Virtual fencing applications : implementing and testing an automated cattle control system
journal contribution
posted on 2017-12-06, 00:00authored byG Bishop-Hurley, David SwainDavid Swain, D Anderson, P Sikka, C Crossman, P Corke
Managing livestock movement in extensive systems has environmental and production benefits. Currently permanent wire fencing is used to control cattle; this is both expensive and inflexible. Cattle are known to respond to auditory and visual cues and we investigated whether these can be used to manipulate their behaviour. Twenty-five Belmont Red steers with a mean live weight of 270 kg were each randomly assigned to one of five treatments. Treatments consisted of a combination of cues (audio, tactile and visual stimuli) and consequence (electrical stimulation). The treatments were electrical stimulation alone, audio plus electrical stimulation, vibration plus electrical stimulation, light plus electrical stimulation and electrified electric fence (6 kV) plus electrical stimulation. Cue stimuli were administered for 3 s followed immediately by electrical stimulation (consequence) of 1 kV for 1 s. The experiment tested the operational efficacy of an on-animal control or virtual fencing system. A collar–halter device was designed to carry the electronics, batteries and equipment providing the stimuli, including audio, vibration, light and electrical of a prototype virtual fencing device. Cattle were allowed to travel along a 40 m alley to a group of peers and feed while their rate of travel and response to the stimuli were recorded. The prototype virtual fencing system was successful in modifying the behaviour of the cattle. The rate of travel of cattle along the alley demonstrated the large variability in behavioural response associated with tactile, visual and audible cues. The experiment demonstrated virtual fencing has potential for controlling cattle in extensive grazing systems. However, larger numbers of cattle need to be tested to derive a better understanding of the behavioural variance. Further controlled experimental work is also necessary to quantify the interaction between cues, consequences and cattle learning.