Gray Wolf Optimization-Long Short-Term Memory Based Temperature Estimation and Closed-Loop Control Method in...

Area of Science:

• Biomedical Engineering
• Analytical Chemistry
• Artificial Intelligence

Background:

• Accurate temperature control is a critical challenge in miniaturized chemiluminescence immunoassay systems for point-of-care testing (POCT), especially in aging populations.
• Conventional indirect temperature control methods lead to systematic deviations, hindering reliable diagnostic results.

Purpose of the Study:

• To develop an integrated solution for precise temperature regulation in microfluidic chemiluminescence immunoassay systems.
• To overcome the limitations of conventional control strategies and improve the performance of POCT devices.

Main Methods:

• Coupling multiphysics simulation (COMSOL heat transfer model) with data-driven modeling (Grey Wolf Optimization enhanced Long Short-Term Memory network).
• Developing an embedded hardware design for real-time temperature estimation and closed-loop control using a fuzzy PID controller.
• Validating the system's performance using cardiac troponin I (cTnI) detection assays.

Main Results:

• The GWO-LSTM model achieved temperature estimation errors within 0.3 °C.
• Steady-state temperature control accuracy reached ±0.2 °C under varying ambient conditions.
• Reduced coefficient of variation for cTnI detection from 10.77% to 2.69%, shortening incubation time.

Conclusions:

• The integrated approach effectively addresses the bottleneck of precise temperature control in microfluidic chemiluminescence analyzers.
• This technology provides robust support for developing next-generation high-performance POCT instruments.
• Improved temperature control enhances diagnostic accuracy and efficiency in point-of-care testing applications.