A Practical Approach to Optimal Control Problem for Atomic Clocks.

Area of Science:

• Atomic clocks
• Control theory
• Signal processing

Background:

• Linear Quadratic Gaussian (LQG) control is a standard method for steering atomic clocks.
• Selecting LQG cost function weights presents a significant challenge due to uncertainties.
• Accurate modeling of steering algorithms for specific noise characteristics is often required.

Purpose of the Study:

• To develop an algorithm for calculating LQG weight and control loop gain coefficients.
• To bypass the need for complex system modeling in LQG control for atomic clocks.
• To determine coefficients using only initial frequency stability data.

Main Methods:

• Proposed a novel algorithm for calculating weight coefficients without system modeling.
• Utilized initial frequency stability information of reference and controlled oscillators.
• Analyzed the proposed LQG algorithm alongside two alternative gain coefficient calculation strategies.

Main Results:

• Successfully calculated control loop gain coefficients without prior system modeling.
• Demonstrated a method to determine necessary coefficients from initial frequency stability data.
• Provided comparative analysis with alternative control loop gain calculation methods.

Conclusions:

• The proposed algorithm simplifies LQG control for atomic clock steering.
• It effectively determines control parameters using readily available frequency stability data.
• This approach offers a practical alternative to model-based methods for atomic clock control.