SBIR Phase II: Building a Nature Monitoring Network for Birds

这个小企业创新研究第二阶段项目的更广泛的影响是进一步开发一种产品，使消费者能够倾听和了解他们后院的鸟类。 该产品将提高公众的科学素养和参与，通过有趣和非威胁的方式介绍有关鸟类的信息。 此外，它将使环境研究更容易理解，有形的，更有价值的人，包括妇女和个人从代表性不足的群体这个项目旨在帮助休闲观鸟者更多地了解自己的后院生态系统，提高他们的能力，以识别鸟类的歌曲，并认识到生活的多样性在他们的窗外。客户将通过成为以鸟类为中心的研究的重要合作伙伴，并与分享他们兴趣的其他人建立联系，了解环境管理的重要性。2020年，参与鸟类计数挑战的人数大幅增加，并在2021年继续创下纪录。主要的技术挑战将是在低成本的边缘硬件上设计和运行鸟鸣探测器，同时忽略非鸟类的声音，如狗、昆虫、警报器和机器。 将开发边缘算法和原型硬件，以尽可能低的成本提供最准确的鸟鸣识别，这将导致广泛的商业化。 这是一个重大的技术挑战，因为该产品的神经网络远远大于人工智能芯片目前所能支持的。 为目标市场生产定价的产品将需要创建一个全面的鸟类和环境声音标记库，一个训练有素的神经网络，优化的硬件，一个全面的beta测试程序，以及一个信息丰富的，易于使用的网站和智能手机应用程序。该奖项反映了NSF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进行评估，被认为值得支持审查标准。

The broader impact of this Small Business Innovation Research Phase II project is to further develop a product which allows consumers to listen and learn about the birds in their backyards. The product will improve the general public’s scientific literacy and engagement by presenting information about birds in fun and non-threatening ways. Additionally, it will make environmental research more understandable, tangible, and valuable to more people and include women and individuals from underrepresented groups This project seeks to help casual birdwatchers know more about their own backyard ecosystems, improve their ability to identify birds by their songs, and recognize the diversity of life outside their windows. Customers will learn the importance of environmental stewardship by becoming important partners in bird-centered research and connecting with others who share their interests. Participation in bird count challenges saw large increases in 2020 and record-setting participation continued into 2021. The primary technical challenge will be to design and run birdsong detectors on low-cost edge hardware, while ignoring non-bird sounds such as dogs, insects, sirens, and machines. Edge algorithms and prototype hardware will be developed to provide the most accurate birdsong identifications at the lowest possible cost, which will result in wide commercialization. This represents a significant technical challenge because the product’s neural net is much larger than artificial intelligence chips can currently support. Producing a product priced for the target market will require the creation of a comprehensive labeled library of bird and environmental sounds, a well-trained neural net, optimized hardware, a comprehensive beta testing program, and an informative, easy to use website and smartphone app.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.