Tricorder Array Technologies Animal Welfare Sensor

项目摘要 Tricorder阵列技术有限责任公司（Tricorder）将与密歇根大学合作开发 亚拉巴马，伯明翰（UAB），SIDECARe™（三录仪阵列技术动物福利 传感器）。SIDECARe™旨在改善大规模动物实验中笼子级别的研究动物护理。 通过自动收集相关数据来提高一致性， 动物模型的可重复性。SIDECARe™将彻底改变监测健康和良好- 作为第一个对小鼠超声发声（mUSV）进行分类的小鼠。在这 第二阶段应用，目标1完成连续mUSV监测的集成 SIDECARe™设备中。目标2扩展SIDECARe™人工智能（AI） 使分类能力能够覆盖研究中断的主要原因。目标3 通过在UAB动物园工作流程中进行测试部署，优化SIDECARe™系统。最后， aim 4在外部测试点部署beta系统，进行售前试验。Tricorder的创新应用 的超声波传感器允许传感器阵列拾取鼠标发声，并使用它们来 解释笼子里的健康和社会事件。 SIDECARe™采用分布式方法部署所需的AI算法， 有效分割和解释mUSV。当前人工智能算法的重点是 SIDECARe™是检测条件，如小鼠幼崽在笼子里，淹没的笼子，战斗 动物和溃疡性皮炎。不断增加的mUSV数据采集 然而，笼子的数量代表了未来可利用的大量未开发数据。 新算法的开发和对小鼠行为和健康的新理解。的 SIDECARe™系统的一个特别重要的功能是集成连续的机器学习 算法，它不断发展，以更好地确定其他因素和趋势， 发病率和死亡率的上升。记录的特定环境的存在 条件（笼照明水平、温度变化、湿度、环境噪声水平） 与小鼠行为相关的研究结果和/或 再现性

PROJECT SUMMARY Tricorder Array Technologies, LLC (Tricorder), will develop, in partnership with the University of Alabama at Birmingham (UAB), SIDECARe™ (Tricorder Array Technologies Animal Welfare Sensor). SIDECARe™ is set to improve research animal care at the cage level on a massive cost-effective scale by automating the collection of relevant data to enhance consistency and reproducibility of animal models. SIDECARe™ will revolutionize monitoring the health and well- being of mice by being the first to categorize mouse ultrasonic vocalizations (mUSV). In this Phase-II application, aim 1 completes the integration of the continuous mUSV monitoring solution into SIDECARe™ devices. Aim 2 expands SIDECARe™ Artificial Intelligence (AI) enabled classification capabilities to cover the top causes of research disruption. Aim 3 optimizes the SIDECARe™ system via beta deployments in the UAB vivarium workflow. Finally, aim 4 deploys beta systems at external test sites for pre-sales trials. Tricorder’s innovative use of ultrasonic sensors allows the sensor array to pick up mouse vocalizations and use them to interpret health and social events in the cage. SIDECARe™ utilizes a distributed approach to deployment of the AI algorithms required for effective segmentation and interpretation of mUSVs. The focus of the current AI algorithms for SIDECARe™ is to detect conditions such as mouse pups in a cage, flooded cages, fighting animals, and ulcerative dermatitis. The continuous acquisition of mUSV data across increasing numbers of cages, however, represents a wealth of unexplored data available for future development of novel algorithms and new understandings of mouse behavior and well-being. Of particular importance to the SIDECARe™ system is to integrate continuous machine-learning algorithms, which constantly evolve to better identify additional factors and trends that contribute to increased rates of morbidity and mortality. The recorded presence of specific environmental conditions (cage lighting levels, temperature variations, humidity, ambient noise levels) correlated with mouse behaviors are useful for studying research outcomes and/or reproducibility.