Real-time automatic integrated monitoring of barn environment and dairy cattle behaviour: Technical implementation and evaluation on three commercial farms

Due to increasing herd sizes and automation on dairy farms there is an important need for automated monitoring of cow production, health, and welfare. Despite much progress in automatic monitoring techniques, there is still a need to integrate data from multiple sources to create a comprehensive overview and accurate diagnosis of a cow’s state. To aid the technological development of data integration, a prototype of an open and customizable automatic system that integrates data from multiple sensors relating to barn environment and cow behaviour was developed. The system integrates data from sensors that measure barn climate (e.g., temperature, humidity, wind speed), air quality (e.g., CO2 concentration), water use and temperature, the moisture and temperature of the litter and cow behaviour (e.g., lying, eating, ruminating). An external weather system and video recording system are also included. The system’s architecture consists of four main elements: sensors, nodes, gateways, and backend. The data are recorded by sensors, then locally processed on custom-developed sensor nodes, and then transmitted via radio channels to local gateways that combine the data from multiple nodes and transmit them to distributed digital storage (“the cloud”) via a 3G/4G cellular network. On the cloud, the data are further processed and stored in a database. The data are then presented to the user continuously and in real time on a dashboard that can be accessed via the internet. In the design of the local wireless network, care was taken to avoid data packet collision and thus to minimize data loss. To test the system’s performance, the system was installed and operated on three commercial dairy cattle farms for one year. The system provided high data stability with minimal loss and outliers, showing that the system is reliable and suitable for long term application on commercial dairy farms. The system’s architecture, communication network, and data processing and visualization applications form an open framework for research and development purposes, allowing it to be customized and fine-tuned before being deployed as a management assistant on commercial dairy farms. Missing elements that should be added in the future are the integration of the data from the milking parlour and cow identification. Algorithms to integrate information from multiple sensors can be added to provide a comprehensive system that monitors all aspects related to cow welfare, health, and production automatically, remotely and in real time, thereby supporting farmers in important management decision-making.