In-Context Learning for Zero-Shot Speed Estimation of BLDC motors

2025-04-01

AI Summary

To address the low speed estimation accuracy of sensorless BLDC motors under nonlinearities and parameter uncertainties—particularly in the low-speed regime—this paper introduces, for the first time, in-context learning to motor speed estimation, proposing a Transformer-based zero-shot real-time speed estimation framework. The method leverages only time-series electrical signals (e.g., voltage and current), pre-trained offline on simulated trajectories, and enables cross-operating-condition zero-shot transfer without online identification or fine-tuning. Experimental validation on a physical motor platform demonstrates that the average estimation error in the 0–300 RPM low-speed range is reduced by 42% compared to Kalman filtering, with inference latency under 1 ms. This significantly improves dynamic response during startup and enhances estimation robustness against model uncertainty and operational variations.

Technology Category

• Intelligent Robots: State Estimation
• Search and Optimization: Learning to Search
• Machine Learning: Multimodal Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Accurate speed estimation in sensorless brushless DC motors is essential for high-performance control and monitoring, yet conventional model-based approaches struggle with system nonlinearities and parameter uncertainties. In this work, we propose an in-context learning framework leveraging transformer-based models to perform zero-shot speed estimation using only electrical measurements. By training the filter offline on simulated motor trajectories, we enable real-time inference on unseen real motors without retraining, eliminating the need for explicit system identification while retaining adaptability to varying operating conditions. Experimental results demonstrate that our method outperforms traditional Kalman filter-based estimators, especially in low-speed regimes that are crucial during motor startup.

Problem

• Zero-shot speed estimation for sensorless BLDC motors
• Overcoming nonlinearities and parameter uncertainties
• Eliminating system identification with in-context learning

Innovation

• Transformer-based models for zero-shot estimation
• Offline training on simulated motor trajectories
• Real-time inference without system identification