I-Corps: Automated Audio Monitoring

I-Corps项目的更广泛的影响/商业潜力在于能够利用声音对环境进行实际的长期监测，例如在畜牧业生产、工业环境、老年人护理和监视/军事应用中。虽然消费者声控数字助理已经推动了消费电子产品的转型，但这项工作有可能通过显著减少分析新环境所需的人力和计算资源，推动非语音听觉数据的平行转型。在农业中，这种类型的声学监测可以更好地了解动物的健康状况，更一致和人道地对待动物，并获得更好的总体结果。对于工业和监视应用，它可以促进对问题或威胁的预防和快速响应。这个I-Corps项目基于一种技术，该技术可以了解环境的正常情况，然后在潜在的异常事件或情况发生时突出显示。如果需要，可以标记这些异常条件，使系统能够区分不同的已知条件。在过去六年中，在监测家禽生产设施的范围内，已经开发和广泛测试了这项技术。在从几只鸟到数万只鸟的许多环境中，它已被证明可以有效地突出异常情况，如疾病、过热、人类入侵、照明变化和异常设备噪音。该技术使用专有的基于稀疏编码的算法，其计算效率足以在边缘硬件上运行。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值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 / commercial potential of this I-Corps project lies in enabling practical, long-term monitoring of environments, using sound, such as in livestock production, industrial settings, elder care, and surveillance/military applications. While consumer voice-activated digital assistants have driven transformation in consumer electronics, this work has potential to drive a parallel transformation for non-voice auditory data by significantly reducing the human effort and computational resources required to analyze new environments. In agriculture, this type of acoustic monitoring could result in a better understanding of animal well-being, more consistent and humane treatment of animals, and better overall outcomes. For industrial and surveillance applications, it could facilitate prevention of and rapid response to problems or threats. This I-Corps project is based on technology that learns what normal conditions sound like for an environment, then highlights when potentially abnormal events or conditions occur. If desired, these abnormal conditions can be tagged to enable the system to distinguish between different known conditions. This technology has been developed and tested extensively over the past six years within the context of monitoring poultry production facilities. In many environments ranging from a few birds to tens of thousands of birds, it has proven effective in highlighting anomalous conditions such as the presence of disease, excessive heat, human intrusions, lighting changes, and abnormal equipment noise. The technology uses proprietary sparse coding based algorithms that are computationally efficient enough to run on edge hardware.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.