HeroLabs d.o.o. is a Slovenian technology company specialised in the development of advanced solutions in the fields of artificial intelligence, data analytics and sensor systems. In cooperation with EDIH DIGI-SI, the company validated an AI-supported solution for monitoring the behaviour and health status of cattle based on data from wearable accelerometer sensors.
As part of the EDIH DIGI-SI project, a comprehensive expert process was carried out, combining a research and analytical phase with the transition towards a production environment. The solution was based on time-series processing, feature extraction, the development and comparison of several machine learning models, and the interpretation of results in the context of detecting behavioural and health-related anomalies.
Special emphasis was placed on the reliability of data flows, time synchronisation of sensors, model validation under real-life conditions, and the definition of appropriate performance indicators. This established a solid methodological and technical foundation for the further development of a production-ready AI solution in the field of smart agriculture.
The process began with an analysis of the existing data flows and solution architecture of HeroLabs d.o.o. In the first phase, a proof of concept (PoC) was developed, covering the synchronisation of heterogeneous data, cleaning and preparation of time series, data segmentation, and extraction of relevant features for machine learning. Subsequently, various models were developed and evaluated, including Random Forest, XGBoost and logistic regression, with particular attention given to model interpretability and understanding the impact of individual features. The second phase focused on the transition towards a production environment, addressing the architectural aspects of integrating AI modules, automation of model training, performance monitoring and management of model degradation under real-life conditions.
Cooperation with EDIH DIGI-SI enabled HeroLabs d.o.o. to significantly reduce the risks associated with developing and implementing an AI solution in practice. The company gained a clearly structured, validated and explainable analytical process that supports the further commercialisation of the solution.
Key benefits include:
- a validated PoC with clearly demonstrated practical value,
- improved reliability and quality of data flows,
- strengthened capability to transition from the research phase to production,
- better understanding of animal behavioural and health patterns,
- increased innovation and technological maturity of the company.
The HeroLabs d.o.o. example contributes to the more effective adoption of artificial intelligence in agriculture and supports sustainable, data-driven livestock management. It helps bridge the gap between the development of advanced AI technologies and their practical application.
