PoultryTech Newsletter, Volume 26, Issue 1, Spring, 2014

Background

Poultry growout houses require daily monitoring to ensure bird health and proper house operation. Current practices require growers to walk through multiple houses several times a day. During these walks, the growers observe the birds’ movements and energy levels and even listen to their vocalizations. They also inspect feed/water and ventilation systems and remove cadavers. These activities are crucial to ensuring the well-being of the birds. However, as one may imagine, they are also time-consuming. Could an automated monitoring system be the answer? Researchers with the Georgia Tech Research Institute’s Agricultural Technology Research Program are exploring the use of small, low-cost unmanned air and ground robots to monitor conditions within the growout house.

 

Project Overview

The initial concern for the researchers is to first do no harm, explains Colin Usher, project director and senior research scientist. Therefore, the primary goal of the project is to explore the impact on animal well-being caused by operating robotic vehicles in the growout house. The researchers are also investigating the feasibility of using various sensing modes to extract information from the environment to both control the vehicles and provide information to the farm manager. They are collaborating with fellow researchers in the University of Georgia’s (UGA) Department of Poultry Science.

 

Air and Ground Robots

The team has outfitted commercially available air and ground robots with 2D and 3D sensors/cameras. These robots were recently operated in an experimental growout facility at UGA’s Poultry Research Farm in Athens, Georgia. Video and audio data was collected on 500 birds for a full growout cycle (6 weeks) in a miniature-scale house.

 

The researchers developed and are currently analyzing a set of metrics that allows for quantitative analysis of bird reactions. These metrics include average avoidance distance, flight response, and recovery time. The avoidance distance metric is the average distance of the birds from a stimulus such as the robots or a human. Flight response contains data such as the average speed a bird runs away from a stimulus, as well as how far away the birds exhibit a reaction. The recovery time metric involves calculating mass of motion and determining how long it takes before the birds resume normal behavior after a stimulus leaves. All of these metrics are calculated for each robot and for humans.

 

Initial Results and Next Steps

According to Usher, preliminary results indicate that there is not a negative impact on the birds due to the operation of the robotic systems when compared to their reactions to humans. The graph below illustrates the average distance metric over the 6-week growout cycle with the air, ground, and human stimuli. Here, the data indicates that the birds are more comfortable with the robotic systems than with a human in the growout environment.

 

Armed with the confidence that the operation of robotic systems is not detrimental to bird well-being, the researchers are preparing to enter the next phase of the research effort. Phase 2 efforts will focus on three separate development goals: (1) develop and test autonomous routines to allow the robots to navigate a poultry house environment, (2) develop sensing systems to inspect equipment and birds, and (3) design and develop manipulators and a robotic system capable of carrying out various tasks.

 

“The goal of such a system would be to reduce the requirement of a human entering a poultry growout house as much as possible,” says Usher.

 

 

 

Groung Robot

 

 

Combined Distance Data Chart

 

 

Air Robot