The swine industry faces a significant challenge in the form of preweaning mortality (PWM), which causes substantial economic losses and animal welfare concerns

The Department of Agriculture National Institute of Food and Agriculture awarded Texas A&M Agrilife Research $603,169 on March 15, 2026, under the Agriculture and Food Research Initiative (AFRI) program to develop a cyber-physical system for reducing preweaning mortality in swine production.

The project, running through March 14, 2029, and performed at College Station, Texas, combines real-time image processing and machine learning algorithms to predict farrowing time, detect farrowing difficulties, and identify stillborn piglets. The system employs a novel image processing technique to evaluate sow physiological indicators for farrowing onset and difficulty, enabling farmers to select sows for farrowing induction and receive alerts for prolonged birth intervals. A complementary automated anti-crushing device identifies danger zones where piglets risk crushing by the sow and relocates piglets away from those regions. The human-in-the-loop design targets 70 percent reduction in stillborn-related preweaning mortality; the closed-loop anti-crushing mechanism targets 70 percent reduction in crushing-related preweaning mortality. Researchers at University of Missouri-Columbia collaborate on the project.

Stillborn and crushing together account for 70% of the total PWM. Researchers at texas a & m university, in collaboration with researchers from university of missouri-columbia, propose a research project towards a novel cyber-physical system to substantially reduce PWM and revolutionize swine production practices. This innovative system employs real-time image processing and machine learning algorithms to predict farrowing time, detect farrowing difficulties, identify stillborn piglets, and incorporates a cost-effective robotic system to minimize piglets' risk of being crushed by the SOW. A novel image processing technic will be developed to evaluate SOW's physiological indicators for upcoming onset of farrowing and presence of farrowing difficulty. This would enable farmers to make rational selection of the sows for farrowing induction in order to minimize the number of unattended farrowing events. Moreover, farmers will be alerted to provide timely treatment to the sows when prolonged birth interval was detected. This human-in-the loop design may help reduce stillborn-related PWM by 70%. Furthermore, the system will identify the danger area where piglets can be crushed by the SOW. An automated anti-crushing device will proactively prevent crushing by relocating the piglets away from the identified danger region. This closed loop configuration may reduce crushing-related PWM by 70%. This comprehensive solution is expected to save millions of piglets annually and generate substantial economic benefits for swine producers.