Lin Xu

Lin Xu

Auto-generated profile, based on publicly available information

Lin Xu

Hangzhou Institute for Advanced Study, UCAS, Hangzhou, China

Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
12 March 2026 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
WYWenHang Yang
WenHang Yang
LXLin Xu
Lin Xu
YZYangyang Zhang
Yangyang Zhang
SZShouzheng Zhu
Shouzheng Zhu
HQHongxing Qi
Hongxing Qi
JWJian-yu Wang
Jian-yu Wang
Temperature control of large-area blackbodies is crucial for the metrological performance of radiometric instruments, yet it remains challenging due to the complexities of traditional PID-based algorithms, which struggle with parameter tuning and fail to account for thermodynamic differences across channels in multichannel systems. To enhance the measurement accuracy and stability of such instruments, this study proposes a novel reinforcement-learning-based, parameter-free temperature control reinforcement-learning temperature controller (RLTC) method aimed at balancing heating rates across the blackbody surface and achieving precise temperature regulation. We first established a real temperature data acquisition system and developed a simulation model to generate temperature data, supplementing the real dataset. Thereby, we collected and labeled the first extensive temperature dataset for a large-area blackbody system, which includes heating, residual heat, and cooling characteristics. The RLTC algorithm was tested on a large-area blackbody ( $2.2\times 2.2$ m) with 64 temperature control channels. Compared to traditional PID algorithms, RLTC reduces stabilization time by 15.23%. Additionally, the integrated temperature balancing module reduces the maximum stabilization time difference between channels by 93.51%. The RLTC algorithm not only improves individual channel heating rates but also ensures more consistent heating across multiple channels, thereby providing faster and more balanced temperature adjustments in complex environments. This research introduces an advanced reinforcement learning (RL) control framework that directly enhances the metrological performance of radiometric systems, contributing to more precise and long-range detection capabilities for large-aperture optical detection applications.