THE INVESTIGATION ON ROBUST CONTROLLER OF O2 AND CO2 IN MINIATURE PROTOTYPE OF BLSS BASED ON SUGENO-TYPE ADAPTIVE NEURAL FUZZY INFERENCE SYSTEMS

THE INVESTIGATION ON ROBUST CONTROLLER OF O2 AND CO2 IN MINIATURE PROTOTYPE OF BLSS BASED ON SUGENO-TYPE ADAPTIVE NEURAL FUZZY INFERENCE SYSTEMS

© DaweiHu, HongLiu, MingLi, LingTong
© Государственный музей истории космонавтики им. К.Э. Циолковского, г. Калуга
Секция "К.Э. Циолковский и проблемы космической медицины и биологии"
2010 г.

The miniature prototype of bioregenerative life support system (BLSS) is a united system concluding human, silkworm, salad and microalgae acting together to perform a specific objective, one of significant problems is of how to maintain gases (mainly refer to O2 and CO2) to keep equilibrium at desired levels. The united system, however, is a complex system and have highly nonlinear and time-varying characteristics, hence modeling it precisely and based on model to investigate proper control algorithms for gases dynamics are almost impossible. The objective of our research is to resort unconventional method to develop a potent algorithm and realize it as actual controller for real-time application to regulate gases variation in system. The method is to establish Sugeno-type adaptive neural fuzzy inference systems (ANFIS) applying fuzzy logic based on expertise and experiences in combination of artificial neural network (ANN) algorithm to map the intricate relationship between the concentration of gases and light intensity illuminating on photo bioreactor where microalgae was cultivated continuously, and use ANN to indentify parameters. After designing control algorithm by digital simulations, conduct real-time simulation to validate the algorithm on the platform of MatLab/Simulink and LabView. The result show that ANFIS controller could regulating light intensity properly, which resulted in the growth rate of microalgae changed accordingly, and therefore indirectly regulated the dynamics of O2 and CO2 holding at prescribed levels and possessing satisfactory transient and steady performances. The conclusion could be drawn that the controller designed through ANFIS could robustly, self-adaptively affect the concentrations of O2 and CO2 in united system and the results provide a new way and feasible approach to control application for BLSS and its subsystems.