Digitization in livestock farming is developing rapidly. However, the complexity of the sensor systems and their data post-processing are still far from meeting the requirements of the multifactorial variables that determine animal behavior, especially in poultry and pig farming.
When analyzing animal behavior with regard to well-being, in many cases only the influence of individual parameters is described due to a lack of available methods, without being able to evaluate the complex interrelationships of the individual parameters.)
The human approach – i.e. the targeted use of one's focus and senses to analyze information – is highly important for smart digital information retrieval. Especially when dealing with livestock, this model allows to design sensor systems and to train computer-based algorithms in order to uncover multifactorial correlations and new relationships. These can then be used as a support for farmers, namely to propose actions to increase animal welfare.
The objective of the animal welfare projects at Fraunhofer IZFP is to clarify whether computer assistance can enable humans to sustainably improve animal husbandry.
To determine environmental data, sensors are standard in livestock buildings due to the manageable investment costs. However, data acquisition with specific regard to the animals themselves or their behavior is not yet a general standard. Today, the major role in this context is played by the economic value of the individual animal. This is also reflected in the number of scientific papers related to the different types of livestock. A lot of data is collected around the barn and the animals, which is mainly used to increase profitability. However, as animal welfare is becoming more and more an issue in the minds of consumers, an economic factor is gradually emerging here that the producing farms cannot ignore and that is gradually showing up in research.
The results of our research projects are particularly important for small and medium-sized enterprises, since the imbalance of animal welfare or the loss of even a few animals has a strong economic impact on the farms. As a result, the welfare of each individual animal is a high priority, even for smaller animals such as poultry, and farms are particularly keen to implement measures to maintain and increase animal welfare. For the successful applicability of the proposed approach in livestock monitoring, a high flexibility of the targeted solutions is required, as farms differ in their barn architecture and in equipment but also in environmental conditions. This challenge will be actively addressed in one of the projects and realized by easy-to-implement, customized training of the systems.
The innovation of the projects concerns the development of assistance concepts and systems that support farmers in their work, providing them with understandable holistic insights into the multifactorial causes of animal welfare imbalances under real-life conditions in their barns. Our intelligent, non-contact sensor systems convert the multi-dimensional correlations of their acquired data into information that is understandable to the user. The systems are trained based on the farmers' experience by linking the data to the condition of the animals.