A Reinforcement Learning Approach for Optimal Placement of Sensors in Protected Cultivation Systems

Abstract

Optimal placement of sensors in protected cultivation systems to maximize monitoring and control capabilities can guide effective decision-making toward achieving the highest levels of productivity and other desirable outcomes. Reinforcement learning, unlike conventional machine learning methods such as supervised learning does not require large, labeled datasets thereby providing opportunities for more efficient and unbiased design optimization. With the objective of determining the optimal locations of sensors in a greenhouse, a multi-arm bandit problem was formulated using the Beta distribution and solved by the Thompson sampling algorithm. A total of 56 two-in-one sensors designed to measure both internal air temperature and relative humidity were installed at a vertical distance of 1 m and a horizontal distance of 3m apart in a greenhouse used to cultivate strawberries. Data was collected over a period of seven months covering four major seasons, February (winter), March, April, and May (spring), June and July (summer), and October (autumn) and analyzed separately. Results showed unique patterns for sensor selection for temperature and relative humidity during the different months. Furthermore, temperature and relative humidity each had different optimal location selections suggesting that two-in-one sensors might not be ideal in these cases. The use of reinforcement learning to design optimal sensor placement in this study aided in identifying 10 optimal sensor locations for monitoring and controlling temperature and relative humidity.

A genetic programming-based optimal sensor placement for greenhouse monitoring and control

“…Therefore, the locations of the sensor that are featured in the resulting GP model are the optimal sensor locations required to facilitate monitoring and control of the entire greenhouse. Consistent with the findings in Uyeh et al (2021), the results show that different optimal sensor locations are representative of the entire environmental condition across different months and different micro-climate. Furthermore, the economic impact of the results is reflected in the observation that only eight (8) sensors are required to monitor and control the controlled cultivation system.…”

“…This work leverages on the same data used in (Uyeh et al, 2021;Uyeh et al, 2022a). The dataset contains temperature and relative humidity measurements collected remotely from a research cultivation-controlled system in Kyungpook National University, South Korea.…”

“…Although several techniques (Kubrusly and Malebranche, 1985;Alonso et al, 2004;Flynn and Todd, 2010;Yi et al, 2011) for optimal sensor placement have been proposed in the literature for different applications, some of the proposed methods are not directly applicable for highly non-linear setups (complex systems) such as controlled cultivation systems. In the context of controlled cultivation systems, optimization, and machine learning-based algorithms have been proposed (Yeon Lee et al, 2019;Wu et al, 2020;Uyeh et al, 2021;Uyeh et al, 2022b).…”

“…However, the aforementioned works are limited because they were investigated over a limited period of time which does not account for different planting seasons and weather conditions. To address these issues, (Uyeh et al, 2021) proposed a Reinforcement Learning (RL) based method to optimally place sensors in a greenhouses using a robust dataset which features different planting seasons. The dataset consists data from 56 dual temperature and humidity sensors distributed within a greenhouse.…”

Optimal Placement of Laying Hen House Temperature Sensors Based on Genetic Algorithm

“…Daniel D.U. [14] used reinforcement learning algorithm to determine the best sensor position of temperature and humidity sensors in greenhouse, which reduced the number of sensors from 56 to 10. Kang H. K. [15] applied neural network modeling to optimize the sensor position for measuring flue gas in industrial pipeline.…”