The UTS Robotics Institute delivers truly innovative robotic solutions to make Australian agriculture more competitive and productive in a global market, from using robots to augment productivity through to smart tools to extract/ generate business information and optimize supply chain management.
Agriculture robotics
Director of Robotics in Agriculture: Dr. Alen Alempijevic
Showcase under the robotics in agriculture nodes:
Trait estimation of livestock
Research Lead/s :
Dr. Alen Alempijevic,
A/Prof. Teresa Vidal Calleja
Cattle non-compliance with processor grid specifications causes a financial loss of $51M per annum for the sector. This problem is aggravated through the failure to improve production efficiencies because data on cattle traits are not available in real-time. The estimated target market for objective real-time assessment of live cattle to assist producers ‘meet market specifications’ for southern beef production would be 11.8M head. The industry opportunity is to have objective real-time assessment of live cattle to assist producers achieve greater production efficiencies by meeting “market specifications”. Potential benefits also exist in production efficiencies and being able to apply the technology to objectively assess welfare of cattle in terms of body condition score.
UTS:RI, in collaboration with NSWDPI and Local Land Services (LLS) is working on developing objective cattle assessment using 3D imaging and introducing the measurements into management tools such as BeefSpecs provide producers with real-time assessments that can be integrated into management tools such as the BeefSpecs drafting tool. The impact to beef producers is that real-time objective assessments of frame score, P8 fat and muscle score assist management decisions to improve beef production efficiencies and increased compliance rates.
See more via YouTube -
Research Strength : Sensing, perception and estimation
Research Partners : NSW Department of Primary Industries, Local Land Services & Meat and Livestock Australia
Research Team : Dr. Behnam Maleki
Trait estimation of carcasses
Research Lead/s :
Dr. Alen Alempijevic,
A/Prof. Teresa Vidal Calleja
Estimating lean meat yield for both the sheep and beef industries has the potential to increase profitability for the red meat industry by implementing a quantitative grading method into the meat processing sector.
Buildings upon previous research using 3D cameras to estimate phenotypic prediction traits in live beef cattle, we and extends the work to estimate protein in beef and lamb carcasses. In developing the technology for point of measurement carcass trait estimation, this project will enable quantitative measures for yield assessment and inputs to grading/condition scoring process of live animals. This will in turn make lifecycle monitoring of animal's traits and health monitoring in on farm management practices feasible in the near future.
Research Strength : Sensing, perception and estimation, Platforms
Research Partners : NSW Department of Primary Industries & Meat and Livestock Australia
Research Team : Raphael Falque, Cedric Le'Gentil, Lan Wu & Jasprabhjit Mehami
Efficiency and automation in wool handling
Research Lead/s :
Prof. Robert Fitch,
Dr. Alen Alempijevic,
Dr. Mickey Clemon,
Dr. Nick Bennett
Wool handling is currently a highly manual and repetitive task. The process currently contains many buffers, orbuffers or waiting zones. These waiting zones exist to enable human workers to perform multiple duties within the shed.
The aim of this project is to develop automated inspection, handling, classing, and baling for which no commercial product exists, thereby reducing and redistributing labour costs, increasing productivity and mitigating muscular strain injuries.
The design and development of such a system requires application-specific mechanical design, robotic control development, and sensor integration. The focus of the project is on inspecting and skirting wool after collection by a handler, classifying for baling, and baling into standard weight bales.
Research Strength : Sensing, perception and estimation, Control, planning and coordination, Platforms
Research Partners : Australian Wool Innovation
Research Team : Dr. Tim Patten, Solomon Ould
Body condition scoring of sheep
Research Lead/s :
Dr. Alen Alempijevic
Australian producers adjust the nutrition plan for groups of sheep based on their condition, called a Body Condition Score (BCS). Current best practice for assessing BCS is to manually palpate the lower lumbar region of a sheep's back to feel for fat, muscling, and the prominence of bone. This project is aimed at developing and validation a device and machine learning models to measure BCS more objectively.
UTS:RI has been working with DPI Researcher Dr. Gordon Refshauge in developing a digital device called “Score 4 Sure” that standardises scoring and allows for consistency in measurement for industry. It can be used on farm and can be produced affordably. It is less invasive and can be integrated with other animal practices.
The device has high repeatability in measurement, highly accurate, easy to use, faster than status quo, providing data to support individual animal management, including connectivity to auto-drafting equipment and decision support.
Score 4 Sure” takes the guess work out of sheep body condition assessment. This provides immediate confidence in the marketplace for the supply chain and greater precision for breeding programs. DPI is commercializing this tool with the private sector.
'Score 4 Sure' can be utilised globally by the sheep industry. DPI is fast-tracked this technology to give supply chain consistency, to give farmers knowledge to improve their breeding programs and to provide quality research data to understand many other sheep management challenges.
See more via YouTube -
"Score 4 Sure" Pitch - GATE 2019
Research Strength : Sensing, perception and estimation
Research Partners : NSWDPI (NSW Department of Primary Industry)
Research Team : Gordon Refshauge, Craig Borrows
