TinyML Enhances CubeSat Mission Capabilities

2026-03-20

AI Summary

This work addresses the limitations of traditional Earth observation systems that rely on ground-based processing, which is inefficient for CubeSats due to their constrained computational resources, power budgets, and communication bandwidth. To enable efficient on-orbit image classification, the authors propose an end-to-end TinyML optimization pipeline tailored for CubeSat platforms, implemented on an STM32N6 microcontroller featuring a Cortex-M55 CPU and a neural processing unit (NPU). The approach innovatively integrates structured iterative pruning, post-training INT8 quantization, and hardware-aware operator mapping to drastically compress models while aligning with the heterogeneous architecture. Experimental results demonstrate an 89.55% reduction in RAM usage and a 70.09% decrease in Flash footprint, with inference energy consumption ranging from 0.68 to 6.45 mJ and latency between 3.22 and 30.38 ms, while limiting accuracy degradation to 0.4–8.6 percentage points.

Technology Category

• Machine Learning: Hardware-aware ML
• Computer Vision: Learning & Optimization for CV
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

Earth observation (EO) missions traditionally rely on transmitting raw or minimally processed imagery from satellites to ground stations for computationally intensive analysis. This paradigm is infeasible for CubeSat systems due to stringent constraints on the onboard embedded processors, energy availability, and communication bandwidth. To overcome these limitations, the paper presents a TinyML-based Convolutional Neural Networks (ConvNets) model optimization and deployment pipeline for onboard image classification, enabling accurate, energy-efficient, and hardware-aware inference under CubeSat-class constraints. Our pipeline integrates structured iterative pruning, post-training INT8 quantization, and hardware-aware operator mapping to compress models and align them with the heterogeneous compute architecture of the STM32N6 microcontroller from STMicroelectronics. This Microcontroller Unit (MCU) integrates a novel Arm Cortex-M55 core and a Neural-ART Neural Processing Unit (NPU), providing a realistic proxy for CubeSat onboard computers. The paper evaluates the proposed approach on three EO benchmark datasets (i.e., EuroSAT, RS_C11, MEDIC) and four models (i.e., SqueezeNet, MobileNetV3, EfficientNet, MCUNetV1). We demonstrate an average reduction in RAM usage of 89.55% and Flash memory of 70.09% for the optimized models, significantly decreasing downlink bandwidth requirements while maintaining task-acceptable accuracy (with a drop ranging from 0.4 to 8.6 percentage points compared to the Float32 baseline). The energy consumption per inference ranges from 0.68 mJ to 6.45 mJ, with latency spanning from 3.22 ms to 30.38 ms. These results fully satisfy the stringent energy budgets and real-time constraints required for efficient onboard EO processing.

Problem

• CubeSat
• Earth observation
• onboard processing
• resource constraints
• communication bandwidth

Innovation

• TinyML
• CubeSat
• model compression
• hardware-aware deployment
• onboard image classification