From GPU to Microcontroller: Online Ridge Regression for Edge-Deployable Traffic Prediction

2026-06-16

AI Summary

This study addresses the challenge of deploying traffic flow prediction models on resource-constrained edge devices, where existing approaches relying on centralized GPU-based training are impractical. To overcome this limitation, the authors propose an ultra-lightweight linear method that constructs an independent ridge regression model for each sensor, enhanced with periodic feature engineering and updated online via recursive least squares (RLS) for adaptive learning—entirely eschewing neural network architectures. Evaluated on four PEMS benchmark datasets, the approach achieves state-of-the-art performance in three MAPE metrics and trails by less than 1% in the fourth. Notably, it enables cold-start training in just 7.4 seconds and single-prediction updates under 2 milliseconds on an ESP32 microcontroller, striking an exceptional balance between prediction accuracy and edge deployability.

Abstract

State-of-the-art traffic flow forecasting models, including Graph Convolutional Networks and graph-less MLPs, require centralized GPU training across all sensors, making them impractical for resource-constrained intelligent transportation deployments. We show that much of this complexity is unnecessary. A parametric analysis of the recent graph-less model GLMST reveals that reducing its internal embedding dimension from 64 to 4 degrades MAPE by less than one percentage point, suggesting that the model's effective capacity far exceeds what the task requires. Motivated by this finding, we replace the neural architecture entirely with per-sensor Ridge regression using horizon-aligned periodic features, combined with Recursive Least Squares (RLS) for online adaptation. With only 444 parameters per sensor (80x fewer than GLMST) and test-time online adaptation, our method achieves the best MAPE on three of four PEMS benchmarks, and remains within one percentage point on the fourth. Because each sensor's model is self-contained and involves only elementary linear algebra, the entire pipeline (training, inference, and online adaptation) runs on edge hardware without a GPU. An ESP32 microcontroller (160 MHz, 520 KB SRAM) completes cold-start training in 7.4s and each predict-and-update in under 2ms with zero heap allocation; a single Raspberry Pi 5 core completes cold-start training in 0.21s and each predict-and-update in 0.26ms.

Problem

• traffic prediction
• edge deployment
• resource-constrained
• GPU dependency
• intelligent transportation

Innovation

• Ridge Regression
• Recursive Least Squares
• Edge Computing
• Traffic Prediction
• Model Compression