SBIR Phase I: Object Pose Estimation System for Pick and Place Robots

点击翻译按钮获取中文摘要

The broader impact/commercial potential of this project will enable improved cost-efficiency and industrial automation in manufacturing, increasing worker productivity and reducing injuries. The end-users of the robots, i.e., automotive original-equipment-manufacturers and subassembly suppliers, will be able to achieve significant cost advantages by automating new assembly tasks with more inexpensive systems. Of the non-fatal injuries and illness cases reported in the U.S. workforce, 43% of injuries were due to bodily reaction/exertion, and 62% of illness cases were due to repetitive trauma. This innovative solution will facilitate the automation of repetitive, injury-prone manual tasks and greatly improve the speed, accuracy, and cost-efficiency of current robotic handling systems. Beyond handling, there is significant market potential in packaging and warehousing, hazardous materials handling, medical device and other precision manufacturing, and military applications such as bomb defusal and evacuation robots. By enabling new robotic applications and increasing productivity in current automation, this sensor will help the U.S. (and other developed economies) maintain a competitive domestic manufacturing sector.This Small Business Innovation Research (SBIR) Phase I project?s goal is to develop a visual-tactile sensing package for parts handling. The solution is two-fold: (1) A new flexible tactile sensor that can be tailored to a wide variety of form factors; (2) Software to fuse the tactile data with a vision system to estimate pose of objects in pick-and-place tasks. Object grasping and manipulation by robotic hands in unstructured environments demands a sensor that is durable, compliant, and responsive to various force and slip conditions. The goal is to be the first commercially available sensing package that integrates tactile and visual data with accompanying software for state estimation. A large software and gaming company was able to greatly impact the machine vision space by introducing an inexpensive, easily calibrated robust visual sensor; this will do the same for touch sensing ? our team has studied the desirable properties of such tactile sensors for years and discovered a way to produce them in an inexpensive, robust format.