Do cows come when called? Training cows to visit the milking robot in response to a cue
Francesca Pells Johansen, Gareth Arnott, Stephanie Buijs
- 发表年份
- 2025
- 引用次数
- 2
摘要
Voluntary movement of cows through the milking robot of an automatic milking system (AMS) is essential to its productivity. However, challenges remain regarding optimising visit frequency, particularly for systems aiming to combine pasture-based grazing and AMS. Here we explore an approach to enhance visit frequency through associative learning of an audio and tactile cue to encourage cows to voluntarily visit the milking robot. The current study aimed to assess 1) training stage reached and response to cue, 2) the effects of training upon milking frequency, 3) the effects of feed choice upon training stage reached and response to cue, and 4) the effects of cow location, light conditions and weather conditions upon the cows’ response to the cue. Twenty-eight cows from a larger herd of forty-four were trained in two batches (early and late). Cows were feed choice tested during week four of the study. The early group was trained to respond to the cue (move to the milking robot in response to an audio and tactile cue) during weeks 11-12 and tested in week 13, and the late group was trained during weeks 14-15 and tested in week 16. Descriptive statistics and GLMMs were utilised to answer the study questions. Four of the cows did not progress from the lowest level of the predefined five-step training plan. None of the cows completed the training plan, but most cows reached steps three (nine cows) or four (thirteen cows). Twenty cows correctly responded to the cue (trafficked to the milking robot) once or more, but none responded to the cue more than five times (of a possible 10). Milking frequency was higher in both groups during training periods compared to testing periods. There were significant associations between feed consumption during the feed choice test and response to the cue, where cows with the feed they consumed the most of (during the feed choice test) offered in the robot achieving an improved response to the cue. Finally, cows were more likely to correctly respond to the cue during dark hours. Training cows to approach the robot in response to an audio and tactile cue is a promising intervention to increase productivity in AMS. However, our study was unsuccessful in achieving this, and further development of training protocols and research on factors influencing training and cue response is needed to improve success. • Achieving optimal milking frequencies is a challenge in automatic milking systems • We aimed to train cows to visit the milking robot in response to an audio cue • No cows completed the training plan, and the best result was 5/10 correct responses • Cows visited the robot more frequently during training compared to during testing • Links were found between personality traits, feed preference, and task performance
关键词
相关论文
Statistical Learning Theory
Yuhai Wu, Vladimir Vapnik
1999
Fractional Differential Equations
Igor Podlubný
2025
Applied Nonlinear Control
Jean-Jacques Slotine, Weiping Li
1991
Genetic Programming: On the Programming of Computers by Means of Natural Selection
John R. Koza
1992