Candidate: Farrar, Teneil

Development and evaluation of an on-animal sensor that is able to remotely monitor breeding animals and generate alerts during mating events.

Farrar, Teneil
University of Sydney
jessica.rickard@sydney.edu.au

Abstract

The following is a summary submitted with the application. It will be updated when the student’s abstract/thesis is received.

The overall aim of this research is to develop and evaluate an on-animal sensor that is able to remotely monitor breeding animals and generate alerts during mating events. The smart tag will utilise GNSS, accelerometer and Bluetooth technology to determine geolocation, behaviour and within-flock interactions of individual sheep. Our initial aims in the development of the smart tag are:

  1. To establish a signature of male breeding behaviour in accelerometer data; an
  2. To compare the quality of mating signatures obtained from smart tags fixed to the neck via a collar or to the ear.

Two trials will be conducted to satisfy the aims of the project: a controlled pen trial and a field trial.

Pen trial:
Rams (n=15) will be fitted with ear and collar smart tags to record and compare accelerometer signatures of a ram mounting a restrained ewe in an artificial pen environment. Each ram will be allowed to interact and mate with the restrained ewe for a total of three minutes. Tags will record a broad range of mating behaviours in addition to mounting including courting, sniffing and pawing. Video footage will also be recorded to validate and label the tag data against observed behaviours. This will be replicated twice for each ram over the course of a day. Data will be analysed via machine learning models (such as random forests) generated in R while expertise on data management, data mining and machine learning will be sought from the Sydney Informatics Hub, The University of Sydney. The collar and ear tag type smart tags will be evaluated for their precision and sensitivity. The type of tag that returns the most effective signature of mating activity will then be tested in a field trial.

Field trial:
Rams (n=5) will be fitted with the selected smart tag (ear or collar type tag) and introduced in-paddock to a flock of oestrus-synchronised ewes (n=30). Data will be collected wirelessly from the tags for a period of 12 hours during peak oestrus (48 hours post-CIDR removal). A remote surveillance system will record video footage for the validation and labelling of tag data with observed behaviours. Data will be analysed as per methods described in the pen trial.

The current study represents the first step in the development of a smart tag for the remote reproductive monitoring of sheep and will achieve the following outcomes:

  • The establishment of the first-ever signature of ram mating activity in accelerometer data
  • Determination of whether ear tags or collar type tags are more effective at monitoring mating activity
  •  Incorporation of the mating signature into the hardware of the smart tags by the Australian Wool Innovation and Digibale
  • The formation of a multi-disciplinary collaboration between the Australian Wool Innovation, Digibale, the Sydney Informatics Hub and the Animal Reproduction Group, University of Sydney.

The ability to detect mating events between individual sheep would allow for the identification of infertile and sub-fertile individuals, low libido rams, estimation of conception date and fetal age as well as the determination of oestrus onset. This would revolutionise the sheep production industry, providing producers with extensive and precise information on the performance, welfare and productivity of each of their individual sheep, enabling them to implement management decisions tailored to the individual animal, rather than the entire flock.