My name is Congcong Sun, I am an engineer of Computer Science and Automatic Control. After joining Wageningen University from May of 2021, I have been focusing on Intelligent Control System for optimal, sustainable and autonomous agro-food production.

Intelligent Control System

Intelligent control system is a class of control methods which use various artificial intelligence approaches like neural networks, Bayesian probability, fuzzy logic, reinforcement learning, evolutionary computation and genertic algorithm. Comparing with classical and modern control approach, intelligent control system empowers dynamic learning capacities into control process and explore optimal solutions out of boundaries set by classical and modern control approaches. Such as, intelligent control can based on data, does not always need a good model, intelligent control can explore optimal solutions from a wider state space. 

The motivation of using intelligent control is because, comparing with other domains, agricultural and food production is complex full of dynamics, uncertainties and variations. Intelligent control system has potentials to achieve optimal, reliable and robust control. As a complete control system used to involve sensing, modelling, control and planning different aspects, so that my contributions to intelligent control can be elaborated in sensing, modelling, control and planning four different pieces.

Sensing

As intelligent control is mostly based on data, sufficient and high quality data input is crutial for performance of modelling and control processes. In the sensing part, I am contributing to design and develop optimal sensing systems using green sensors, soft sensing techniques, optimal sensor placement and optimal sensor usage methods, etc. The objective of the sensing system is mainly collecting and providing high quality data for modelling and control applications, in an efficient and sustainable way. In this part, Prof. Eldert van Henten and I have won a 4TU Green Sensors project, where we are planning to develop and apply biodegradable soil sensors for sustainable agriculture.

Modelling

Besides sensing, intelligent control can also contribute to develop different models, like data-based model, hybrid model using netural networks, etc. My ongoing projects include learning animal behaviors in livestock buildings for better animal robot interaction. This applications is involved in the NWO DurableCase project.

Control

As presented previously, there are different intelligent control methods including fuzzy logic, reinforcement learning and genetic algorithms. In the control section, I am currently working on projects of climate control in greenhouse, vertical farm and plant factories to have efficient crop cultivation; environment control of livestock building for animal welfare and emission mitigation. As well as optimal control of irrigation systems to have efficient usage of water resource, etc.. Reinforcement learning and Genetic algorithms are my most commenly used approaches.

Planning

Intelligent control can also contribute to planning. In NWO DurableCase project, I am now working on optimal logistics planning of multi-agent harvesting robots in order to achieve autonomous harvesting, with optimal energy usage, and less soil compaction. In the PDeng Robotic Interactions in Livestock Systems project, I am now exploring optimal design of mission and maneuver planning for collaborative manure removing robots in dairy barns.

DurableCASE: Durable Cooperative Agrobotics Systems Engineering

DurableCASE is a collaboration to develop solutions for collaborative robotic vehicles in the agricultural sector. DurableCASE stands for Durable Cooperative Agrobotics Systems Engineering: sustainable solutions for collaborative robots in the agricultural sector. 

The use of robots as a replacement for labor in this sector is increasingly becoming a necessity due to a lack of manpower. The perspective offers many possibilities, for example in the field of sustainability. Multiple compact robots can take over the tasks of one large machine. That saves soil compaction. It also provides greater operational reliability: if one robot fails, other robots can take over tasks. The robots have to work well together. 

DurableCASE develops solutions for communication between collaborating robots. The goal: robust cooperativity, where robust stands for safe, secure & performant. In other words, safe, secure and with optimum performance, under all circumstances. For robotics, there is a lot available in the rich ecosystem of ROS (Robot Operating System). DurableCASE makes grateful use of this. 

AutoFarming: Autonomous production control of greenhouse farming system using Reinforcement Learning

Greenhouse is an important protected horticulture system in providing fresh food for the growing global population. However, it is also a big resource consumer, such as the large amount of energy usage from LED lighting and heating to maintain ideal growing climate for the crops. The main objective of greenhouse production control is realizing resource-effective crop growth and development through operating indoor climate actuators (e.g. lighting, heating, CO2 dosing, ventilation, screening, watering). In high-tech greenhouses, the modern sensing, control and computing systems are equipped but the operating set-points for the climate actuators are still rely on growers. However, the number of experienced growers is rare while the scale of greenhouse production system is expanding worldwide. Therefore, using advanced techniques to control the greenhouse production system with more optimality and autonomous in a reliable and robust way is in need nowadays challenge.

Reinforcement learning is the kind of learning system, which uses continuous feedback to adjust its own actions to obtain the best. It is also a dynamic control strategy which can update automatically the current control algorithm (policy) through incorporating newly developed knowledge learning from historical and real time data. The motivation of this project is developing autonomous production control of greenhouse farming systems using reinforcement learning, aiming to optimize the production system with desired growing climate, efficient resource usage, and at the same time adaptability to variable conditions (amongst individual plants, species and growing stages).

In this project, you will develop control-oriented crop-climate model based on crop/climate sensor measurements; develop autonomous control scheme of greenhouse production systems using reinforcement learning to achieve efficient crop growth/yield and energy usage. All the proposed approaches will be validated in real greenhouse for lettuce production. The performance will also be explored and analysed with state-of-the-practice.

SmartFarming: smart vertical farming production using plant monitoring and self-learning control

Vertical farming (VF), the practice of growing crops in vertically stacked layers under a highly controlled environment, represents a solution to sustainably produce high-quality fresh vegetables at any time of the year, especially in urbanized areas where the cultivable lands are scarce. Despite its considerable potential, VF development is currently limited by the high energy use and operating costs, increasing the selling price for the public and reducing the profit margins for shareholders. Therefore, advanced climate control techniques taking advantage of the predictability of plant behavior in VF could improve plant growth and development, whilst decreasing energy consumption. Automation also reduces the dependence on experienced growers that are not in sufficient number to fulfill the growing needs of the VF industry.

Self-learning control is a dynamic control strategy which can automatically update the current control model through incorporating newly developed knowledge learning from historical and real time data. Using self-learning control, the control model of climate in VF could be learned and updated from monitoring measurements with optimal solution adaptively for different stage of plant development. Reinforcement Learning (RL) is the only machine learning algorithm with both learning and optimization abilities, which can give reward feedback for each result to the algorithm to learn from and improve in future. 

This project focus on developing a proof of concept of a smart VF production system using real-time plant monitoring and self-learning control to achieve fully autonomous climate control, as well as optimization of crop growth, energy consumption and operational cost. This project will combine the expertise of two chair groups HPP and FTE.

4TU Green Sensors

Population growth, climate change, resource depletion and soil degradation are placing huge challenges on agriculture. Bringing advanced sensing and communication techniques into agriculture is urgently needed. Successful agriculture decision making requires characterization of soil heterogeneity (e.g. soil compaction, moisture, phosphorus, nitrogen) in real-time in order to optimize precisely time, location and degree of agricultural operation (e.g. sowing, irrigating, fertilizing and harvesting). The necessity of sustainable development and environmental friendly electronic devices forces manufactures to use green sensors. These green sensors are biodegradable and fit naturally into the cycle of nature, with a period of life followed by death and a natural recycling process, which are safe for soil, plants and groundwater.

This project will develop a biodegradable soil sensor system to monitor precisely soil status in terms of soil compaction, moisture, phosphorus and nitrogen, without leaving hazardous e-wastes to environment. In this research, the following research questions will be addressed: 1) How to use green materials to develop biodegradable sensors? 2) How to design IoT and ICT systems for wireless green sensing systems? 3) How to model and control biodegradable sensing system using data and AI-based algorithms? 4) How to realize optimal and sustainable agricultural productions using biodegradable sensors?

EngD project

Design of mission and maneuver planning for collaborative manure removing robots in dairy barns

Automation in the dairy sector is becoming more and more popular. However, up to now, most of the planning and operation of robotic systems have hardly or not involved cow behavior, amount of manure at a specific place and time and cooperation between robots.The objective of this EngD project is to design, develop and plan optimally mission and maneuver for collaborative manure removing robots in dairy barns in order to 1) improve performance of the robotic systems; 2) reduce ammonia emission and 3) improve animal welfare. 

This project is collaborated with the Wageningen Livestock Research.

2019-2020    Postgraduate in Data Science and Big Data, University of Barcelona, Spain

2011-2015    Ph.D. in Automation, Robotics and Vision, Institute of Robotic Industrial (IRI, UPC-CSIC), Spain   

2008-2011    M.Sc. in System Engineering, Tongji University, Shanghai, China

2004-2008    B.Sc. in Computer Science, Nanjing Audit University, Nanjing, China