Nano-U: Efficient Terrain Segmentation for Tiny Robot Navigation
2026-05-11 • Robotics
RoboticsComputer Vision and Pattern Recognition
AI summaryⓘ
The authors created a very small and efficient computer program called Nano-U that can help small robots tell the difference between different types of ground outdoors. Because Nano-U is so tiny, they used a special training method called Quantization-Aware Distillation to make sure it still works well even with limited power and memory. They tested it on real outdoor and farm field data and got good results. They also made Nano-U run smoothly on an inexpensive microcontroller by using a smart software engine that saves memory and speeds up processing. This shows it is possible to do useful robot terrain detection on very simple, low-cost devices.
terrain segmentationmicrocontrollerbinary segmentation networkQuantization-Aware Distillationknowledge distillationquantization-aware trainingTinyMLinference engineESP32-S3MicroFlow
Authors
Federico Pizzolato, Francesco Pasti, Nicola Bellotto
Abstract
Terrain segmentation is a fundamental capability for autonomous mobile robots operating in unstructured outdoor environments. However, state-of-the-art models are incompatible with the memory and compute constraints typical of microcontrollers, limiting scalable deployment in small robotics platforms. To address this gap, we develop a complete framework for robust binary terrain segmentation on a low-cost microcontroller. At the core of our approach we design Nano-U, a highly compact binary segmentation network with a few thousand parameters. To compensate for the network's minimal capacity, we train Nano-U via Quantization-Aware Distillation (QAD), combining knowledge distillation and quantization-aware training. This allows the final quantized model to achieve excellent results on the Botanic Garden dataset and to perform very well on TinyAgri, a custom agricultural field dataset with more challenging scenes. We deploy the quantized Nano-U on a commodity microcontroller by extending MicroFlow, a compiler-based inference engine for TinyML implemented in Rust. By eliminating interpreter overhead and dynamic memory allocation, the quantized model executes on an ESP32-S3 with a minimal memory footprint and low latency. This compiler-based execution demonstrates a viable and energy-efficient solution for perception on low-cost robotic platforms.