APPLICATION OF ADAPTIVE CONTROL SYSTEM FOR SATELLITE NAVIGATION OPTIMIZATION USING NEURAL NETWORK

Subject Area: Control System Engineering

Satellite Attitude Determination and Control Systems (ADCS) are critical for maintaining satellite stability and accurate orientation while in orbit. This paper focuses on the Nigeria ThinSat-1A and addresses challenges posed by nonlinear disturbances, particularly electromagnetic torque, which can destabilize the satellite’s angular velocity and orientation. To mitigate these effects, a Multi-Layered Neural Network (MLNN) was employed to develop an Adaptive Control System (ACS). Real-world data from the NigeriaSat-2 system at NIGCOMSAT were collected to train the MLNN-based ACS and to model the satellite's dynamic behavior for optimized control. The system was implemented and evaluated using a Helmholtz cage for geomagnetic field simulation and further refined on a hardware testbed. Additionally, vector data were utilized to estimate satellite orientation through a fast and reliable Attitude Determination Algorithm (ADA). The performance of the ACS was validated using Simulink with aerospace and neural network toolboxes. Results demonstrated that the ACS effectively minimized the impact of nonlinear dynamic torques, maintaining satellite stability and providing precise angular velocity control over multiple orbits