data processing paradigm from terrestrial stations to in-situ satellite processing. This means we bypass the issues tied to high bandwidth consumption and lags affiliated with raw data transmission to Earth. Instead of raw data, only valuable, processed insights are sent back—drastically reducing latency and transmission costs. For mission designers and satellite operators, this translates to swifter access to processed information and nimble decision-making.
- AI Integration & Adaptability: The Leopard DPU integrates both hardware components and our processing software, coupled with AI algorithms. Given the challenge associated with transmitting extensive raw satellite data, notably from hyperspectral sources, the DPU offers an on-board solution. Adopting this “brain” close to the “eyes” strategy streamlines data handling. The Intuition-1 mission, functioning as a mobile laboratory, utilizes this configuration, enabling the updating and processing of machine learning models in orbit. The comprehensive deployment pipeline is designed to assess both functional and non-functional attributes of machine learning models prior to their satellite integration, ensuring they can withstand various in-orbit data-level interferences. Furthermore, the capacity to upload AI models to the satellite offers adaptability in addressing diverse Earth observation tasks. Currently, the primary focus lies in analyzing soil parameters through hyperspectral imagery, supporting improved agricultural processes and addressing wider challenges such as food security and climate implications.
Hyperspectral Camera (Eagle Camera)
- Instrument Type & Capability: Tailored specifically for hyperspectral imaging, this optical instrument delivers an unprecedented spectral resolution from visible to near-infrared light. Unlike traditional cameras, it subdivides the light spectrum into 192 channels. This vast number of channels offers a richer data set than most existing market offerings, unveiling phenomena hidden from the naked eye.
Hyperspectral Camera (Eagle Camera).
- Hyperspectral Cube: Complementing this is the Leopard Data Processing Unit, which adeptly processes the acquired data, compiling it into what’s known as a “hyperspectral cube”. Additionally, it ensures that these invaluable images are securely preserved in a non-volatile memory.
- Applications of Hyperspectral Imaging: Agriculture benefits from its capabilities in crop health assessment and disease detection. Forestry relies on it for invaluable data on forest health. For geologists and mineralogists, it’s essential in identifying mineral locations. The tool is also pivotal in environmental monitoring, spotlighting issues such as water quality and deforestation. In the realm of climate change studies, it offers comprehensive data on glacier movements and vegetation responses. Additionally, the defense sectors harness its prowess for target identification and environmental surveillance.
OBC Software (Oryx OBCS)
- Flight Software: Oryx is a modular flight software tool that manages all satellite tasks – namely, processing telecommands sent by the operators, monitoring the power budget, executing pre-defined schedules, managing emergencies, and handling data from all the sensors on board.
- Modular Architecture: Benefitting from a modular structure based on building blocks, Oryx streamlines the mission software development process. It taps into a comprehensive library of components such as logging, scheduling, testing, and communication, ensuring rapid and efficient development.
OBC Software (Oryx OBCS).
- SDK: At its core, the SDK is powered by a dynamic Scheduler that orchestrates automated jobs, whether they’re time-triggered or driven by telemetry values, and its capabilities are expanded further with an integrated Lua-based Scripting Engine. This engine not only allows for the seamless execution of precompiled Lua scripts onboard but also provides access to all internal APIs. This ensures that mission objectives and automated tasks can evolve and adapt without the need to reconfigure the entire flight software.
Supporting the Intuition-1 Mission
The Intuition-1 mission will be supported by a ground segment consisting of:
- KP Labs Mission Operations Center,
- KP Labs Ground Station,
- Ground Network Stations.
To ensure seamless communication with Intuition-1, our facility is integrated with two specialized antennas, enhancing our real-time communication capabilities. The station will support Intuition-1 operations via the TM&TC link in the UHF band, as well as payload data transfer in the S-band and X-band.
- The UHF band serves as the backbone for telemetry and telecommands, stabilizing connections at a rate of 9.6 kb/s.
- The S-band, responsible for software modifications, ensures that Intuition-1 remains adaptable, transferring data at up to 256 kb/s.
- For efficient data retrieval, the X-band stands ready, supporting rapid data rates of up to 50 Mb/s.
The Ground Station and Mission Operations Center are located at KP Labs’ headquarters in Gliwice, Poland.
As November approaches, we at KP Labs are eagerly anticipating the launch of Intuition-1. This project represents a culmination of our dedication to advancing space technology. Intuition-1 is not merely a satellite; it signifies a new era of Earth observation and discovery.
The next part of the article will describe the integration of the satellite’s components and what challenges we had to overcome to be ready for the launch. Stay tuned and follow our social media for more!