Approach

Relevance, Hypothesis, Challenges, and Approach

 

Relevance

Crops are the cornerstones of sustainable food production; however, crop production is currently under immense pressure for several reasons. First, the growing world population demands more food, and simultaneously more renewable resources for non-food products. Second, arable land is limited, thus an increase in space is not an option. The same is true for some essential nutrients for plant production and at the same time, climate change increases the severity of plant stress. Third, the environmental footprint needs to be reduced by reducing the use of agro-chemicals and minimizing soil degradation, water consumption, and pollution. Finally, society increasingly demands high-quality and organic food. Addressing such conflicting demands requires drastic changes in the way we produce crops. To increase crop yields while minimizing the environmental footprint under adverse conditions and in an economically reasonable way, substantial progress in our scientific understanding of novel technologies is required for which we need the following:

1. Improved understanding of crops to achieve increased yield potential with low nutrient and water inputs as well as the identification of the key, relevant complex traits through phenotyping;

2. More detailed and specific information about the current state of crops and their environment for sustainable crop production and farm management using novel technologies;

3. Greater knowledge and better understanding of improved models for handling complex relationships between crop growth and ecosystem, including the ability to formulate forecasts and simulations, so that crop management and breeding can be optimized;

4. Intelligent autonomous robots that enable new approaches to improve crop-soil management, targeted weed control, and plant protection for reducing the environmental impact and increasing yield and to improve breeding by phenotyping for more complex traits at a higher throughput;

5. Better adoption of these technologies for crop management and breeding through a systematic analysis of technological and socio-economic prerequisites and implementation strategies.

PhenoRob aims to address these needs via a technology- and data-driven approach that is integrated with experimental agricultural sciences and economic and ecological impact analysis. This is in line with Walter et al. (2017) stating: “smart farming is key to developing sustainable agriculture”.

Main Hypothesis

The main hypothesis of the Cluster of Excellence is that a major shift toward sustainable crop production can be achieved by two specific approaches: (1) multi-scale monitoring of plants and their environment using autonomous robots with automated and individualized intervention and big data analytics combined with machine learning to improve our understanding of the relation between input and output parameters of crop production, and (2) assessing, modeling, and optimizing the implications of the developed technical innovations in a systemic manner.

We will address this extremely challenging scientific endeavor outlined above using a technology- and data-driven approach. We will design new approaches to overcome technological bottlenecks in crop management and to support crop breeding. We foresee novel ways of growing crops and managing fields, and aim to reduce the environmental footprint, maintain the quality of soil and arable land, tackle the effect of climate change on crop production, and analyze the best routes to improve technology adoption.

PhenoRob is a novel approach characterized by the integration of robotics, digitalization, big data analytics, and machine learning on one hand, and modern methods of plant phenotyping, modeling, and crop production on the other. Various in-field monitoring activities of PhenoRob will generate large amounts of heterogeneous data on plants, crop stands, and the environment targeted towards crop breeding and management. This data will include geometric and semantic models of individual plants, multispectral images at different scales using optical and non-optical sensors, distributions of plant species in and around the fields, the status of plant health, disease detection, nutrient and water contents of soil, weather information, and additional soil-crop-atmosphere variables and fluxes in the environment. PhenoRob will address all aspects of technological challenges by acquiring and analyzing these data, identifying key traits, integrating them into models and innovative crop management concepts, and analyzing socioeconomic requirements for technology uptake, its economic and ecological impact as well as opportunities of technology adoption for practical crop management and breeding support.

Scientific Challenges

The main scientific challenges in such an approach are:

Challenge 1: How to identify relevant relations between inputs (plant genetic resources, crop management, and soil and weather conditions) and outputs (yield, plant growth, phenotypic performance, and environmental impact) in complex and changing field environments so that consequences of changes in input parameters on the output can be reliably forecasted to support the design of improved crop production systems.

Motivation of Challenge 1: Already today, plant and agricultural sciences generate extensive amounts of data; however, predicting better crop management and breeding practices based on these data is difficult. PhenoRob will determine the measurements through which relevant parameters can be obtained to improve breeding capabilities, design effective agronomic management, and improve the predictive power of models for future weather and near real-time management scenarios. PhenoRob will analyze how the relationship between input and output parameters can be identified using big data analytics and machine learning approaches.

Challenge 2: How to obtain relevant information about the current states and parameters of crop farming systems, and how to exploit this information for crop management, robotic in-field intervention and characterization, and breeding support.

Motivation of Challenge 2: Spatial and temporal data are essential for breeding and crop production; early stress detection, farm management, and predictive modeling improve agro- ecosystem management. However, the accuracy and throughput at which such data are generated by current technologies is not sufficient to enable the identification of novel key traits to support breeding and for targeted intervention in an automated way. Therefore, models with forecasting capabilities are needed to automate the management of crops and characterization of complex traits using robots. PhenoRob will enable the development of the relevant technologies. The ability to monitor each plant and address its needs will substantially reduce the chemical input.

Challenge 3: How to design the process of upscaling the innovations to support sustainable crop production by facilitating technology transfer, achieving compatibility to current crop production, and predicting the way in which new technologies will change the ecology and economy of crop farming systems, landscapes, and markets.

Motivation of Challenge 3: New knowledge and exciting technologies for crop production are the first step to achieve sustainability. The adoption of new technologies into practice is key to deliver the required impact; this will be explicitly addressed in our research. Monitoring the technological capabilities and an improved understanding of the underlying processes are not sufficient to meaningfully implement these innovations into crop farming systems. Therefore, we will overcome hurdles and develop efficient routes to adopt novel technologies into agricultural practice. We will identify the risks and benefits of these technological advances to stakeholders, and provide an argumentation to enable decision makers to weigh the risks, challenges, and potential of these technologies effectively.

Approach

Approach: The PhenoRob cluster will address the challenges stated above by combining recent progress and new innovations in technology (such as probabilistic robotics, modern geodesy, computer vision, machine learning, and big data analytics) with breeding and crop management-oriented implementation of distributed sensing, high-precision monitoring, cutting-edge plant and root phenotyping, novel soil and ecosystem research, advanced crop growth and agro-ecosystem modeling, as well as economic and ecological impact analysis.

We will make progress with respect to these challenges by addressing all four central research objectives of PhenoRob through the following set of actions:

(a) Systematically monitor essential parts of crop farming processes with sensor networks and mobile ground and aerial robots, providing spatially and temporally aligned, detailed data on individual plants and soil, nutrient, and ecosystem parameters, such as weather or vegetation biodiversity. This will enable more selective, effective, and sustainable management of input and output parameters of the entire agro-ecosystem comprising soil, crop, and environment and novel approaches to support breeding by quantifying novel traits and improving screening capacity in the field.

(b) Develop novel technologies that enable real-time in-field interventions, such as automated control of weeds and selective spraying and fertilization of individual plants in the field. First, intelligent autonomous systems will be developed to perform automated targeted interventions (which also addresses Objective 1). Second, agro-ecosystem modeling simulating the effects of certain management actions will allow for selecting the most effective approaches. This will enable new procedures that reduce the environmental footprint of crop production by lowering the amount of applied agro-chemicals, improving water and nutrient use efficiency, and minimizing soil degradation and erosion.

(c) Apply and extend modern machine learning techniques to analyze large amounts of acquired data, improve our understanding as well as our models of plant growth, and nutrient and water use efficiency, and identify correlations between inputs and outputs.

(d) Predict the expected impacts of novel approaches on management decisions at the farm level, and investigate the requirements for broad technology adoption considering market interactions and benefits that result from upscaling the developed technologies.

In particular, the approach of addressing crop and agro-/ecosystem modeling in combination with big data analytics and large-scale machine learning is especially novel, innovative, and promising, but also carries high risk. It is an orthogonal approach compared with omits and molecular approaches to modeling. We believe that this cluster will establish the foundations for a new direction of data-driven, sustainable crop production.