{"id":24877,"date":"2018-07-03T12:07:28","date_gmt":"2018-07-03T16:07:28","guid":{"rendered":"https:\/\/www.tun.com\/blog\/?p=24877"},"modified":"2022-03-16T10:45:20","modified_gmt":"2022-03-16T14:45:20","slug":"why-robots-might-pack-your-next-amazon-order","status":"publish","type":"post","link":"https:\/\/www.tun.com\/blog\/why-robots-might-pack-your-next-amazon-order\/","title":{"rendered":"Why Robots Might Pack Your Next Amazon Order"},"content":{"rendered":"

A team of three roboticists from the Queensland University of Technology (QUT) has <\/span>developed a faster and more accurate way<\/span><\/a> for robots to grasp objects in real-time, opening doors for applications in both industrial and domestic settings. <\/span><\/p>\n

Their<\/span> paper<\/span><\/a> was presented at the Carnegie Mellon University\u2019s international robotics conference,<\/span> Robotics: Science and Systems<\/span><\/a>, last week. <\/span><\/p>\n

Taking baby steps<\/b><\/h2>\n

No one can deny that robotics is the way of the future. According to the <\/span>International Data Corporation<\/span><\/a>, worldwide spending on robotics will reach $135 billion by 2019. <\/span><\/p>\n

While robots have long been replacing human labor in assembly lines and computation and have even started to see, hear and think like humans in chess and debates, robots can\u2019t quite grasp one thing — picking up an object in an unstructured environment. <\/span><\/p>\n

Even for humans, although unnoticed, thousands of sensors in human fingers and complex network of circuitry inside human brains allow a three-month-old to pick up his toy rattle. However, he obviously does not need hours of training or hundreds of researchers to teach him. <\/span><\/p>\n

But for robots, grasping objects in unstructured environments is extremely difficult. Yet, to be useful around humans, robots need to know how to adapt to changing environments. <\/span><\/p>\n

The QUT team has developed a new method for robots to grasp items that move randomly more quickly and accurately. <\/span><\/p>\n

\u201c<\/span>We are interested in creating robots that do useful tasks in the real world. I have been interested for a while in getting robots to robustly interact with the world, pick objects up and place them correctly,\u201d said <\/span>J\u00fcrgen Leitner<\/span><\/a>, a postdoctoral research fellow at <\/span>QUT\u2019s <\/span>Science and Engineering Faculty<\/span><\/a>. \u201cIt turns out some things that are hard for us are simple for computers, such as chess, while some things that come easy for us are very tricky for robots, such as grasping.\u201d<\/span><\/p>\n

The new approach<\/b><\/h2>\n

Deep-learning techniques allow robots to learn from thousands of sizes and shapes to calculate the best way to grasp an unknown item in a static condition. <\/span><\/p>\n

For example, a robot with convolutional neural network (CNN), the most widely used method for object recognition, samples and ranks objects individually and thinks of the best possible grip. However, this technique takes long computation time, even in static environments. <\/span><\/p>\n

The researchers\u2019 method, however, is based on a generative grasping convolutional neural network (GG-CNN), which directly scans its environment and maps each pixel it captures using a depth image, and is fast enough to grasp moving objects. <\/span><\/p>\n

\u201cThe Generative Grasping Convolutional Neural Network approach works by predicting the quality and pose of a two-fingered grasp at every pixel. By mapping what is in front of it using a depth image in a single pass, the robot doesn\u2019t need to sample many different possible grasps before making a decision, avoiding long computing times,\u201d Douglas Morrison, a QUT doctoral researcher said in a <\/span>statement<\/span><\/a>.<\/span><\/p>\n

While their previous robot, which won the Amazon Picking Challenge in 2017, simply looked into a bin of objects, calculated the best possible grip and blindly went in to pick them up, their updated version processes images of the objects within about 20 milliseconds. <\/span><\/p>\n

According to Leitner, after picking up objects, their current robot places them at a fixed location and starts looking for the next item at the scene. <\/span><\/p>\n

The unpredictable world<\/b><\/h2>\n

The researchers tested their new approach in multiple ways. <\/span><\/p>\n

\u201cThere were various experimental setups, all with real world robot and objects,\u201d said Leitner. \u201cWe tested with a range of \u2018household objects,\u2019 3D printed objects, and we verified the approach with different noise levels in the robot controller.\u201d<\/span><\/p>\n

First, they performed 10 trials on grasping unseen objects with adversarial geometry and household objects under both static and moving environment. <\/span><\/p>\n

While the robot achieved grasp success rate of 84 percent for adversarial geometry and 92 percent for household objects under static environment, it achieved 83 percent and 88 percent respectively under moving environment. <\/span><\/p>\n

Second, they mixed 10 test objects in a box and poured it in a cluttered pile below the robot for it to grasp them one by one. They then performed 10 trials on grasping under both static and moving environment. <\/span><\/p>\n

While the robot achieved grasp success rate of 87 percent under static environment, it achieved 81 percent under moving environment. <\/span><\/p>\n

\u201cUsing this new method, we can process images of the objects that a robot views within about 20 milliseconds, which allows the robot to update its decision on where to grasp an object and then do so with much greater purpose. This is particularly important in cluttered spaces,\u201d Letiner said in a statement.<\/span><\/p>\n