【摘要】 温度波动对垂直腔面发射激光器（VCSEL）的波长影响显著，在外界温度骤变的复杂环境下采用传统PID进行温度控制存在超调量大、抗干扰能力弱和自适应性能差等问题。文章通过仿真分析VCSEL温控系统的三维稳态热分布特性，建立温度控制数学模型，提出一种基于反向传播（BP）神经网络优化的PID控制方法。在MATLAB/Simulink环境中，将所提出的方法与经典PID控制、模糊PID控制进行对比仿真实验。结果表明，该方法利用BP神经网络实现在线整定PID参数，有效抑制了系统超调，显著提高了系统对环境变化的适应能力与鲁棒性，从而整体提升了温控性能，能够满足高精度原子磁强计对激光器温度稳定性的严格要求。更多还原

【Abstract】 Temperature fluctuations significantly impact the wavelength of vertical-cavity surface-emitting lasers(VCSELs). Traditional PID control in complex environments with abrupt external temperature changes suffers from issues such as large overshoot, weak disturbance rejection, and poor adaptive performance. This paper simulates the three-dimensional steady-state thermal distribution characteristics of the VCSEL temperature control system, establishes a mathematical model for temperature control, and proposes a PID control method optimized using a backpropagation(BP) neural network. Within the MATLAB/Simulink environment, the proposed method is compared with classical PID control and fuzzy PID control through simulation experiments. Results demonstrate that this approach utilizes a BP neural network to achieve online tuning of PID parameters, effectively suppressing system overshoot while significantly enhancing the system??s adaptability and robustness to environmental changes. Consequently, overall temperature control performance is improved, meeting the stringent requirements for laser temperature stability demanded by high-precision atomic magnetometers.更多还原