dr.ir. C (Congcong) Sun PhD MSc

My research objective is developing and implementing Intelligent Control Systems for optimal, sustainable and autonomous agricultural production.

Intelligent Control System

Intelligent control system is a class of control methods that utilize artificial intelligence approaches, such as neural networks and reinforcement learning algorithms, for control purpose. Compared to classical control approaches, intelligent control systems incorporate dynamic learning capabilities into the control process and explore optimal solutions beyond limitations of traditional methods. For example, intelligent control can operate based on data without requiring a precise model and can explore optimal solutions from a broader state space.

The motivation for using intelligent control arises from the complexity, dynamics, uncertainties, and variations inherent in agricultural systems. Intelligent control hold the potential to achieve optimal, reliable, and robust control. A complete control system involves several components: sensing, modeling, control, and planning. My contributions to intelligent control can be detailed across these four areas.

Sensing

Since intelligent control relies heavily on data, obtaining sufficient and high-quality data is crucial for the performance of modeling and control processes. In the sensing domain, I contribute to the design and development of optimal sensing systems using green sensors and soft sensing techniques. The objective of these sensing systems is to collect and provide high-quality data for modeling and control applications in an efficient and sustainable manner.

One notable project in this area is the 4TU Green Sensors project, led by Prof. Eldert van Henten and me, where we aim to develop and apply biodegradable soil sensors for sustainable agriculture.

Modelling

In addition to sensing, intelligent control plays a significant role in developing various models, including data-driven models and hybrid models that integrate neural networks. One of my ongoing projects involve learning animal behaviors in livestock buildings to improve interactions between animals and robots. This application is part of the NWO DurableCase project.

Control

As previously mentioned, intelligent control encompasses methods like neural network and reinforcement learning. In the control domain, I am working on projects related to: Climate control in greenhouses, vertical farms to enable efficient crop cultivation and Environmental control in livestock buildings to enhance animal welfare and reduce emissions.

Planning

Intelligent control also contributes to planning. In the NWO DurableCase project, I am developing optimal logistics planning for multi-agent harvesting robots to achieve autonomous harvesting with efficient energy use and reduced soil compaction. In the Engineering PhD Robotic Interactions in Livestock Systems project, I am exploring the optimal design of mission and maneuver planning for collaborative manure-removing robots in dairy barns.