The use of automatic milking systems (AM-systems) is increasing. The main attraction of AM-systems is that they allow farmers to increase productivity and improve their daily job routines. Milking by robot affects all aspects of farm management, and the daily life of the cow. Cows can go voluntarily to the milking station throughout the day and night, with the result that milking frequency and milking interval lengths are variable in time for each cow and between cows, depending on individual cow productivity, temperament, attitude and health, as well as on the traffic system and other managerial and environmental factors. Frequency of milking, milking interval length and the time of day when milking occurs all influence milk yield and composition, and the capacity and productivity of the system. To optimise machine utilisation, farmers often adopt some form of milking frequency stimulation, such as one-way traffic, feeding rewards, playing music, or drinking restrictions along the path. Thus, the willingness and spontaneous behavior of cows milked by robot are partially controlled by farmers’ behaviors.

The AM-systems are equipped with sensors and technical solutions to check animals, machine functioning and the milking process. The personal computer managing an AM-system automatically records and processes many readings every day, every milking, and for each udder quarter milked. This enormous amount of data can be useful for investigations into dairy cattle behavior. The data concern animal behavior, animal feeding and milk properties, and thus animal behavior can be related to health conditions and system characteristics.

Behavior-related computer data stored in AM-system databases usually include:

• number of visits to the milking station and number of passages through the selection gate
• time and duration of milking station visits
• number of refusals by the milking station