Performance Comparison of Adaptive Neuro-Fuzzy Inference System (ANFIS) and Fuzzy Logic Based Microclimate Control System in Plant Factory

Abstract

Population growth and the reduction of agricultural land necessitate the
application of technology to enhance agricultural productivity. A plant factory
is an advanced agricultural technology that enables indoor plant production by
precisely regulating the microclimate for optimal growth. While fuzzy logic
algorithms have been applied for microclimate control, the use of an adaptive
neuro-fuzzy inference system (ANFIS) has not been explored. This research aims
to develop a microclimate monitoring and control system based on ANFIS and
fuzzy logic in a plant factory and compare their performance. The study involves
five stages: designing control system schemes, developing hardware and
software, testing, analyzing data, and comparing system performance.
Microclimate data from both systems were analyzed using the Mean Absolute
Error (MAE) metric and visualized through performance graphs. The results
indicate that the plant factory with ANFIS control achieved MAE temperature
values of 1.18°C and 1.48°C and MAE humidity values of 14.68% and 12.48%,
while the fuzzy logic control system yielded MAE temperature values of 1.68°C
and 1.60°C and MAE humidity values of 13.02% and 12.31%. Based on the
MAE values, the ANFIS control system demonstrated better temperature
regulation than fuzzy logic; however, neither system provided optimal
microclimate control. These findings highlight the potential of ANFIS for
improving temperature regulation in plant factories, suggesting the need for
further refinement and optimization of control strategies to enhance overall
system performance

The research consists of five stages, namely designing ANFIS and fuzzy logic control system schemes, designing hardware, designing software, testing and analyzing data, and comparing the performance of the two control systems. Microclimate data from both control systems were then analyzed to see their performance by looking at the MAE (Mean Absolute Error) value. Analysis is also done by looking at the graph of running results. The results showed that the plant factory with ANFIS control system showed MAE temperature values of 1.18^oC and 1.48^oC and MAE humidity of 14.68% and 12.48% while the plant factory with fuzzy logic control system showed MAE temperature values of 1.68^oC and 1.60^oC and MAE humidity of 13.02% and 12.31%. The plant factory with ANFIS control system provides better performance in temperature regulation based on the MAE value obtained but has not provided good performance, either using ANFIS control system or using fuzzy logic control system.