APPLICATION OF KOHONEN NEURAL NETWORKS FOR LAWN CARE

摘要

The objective of the study was to determine if Neural Networks could be useful in reducing the cost of private lawn care. A Kohonen Self-Organizing Map (SOM) was developed to search for useful patterns which would allow reduction in caretaker cost. Data collection parameters were length of grass, density of grass, color of grass, presence of weeds, and presence of moss. Fuzzy Logic clustering was required for initialization of the Neural Network. The pattern recognition and mapping were successful. Cost reduction was demonstrated in reduced needs for watering, fertilizer, weed control, moss out, and lime.