SBIR Phase I: Gripper-integrated proximity, contact and force sensing for collaborative robots

这个项目更广泛的影响/商业潜力将极大地推动机器人的应用，提高机器人与人类一起工作时的安全性。将嵌入式接近传感器与先进的用户界面控制系统集成到机器人末端执行器中，为当前可用的传感和控制系统提供了重大飞跃。这个项目在传感和控制方面的进展将扩大目前限制人类与机器人力量互动的界限。协作机器人在控制和传感方面的改进将推动机器人在工业、医疗和家庭应用中的应用。精度的提高将扩大机器人的使用范围，并显著降低许多先进行业的制造成本。这个项目大大提高了机器人的安全性？S系统将允许扩大人与人之间的互动，包括机器人护理，以满足日益增长的老年人口的需求。这个小企业创新研究（SBIR）第一阶段项目将促进我们对触觉在稳健操作中所起作用的理解，并为夹持器内3D传感带来新的手段，补充并在某些情况下取代外部3D感知系统。具体来说，本项目将展示接近感测和零力接触感测在抓取规划、目标配准和目标识别方面的作用。一个关键的挑战将是合并来自两种不同传感器技术的信息，这两种传感器技术在不同的范围范围内都表现出色，但在其他方面精度较差。提出的基于协方差的传感器融合算法将通过与高保真激光和3D扫描系统获得的3D模型进行比较来进行基准测试。第二个挑战将是将产生的能力抽象到易于使用的编程环境中，用于机器人的抓取和操作规划，这将在现场与客户进行测试。

The broader impact/commercial potential of this project will be to greatly advance the application of robots and improve robots safety when working together with humans. Integration of embedded proximity sensors into robotic end-effectors with an advanced user-interface control system provides a significant leap from current sensory and control systems available. The advances of this project in sensing and control will expand the current boundaries which limit human interaction with the power of robots. The improvements to both control and sensing in collaborative robots will advance the use of robots in industrial, medical and household applications. Improved accuracy will expand robot use and significantly lower manufacturing costs in many advanced industries. Vast increases in robot safety with this project?s system will allow expanded human interaction including robot care for the needs of the growing senior population.This Small Business Innovation Research (SBIR) Phase I project will advance our understanding of the role tactile sensation plays in robust manipulation and lead to novel means for in-gripper 3D sensing, complementing and in some cases replacing external 3D perception systems. Specifically, this project will demonstrate the role that proximity sensing and zero-force contact sensing play in grasp planning, object registration, and object recognition. A key challenge will be to merge information from two different sensor technologies that both excel in different range regimes, but have poor accuracy otherwise. The proposed covariance-based sensor fusion algorithms will be benchmarked by comparison with 3D models obtained from high-fidelity laser and 3D-scanning systems. A second challenge will be to abstract the resulting capability into easy-to-use programming environments for robotic grasp and manipulation planning, which will be tested with customers in the field.