Manufacturing Advancements, Manufacturing Technology

Bringing the future forward with advanced manufacturing technology

Advanced manufacturing technology is making headlines left and right. Robotics especially, are commanding a significant amount of attention. While not specific to the manufacturing industry, robotics has revolutionized the way manufacturing operates. With new advances constantly on the horizon, it’s only a matter of time before robotics achieve even greater feats of innovation.

Intelligent, Advanced Manufacturing Technology

Artificial intelligence has always held a certain fascination in the STEM environment. People want to explore the possibilities of technology and determine how far they can push the boundaries. There are two ways most people consider the possibilities of artificial intelligence:

  1. Robots becoming sentient beings
  2. Robots as highly efficient machines with the ability to operate autonomously

The first is more science fiction than science, while the second is the more realistic of the two possibilities. Dr. Herbert Simon, a behavioral scientist who helped pioneer artificial intelligence, stated that, “Machines will be capable in 20 years of doing any work a man can do.” That was in 1965. It’s been over 50 years since Dr. Simon said those famous words, yet technology is still not to the point where it can do all of the things a human is capable of.

Finger dexterity, the ability to make emotional judgements and other key factors mean robotics still trail humans in their capacity to complete tasks.

The Mysteries of Artificial Intelligence

They key to unlocking the mysteries of artificial intelligence may lie in machine learning. Advanced manufacturing technology is heavily focused on machine learning. Everything from supply chain operation and product customization, to learning purchasing algorithms falls under the umbrella of machine learning.

A breakthrough in robotics came when Stefanie Tellex, a computer science professor at Brown University, collaborated with Ashutosh Saxena of Cornell University. Together they demonstrated that robots could learn from the experiences of other robots. Saxena taught her robot to lift and arrange a series of cups in a lab in California. Tellex then downloaded the data and used it to train her robot the same actions while located on the east coast. What was significant about the experiment was that both robots were physically different and learned to complete the same actions in completely separate environments.

Machine learning is far from reaching the level of artificial intelligence, but steps are being taken in that direction. Manufacturing technology is already benefiting from a variety of machine learning capabilities, including the ability to extract relationships within large data-sets in complex environments. Machine learning also helps manufacturers to reduce their cycle time while improving the efficient use of resources.

Addressing Challenges for Manufacturing Technology

When something as complicated as robotics and artificial intelligence is on the line there’s bound to be some challenges.

Common problems associated with robotics:

  • Can be too clumsy to handle certain objects
  • Can have an inability to distinguish between grip variations
  • Vision-based sensing accuracy
  • Optimal gait (walking ability)
  • SLAM (Simultaneous Localization and Mapping)
  • Metacognition

One of the biggest obstacles to the widespread adoption of robotics in manufacturing is the missing element of finesse, especially as it concerns the ability to grip and pick-up objects. No robot can handle objects quite like the human hand. Our ability to judge the size, weight, texture, and fragility of an item is computed instantaneously.

Recent research suggests that robotics is progressing from learning how to manually grip items, to being able to compute the optimal grip based on data derived from 3-D shapes, visual appearance, and the physics of grasping each item. Improvements like these have definitely earned the right to be considered advanced manufacturing technologies.