Â鶹ÒùÔº

As the dairy industry increasingly adopts automation with the use of sensors and robotics, researchers at Texas A&M AgriLife are helping producers harness this evolving technology to help optimize production and improve the health and well-being of dairy cattle.

Sushil Paudyal, Ph.D., an assistant professor of dairy science in the Texas A&M College of Agriculture and Life Sciences Department of Animal Science, is helping to spearhead these efforts. He leads research that applies artificial intelligence, AI, and machine learning technology to gather advanced, real-time data on farms, developing systems that support earlier disease detection, informed decision-making and cost-effective adoption of robotics.

"Sensor-based systems, AI and real-time analytics are transforming how dairies make everyday decisions," Paudyal said. "But to be effective, these technologies must be adaptable, updatable and tailored to individual farm needs."

Building the future of data-driven dairy

Paudyal's lab focuses on practical, technology-based research that helps producers stay ahead of evolving challenges, including rising heat stress and changing labor dynamics. Technology-driven models can detect diseases early, enhance cow management and improve efficiency on dairy farms. Already, he has successfully deployed models to detect lameness, mastitis and heat stress in individual dairy cows, using advanced analytics algorithms trained on camera images and behavioral cues.

"Right now, we're developing farm efficiency models based on machine learning for robotic milking systems, aiming to pinpoint idle time and milking failures," he said.

At the recent in Lincoln, Nebraska, Paudyal and his team highlighted some of their research:

• Evaluating Effects of Heat Stress on the Efficacy of Robotic Milking Systems—This study, led by doctoral student Rajesh Neupane, developed machine learning and computer vision models which determined that managing heat stress is crucial in robotic milking systems, as it significantly affects cow flow, robot usage, milk yield, feed intake and milking performance. Cows in cooler conditions perform notably better. Mitigation strategies, such as improved cooling, ventilation and adjusted feeding protocols, are critical to maintaining productivity and animal welfare.
• AI-Driven Quantification of Heat Stress and Mastitis in Dairy Cattle—This study outlines an automated video monitoring-based system that uses AI to assess heat stress and mastitis in dairy cows through behavioral cues, enabling real-time, scalable monitoring that improves animal welfare and farm efficiency.
• Using Computer Vision to Detect Different Digital Dermatitis Conditions—This research explores the latest advancements in computer vision and machine learning approaches for early detection and prediction of digital dermatitis in dairy cattle, focusing on their potential for real-world application. Computer vision enables early, accurate and noninvasive detection of digital dermatitis in dairy cows, improving health monitoring and reducing dependence on subjective visual scoring.

Innovation designed for real-world use

One of Paudyal's goals is to create noninvasive, cost-effective diagnostic tools that work across diverse production systems. For example, some rely on camera-based systems in place of physical sensors to monitor large groups of cows, lowering start-up costs while expanding impact.

"We are developing sensors in our lab that can help detect diseases without collecting invasive blood samples or milk samples," Paudyal said. "They will monitor behavior and physiological variables to determine sick cows."

His team is currently developing a "DairyBot" virtual assistant, a generative AI tool that will enable producers to evaluate farm data and lab results, as well as ask questions about feed decisions while using AI to interpret herd data in real-time.

"They will have a real-time advisor with a vast domain of knowledge that can pull from their farm's data and dairy-specific literature," Paudyal said. "It won't replace the vet or nutritionist, but it will empower and support them for informed decision-making."

Paudyal presents early findings at the in Louisville, Kentucky, June 22-25. A working prototype of DairyBot is expected within six months.

Right-sized technology for all dairies

Although Paudyal said technology and real-time decision-making are the future of dairies, he emphasizes the importance of flexible, right-sized solutions. However, while many farmers see a return on investment, adoption rates vary.

He believes the camera-based systems, which monitor larger groups of cows, can reduce the upfront cost and increase adoption, ultimately helping to minimize the digital divide.

"I always want to develop solutions to the real-world problems that help dairy farmers," Paudyal said. "As a land-grant university with a mission to support Texas dairy farmers, it is essential to develop research projects that provide practical, immediately applicable solutions. By equipping farmers with the tools and resources they need, we can help address real-world challenges on the farm more effectively."