Follow me on:
Serving as a full professor of Biosystems Engineering at Wageningen University & Research I am passionate about investigating and developing technology that interacts with nature in the context of agri-food production. Technology has played a crucial role in agriculture since its very beginning and novel technologies will definitely be needed to guarantee sustainable agricultural production of food, feed, fibres, fuels etc. for generations to come.
When it comes to application domains within the agri-food chain, I do not have a specific preference for individual domains like arable farming, protected cultivation or livestock. Being a systems thinker, I prefer to think in terms of similarities rather than differences between these application domains. The systems approach allows me to think in terms of generic technical concepts or capabilities shared by all these application domains. This systems approach also opens doors to completely different application domains and associated technologies that act as a great source of inspiration. For me a project is successful once technology has been tested in a practical environment up to proof of principle level, that means Technology Readiness Levels 5 to 6. For sure, when it comes to details, implementations of technology in products will differ. That is a key aspect of product development. I leave that to industry.
In 1987 I completed a MSc in Agricultural Sciences with honours (cum laude) at Wageningen University, followed in 1994 by a PhD in Agricultural and Environmental Sciences at Wageningen University with a thesis entitled ‘Greenhouse climate management: an optimal control approach’. After a career within the contract research organization of Wageningen University & Research, I continued my career in academia in 2005 as head of the Farm Technology Group within Wageningen University until the end of 2021. Currently I am a personal professor in Biosystems Engineering within the Farm Technology Group.
Research output includes (co-)authorship of 2 books, 5 book chapters, 2 special issues on agricultural robotics of the Journal of Field Robotics and more than 150 papers in peer reviewed journals and conference proceedings. So far, they yielded a h-index of 28 in Web of Science, 32 in Scopus, 40 in Google Scholar and 36 on Research Gate. Besides being a member of the editorial boards of Biosystems Engineering, Computers and Electronics in Agriculture and International Journal of Agricultural and Biological Engineering, I serve as a co-chair of the Technical Committee on Agricultural Robotics of euRobotics and as a co-chair of IEEE-RAS Special Interest Group on Agricultural Robotics and Automation.
Below follow some examples of my research activities.
Agro Food Robotics
One line of research focusses on agro food robotics. I have a keen interest in deploying recent advances in robotics technology, machine vision and artificial intelligence in agriculture. Usually, that proves to be a challenging task. Main academic challenges in agro food robotics are in the inherent variation in nature, fruits, animals, plants, having different shapes, sizes and colours. Also the working environment of the robot in agriculture is poorly structured. Yes, we grow crops in rows, but that's about it. Fruits never appear at the same time and at the same place and the size and structure of crops continuously changes. Last but not least, climate conditions in agriculture affect performance of technology. Think of sun light changing during the day, and don't forget the impact of rain, dust and dirt on technology.
The proof of the pudding is in the eating, so in all my projects I have a strong desire to bring technology to TRL levels 4 to 6 and test it in environments that are as close as possible resemble agricultural practice.
For me, the agro food robotics strand of research started in the late nineties of the past century with the development of an autonomous harvesting robot for cucumbers of which a proof of principle test was done in a greenhouse in 2001. Soon it was followed by a robotic device to remove leaves in a high-wire cultivation system for cucumbers. Interestingly enough both functions were implemented on the same robotic platform, only the software and gripper were changed. That multi-functionality embedded in the same robotic system was a first timer, at least in agriculture.
The cucumber harvester was followed by harvesters for sweet pepper in the EU projects CROPS and Sweeper. Both systems were tested in a commercial greenhouse. In both projects my BSc, MSc and PhD students participated during their thesis projects.
Hedge trimming robot
The Trimbot2020 project researches the robotics and vision technologies to prototype the first outdoor garden trimming robot. Within this project my PhD student specifically focusses on the planning of a path for the cutting device. This is a coverage path planning problem in 3D space that proves to be both inspiring and challenging.
Precision Weed control
Machine vision based precision weed control has been on the research agenda for a long time. Having as main aim the reduction of the use of crop protection chemicals, machine vision and AI technologies are deployed together with precise chemical application. Building on previous research in the INTEREG IVa project SmartBot, this line of research was continued in the ECHORD++ project SAGA as well as the current SMARAGD project.
Floor egg collection
Another example of work in robotics is the autonomous collection of floor eggs in poultry farming, for which we developed a robot called PoultryBot. Since battery cage systems are prohibited, laying hens spend their lives in free range systems in which they have much more freedom to strole around. Usually, hens will lay their eggs in specifically designed nests, but occasionally chickens lay an egg on the floor. These eggs have to be collected by the farmer once or twice a day, a demanding task. In the PoultryBot project a prototype robot was developed that navigates through the farm house and collects the floor eggs. The robot was successfully tested in a close-to-real-production environment.
FlexCRAFT - Cognitive Robots for Flexible Agro Food Technology
At the end of 2018 a consortium consisting of Wageningen University (coordinator), Eindhoven University of Technology, Twente University of Technology, Delft University of Technology and the University of Amsterdam received funds from NWO to work on generic principles in active perception, world modelling, motion planning and grasping and manipulation that are desperately needed to move the field of agro food robotics forward. For sure, many challenges lie ahead of this team, not in the least, because the consortium aims to demonstrate the generic nature of the developed principles and technology in three demonstrators including Food Processing Robotics, Food Packing Robotics and Greenhouse operation. Ten PhD's and three postdocs will be the main workforce in this program. A user group consisting of 15 companies provides support to this initiative to guarantee the required knowledgfe transfer to industry.
(Optimal) control of agricultural production processes - Precision Farming
This is a second strand in my research portfolio. It goes back to the times I was working on my Master thesis and my PhD thesis in the late eighties, early nineties of the previous century. For sure this research does not yield the fancy pictures and videos commonly produced in robotics research, it was and still is a challenging research topic.
What is so challenging then? As indicated in the accompanying diagram for greenhouse crop production, there appear to be two control loops when it comes to using the greenhouse climate for controlling the crop production process. The inner loop is implemented on a process computer nowadays and aims to realize preferred settings of the climate variables like temperature, humidity and carbon dioxide provided by the grower using measured data of the climate and the heating valves, screens and ventilation windows as actuators. But growers do not produce climate, they produce fruits and flowers. So the outer loop is more interesting. In that loop the grower identifies the correct target values for the indoor climate based on visual inspection of the (state of the) crop, the current indoor and outdoor climate. That requires long term planning skills, because the crop responds only very slow to changes in the environment whereas the climate changes rapidly in response to changing valve settings. My research addresses this outer control loop. It requires modelling of a non-linear dynamic system with different response times. Variation in the crop and uncertainty in for instance the weather adds to the complexity of this dynamic management or control problem. Efficient use of energy or improving economic return are key drivers for this research.
Ongoing research in this strand includes the NWO funded program Led it be 50% (in Dutch) and the NWO funded project FlexCROP.
Essentially, this line of research addresses the core of precision farming, intending to do the right thing at the right place at the right time, a challenging sensing and decision making process. Substitute in the previous greenhouse example the climate with soil and a tomato crop by grass, one obtaines a similar kind of system and for that system the question might be how to irrigate such that the available water is used as effective as possible, while realizing a healthy and productive crop. Ongoing research in my group addresses optimal irrigation scheduling within this precision farming framework.
(Optimal) model based design of agricultural production systems
Related to the previous line of research, is a line of work addressing optimal model based design of agricultural production systems. In a way developing and implementing a new system design of a production system is a kind of control problem with a very long implementation time, that can last many years. An inspiring first step in this line of work was the PhD project called The Adaptive Greenhouse, in which the question was addressed which greenhouse fits best to local conditions. This question was based on the observation that a wide variety of greenhouse systems does exist around the world and a one size fits all approach has proven not to be a feasible way to go. Using a dynamic model of the greenhouse crop production system and a suitable optimization procedure, different designs could be generated for different locations on earth. This was very valuable first step. In another PhD we investigated the on-farm logistics of work force and machines in a high-tech greenhouse using discrete event modelling. We barely scratched the surface in this field and many fundamental questions still remain. More will follow.