STTR Phase I: Using Audio Analytics and Sensing to Enhance Broiler Chicken Welfare and Performance by Continuously Monitoring Bird Vocalizations

这一小型企业技术转让第一阶段项目的更广泛影响将是提高家禽养殖场鸡的福利，并为养殖者配备有效的工具来监测禽类状况。由于鸡肉是全球广泛消费的活体动物蛋白来源，消费者对合乎道德的饲养动物的偏好越来越高。该项目通过促进改善家禽养殖场的福利做法来满足这一需求。有与主要生产者和机构消费者合作的运动，以建立影响整个供应链的循证福利标准。随着美国农业劳动力的减少，拥有自动化机制来扩展农民的能力是至关重要的。该项目将为满足这些需求的鸟类开发一个智能监测系统，从而改善鸟类的福利并扩大农民的能力。这个小型企业技术转移(STTR)第一阶段项目使用音频监控和机器倾听来测量动物的行为。由于不同农场的家禽操作差异很大，而且在鸡从雏鸟成长为成熟鸟的整个生命周期中，机器学习算法必须适应。监测系统必须像家用电器一样，因为它们不需要专业知识，也不超过农民的最低限度参与。这项研究将导致声学机器学习算法的进步和产品化，这些算法可以搜索出动物在其环境中的异常行为，并基于智能监听和推理为种植者提供痛苦、疾病、不适以及饲料和水问题的早期迹象。声学方法不会干扰动物，比视频更强大，可以在尘土飞扬的环境中长期部署，并且全天候在黑暗中运行。通过向饲养者提供早期可操作的见解，这项技术可以及早纠正问题，从而不仅改善动物的福利，而且提高它们的生产力。通过在成长中的房屋中的多个地点部署廉价的麦克风，可以将活动和问题本地化，将精确的牲畜技术带到基于羊群的动物管理中。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact of this Small Business Technology Transfer (STTR) Phase I project will be in enhancing the well-being of chickens on poultry farms and in equipping growers with effective tools to monitor bird conditions. As chicken is a widely consumed source of live-animal protein globally, there is a growing consumer preference for ethically raised animals. The project addresses this demand by fostering improved welfare practices in poultry farming. There are collaborative movements with major producers and institutional consumers to establish evidence-based welfare standards impacting entire supply chain. With a declining agricultural workforce in the United States, it is essential to have automated mechanisms to extend a farmer’s capabilities. This project will develop a smart monitoring system for the birds meeting these needs, resulting in improved bird welfare and amplification of the farmer’s capacity. This Small Business Technology Transfer (STTR) Phase I project uses audio monitoring and machine listening to measure animal behavior. Since poultry operations differ significantly from farm to farm and over the life of the chicken as it grows from chick to a mature bird, the machine learning algorithms must adapt. The monitoring systems must be appliance-like in that they do not require expertise or any more than minimal involvement on the part of the farmer. This research will result in the advancement and productization of acoustic machine learning algorithms which search out unusual behaviors in the animals in their environment and provide early indications of distress, sickness, discomfort, and feed and water issues to the grower based on intelligent listening and inference. Acoustic approaches do not disturb the animals, are more robust than video for long-term deployment in dusty environments, and operate around the clock and in the dark. By providing early actionable insights to the grower, this technology can correct problems early, thereby improving not only the animal’s welfare, but their productivity as well. By deploying inexpensive microphones at multiple locations in a grow-out house, activities and problems can be localized, bringing precision livestock technology to flock-based animal management.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.