Programmable Sensing Composites

大多数人造物体仍然处于模拟时代：很少有人造物理物体包含传感器，更少的物体在传感器旁边嵌入了计算。这种情况很像世纪之交的模拟摄像机。在物体制造过程和组装结构中嵌入数字计算和传感，可以实现现场计量、真实世界使用期间的现场分析以及来自每个单独物理物体的有价值的统计数据。这种对物体内部物理历史和使用情况的现场处理数字摘要可以从根本上改变我们设计、制造和使用物理物体的方式。对物理物体内部条件的综合数字摘要可以实现与无处不在的数码相机实现的计算机视觉数据驱动方法的革命同等规模或更大的革命。由于它们将使嵌入传感与本地就地信号处理相结合，传感器增强的制造工艺和传感器增强的制造物体可以实现一场革命，对社会(更适合用途的更好的产品)和环境(用更少的废物制造的产品)具有潜在的深远好处。嵌入式传感和计算还可以使材料特性在计算的控制下适应物体的使用模式的未来。可嵌入感知和计算设备以及来自该项目的相关研究成果和方法可以使我们第一次能够在物理对象的生命周期内自主地监测这些现象，基于在整个产品批量中监测的实时数据来执行全新形式的结构完整性分析，并支持新的结构能力和商业产品能力，例如感知应力驱动的自适应召回和数据驱动的产品定制。我们在该项目中的目标是研究将计算和感知嵌入到添加制造的对象中的新的基本方法，并使用现场分析的传感器数据来改进它们的材料配方、设计和制造。在实际使用过程中，来自连续物体内计量的数据可能会使制造高价值物品的全新和潜在颠覆性方法成为可能。今天，在其他工程领域，来自真实世界使用的大量数据正在给计算机视觉和语音识别等任务带来革命性的变化，与之不同的是，制造的对象在很大程度上仍然没有仪表化，数据贫乏，并且错过了数据驱动的用法知情设计、材料配方和制造的机会。

Most man-made objects are still in the analog era: Few man-made physical objects contain sensors and fewer still have any embedded computation alongside sensors. This situation is much like analog cameras at the turn of the century. Embedding digital computation alongside sensing in object fabrication processes and in fabricated structures could enable in situ metrology, in situ analysis during real-world use, and valuable statistics from every individual physical object. Such in-situ-processed digital summaries of the physical histories of internals and usage of objects could enable a fundamental shift in how we design, fabricate, and use physical objects.Synthesized digital summaries of the internal conditions within physical objects could enable a revolution at the same scale or greater than the revolution in data-driven methods for computer vision enabled by ubiquitous digital cameras. Because they will enable embedding sensing combined with local in situ signal processing, sensor-augmented fabrication processes and sensor-augmented fabricated objects could enable a revolution with potential far-reaching benefits to society (better products that are more fit for purpose) and the environment (products manufactured with less waste). Embedded sensing and computation could also enable a future where materials properties adapt, under control of computation, to the modes of usage of objects. Embeddable sense-and-compute devices and the associated research results and methods from this project could allow us, for the first time, to monitor those phenomena autonomously over a physical object's lifetime, to perform fundamentally new forms of structural integrity analysis based on real-time data monitored throughout the volume of products, and to enable new structural capabilities and commercial product capabilities, such as sensed-stress-driven adaptive recalls and data-driven product customization.Our goal in this project is to investigate new fundamental methods for embedding computation and sensing into additively-manufactured objects and to use sensor data, analyzed in situ, to improve their materials formulation, design, and manufacturing. Data from continuous in-object metrology during real-world use could enable fundamentally-new and potentially-disruptive methods for manufacturing high-value items. Today, unlike in other areas of engineering, where large amounts of data from real-world use are revolutionizing tasks such as computer vision and speech recognition, manufactured objects are still largely un-instrumented, data-poor, and missing out on opportunities for data-driven usage-informed design, materials formulation, and manufacturing.