AI Method of optimize plant factory nutrient solution control system using reinforcement learning AI

• Main Author: BAN, BYUNG HYUN
• Format: Patent
• Language: English; Korean
• Published: 22.07.2022
• Subjects: AGRICULTURE; ANIMAL HUSBANDRY; CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; CULTIVATION OF VEGETABLES, FLOWERS, RICE, FRUIT, VINES, HOPSOR SEAWEED; FERTILISING; FISHING; FORESTRY; HORTICULTURE; HUMAN NECESSITIES; HUNTING; PHYSICS; PLANTING; SOWING; TRAPPING; WATERING

The present invention relates to a server for optimizing nutrient solution control of a plant factory using reinforcement learning AI connected to a plurality of nutrient solution control devices and nutrient solution measurement devices. The server for optimizing nutrient solution control of a plant factory using reinforcement learning AI comprises: a server communication unit receiving a plurality of nutrient solution information about a nutrient solution stored in each of a plurality of nutrient solution control devices, which is acquired by a nutrient solution measurement device, and transmitting a control signal corresponding to each of the plurality of nutrient solution control devices to the plurality of nutrient solution control devices; and a server analyzing unit, based on the received plurality of nutrient solution information, deriving the plurality of control signals having a nutrient solution concentration vector at t+1, which is the time following the current time t, directed to the same direction as an initial nutrient solution concentration vector by using reinforcement learning Al learned with at least one of the initial nutrient solution concentration vector for nutrient solution stored in each of the plurality of nutrient solution control devices, the nutrient solution concentration vector, and the control signal corresponding to the nutrient solution concentration vector. Accordingly, it is possible to achieve a nutrient solution supply condition for optimum cultivation of crops by analyzing a ratio of optimal nutrient solution materials supplied to the nutrient solution and controlling supply of the nutrient solution materials. 본 발명은 복수의 양액제어 장치 및 양액측정 장치와 연결된 강화학습 AI를 활용한 식물공장 양액제어 최적화 서버에 관한 것으로, 상기 양액측정 장치가 획득한 상기 복수의 양액제어 장치 각각에 저장된 양액에 대한 복수의 양액 정보를 수신하고, 상기 복수의 양액제어 장치로 상기 복수의 양액제어 장치 각각에 상응하는 제어신호를 송신하는 서버 통신부; 및 상기 수신한 복수의 양액 정보를 바탕으로 상기 복수의 양액제어 장치 각각에 저장된 양액에 대한 초기 양액 농도 벡터, 양액 농도 벡터 및 상기 양액 농도 벡터에 상응하는 상기 제어신호 중 적어도 어느 하나로 학습된 강화학습 Al를 활용해 현재시간 t의 다음시간인 t+1의 상기 양액 농도 벡터가 상기 초기 양액 농도 벡터와 동일한 방향을 가리키게 하는 복수의 상기 제어신호를 도출하는 서버 분석부;를 포함한다.