Thesis subject

ELSA compliant Autonomous Greenhouse Control (MSc)

Artificial Intelligence (AI) has become an important enabler of the world-wide transition towards circular and sustainable food systems, but faces hurdles in ethics, legality, and societal acceptance. Recommender Systems (RS), though widely successful, prioritize efficiency and profit over societal well-being and lack ethical and legal considerations. The application of AI and RS in this thesis focuses on greenhouse automation, highlighting their potential impact on sustainable agriculture.

Short description

Greenhouses are an important source of healthy food, and with a growing population, decreasing number of expert growers, and serious challenges in sustainability, there is a rising interest in autonomous greenhouse control. Controlling a greenhouse is a complex multidisciplinary problem and it can be approached from many perspectives, with algorithms from classical control, AI or even expert-driven heuristics, with digital tools such as data platforms, dashboards, digital twins and RS. However, the ethical, legal and social aspects (ELSA) are largely missing when designing or implementing the algorithms and tools for autonomous greenhouse control. The goal of this thesis is to design, develop and validate a RS for autonomous greenhouse control that is ELSA compliant. One potential use case is greenhouse climate control considering sustainable energy transition.

Objectives

  1. Investigate the current state of the art in use of recommender systems (RS) in greenhouse control
  2. Identify the ELSA challenges in the design, development and deployment of RS for greenhouse control
  3. Design an ELSA-compliant RS for greenhouse control

Tasks

The work in this MSc thesis entails:

  • Literature review on present state of greenhouse control (including the user-greenhouse interaction), and the future potential of RS.
  • Assess the ELSA issues involved in building a RS, and propose solutions to amend the identified issues.
  • Design and develop an ELSA compliant RS.
  • Validate the design in a greenhouse. This task may also be supplemented through interviews with ELSA experts.

Requirements:

  • Courses: Greenhouse Technology (FTE31306), Programming in Python (INF-22306), (Optional), Big Data (INF-34306), Software Engineering (INF-32306) or Machine Learning (FTE-35306)
  • Required skills/knowledge: Basic data analytics/machine learning and willingness to learn new software tools, interest about autonomous greenhouses control and sustainable energy transition.

Contact person(s)

Varsha Kalidas (INF/ABE)(varsha.kalidas@wur.nl)

Tarek Alskaif (INF) (tarek.alskaif@wur.nl)