Visual Robots - Peek into the Future of Material Handling

Visual Robots - Peek into the Future of Material Handling

 
By Bill Lydon - Contributing Editor
 
Autonomous Robots...
At ProMat 2009, Seegrid Corporation provided a peek into the future capabilities that advanced technology will deliver to material handling with its GP8 robotic pallet truck. The product is easy to use, simply walk it through where it needs to go and it learns the path by taking pictures and building 3D map of its surroundings to create guide paths.
 
 
...learns the path by taking pictures and building 3D map of its surroundings...
 
Seegrid robots use no lasers, tape or wires and Seegrid terms the learning process, WalkThroughThenWork™, set up is one walk through your facility and the machines are ready to work in minutes.   Sort of a visual GPS, except that it is learning in the way a person remembers how to get places - visually.    Software translates physical information into data and builds 3D maps and creates guide paths from the information.
 
Seegrid robots use no lasers, tape or wires!
What makes Seegrid’s robots different from today’s AGVs (Automated Guided Vehicles) is the company’s IMR technology.  Seegrid is the first to provide early-stage capabilities of an autonomous robot behavior with Sense, Move, Analyze, Interact and Repeat capabilities. Current AGVs are at the peak of their technology capacity, but Seegrid’s IMR technology is at the beginning - providing AGV-like abilities but with greater flexibility and at a considerably lower cost. Flexibility at a lower cost is critical to companies who want the ability to do flexible manufacturing, changing work processes on the fly and working in real time to adjust to work volume.   The picture shows a computer image of a model built with the onboard cameras.
 

 
Carnegie Mellon University’s Mobile Robot Lab Roots
Seegrid Corporation was founded in Pittsburgh, PA in February 2003 as a spinout of Carnegie Mellon University’s Mobile Robot Lab.  Dr. Hans Moravec and Dr. Scott Friedman co-founded Seegrid to develop a new class of affordable industrial mobile robots that operate reliably in dynamic environments.
 
Seegrid’s core Industrial Mobile Robotics (IMR) technology has over 30 years of innovation and research behind it and overcomes the traditional hurdles of robotic adoption.
"The lift truck industry grew out of the need for warehouse workers to have durable equipment to do their work effectively. Industrial mobile robotics means that the same durable equipment now can do some of the work for the warehouse workers," said Scott Friedman, CEO of Seegrid. "It allows warehouse operators to maintain their hard learned and hard-earned best practices, while giving them a new tool to address the challenges they face today: Labor shortages, ergonomic and safety issues, and cost reduction."   Seegrid CEO, Scott Friedman
Dr. Hans Moravec, Seegrid’s Chief Scientist, invented and pioneered the use of 3D evidence grid technology for machine navigation and perception.  Evidence Grid Technology is a software mechanism that breaks the world down into a probabilistically weighted grid. Evidence grid technology, properly applied, allows robots to sense and interpret their environment.
 
From inception Seegrid has been solely focused on creating affordable, dependable robots for material handling applications. Seegrid’s team brings together many disciplines including physics, mathematics, computer programming, electrical and mechanical engineering to create a reliable, cost effective IMR solutions.  The minds at Seegrid have been at work on robotics technologies since the late 1970s, and a series of breakthroughs and evolutionary developments have worked together to allow Seegrid to deliver the future of industrial mobile robotics today.
 
Technological Flashpoint
Increased robot capabilities can be linked directly to the evolution of computer power. Software which has been developed over the last 30 plus years had to wait until computer size, cost and power caught up to make industrial mobile robotics commercially and technically feasible. IMR technology is not heuristic but probabilistic which increase in capability as CPU performance and storage increases.
 
Tip of the Iceberg
Seegrid has a vision…in order for robots to be truly effective they need to become autonomous or self sufficient and “learn” the required tasks directly from the people that work with them each day.
 
Industrial Mobile Robotics (IMR) technology employs sophisticated, state-of-the-art artificial intelligence, Bayesian statistics, image processing and machine learning methods to enable vehicles to automate core material handling processes. IMR technology acts as a platform for truly autonomous behavior, allowing material handling vehicles to work in environments that were previously not economically or technically feasible.
 
Seegrid’s core IMR technology relates to the development and exploitation of 3D evidence grids for machine perception. In an evidence grid, many views of a scene are taken (with stereo cameras or other ranging sensors), and the data from each individual snapshot are accumulated into one 3D model of the scene. A 3D evidence grid breaks the world up into individual cubes called voxels, and then statistically weights those cubes with certain kinds of evidence (information about the world).
 
These grids allow Seegrid’s technology to deal with environmental uncertainty better than the current generation of commercial automation offerings. Where previous robotics and machine vision industrial solutions needed tight precision and close proximity in the environment, Seegrid’s solutions deal with variability and uncertainty in both the environment as well as in the work to be completed, which is more realistic and true to life in the “Murphy’s Law” world of material handling.
 
 
Seegrid industrial mobile robots sense, move, analyze, interact…and repeat to automate your everyday material handling tasks.
 
Software Centric Approach
Seegrid’s approach to solving challenging robotic tasks is to rely on intelligent software rather than expensive hardware. This fresh approach uses inexpensive sensors, draws solid conclusions from incomplete or bad data, and grows and improves robot capabilities with the increasing computational power of commercially available chipsets. The end result is a flexible, affordable and reliable robot that functions well in material handling environments.
 
Real World
Seegrid’s IMR-enabled robots automate material handling work processes without any changes to existing facilities, enabling the automation to be incrementally applied, and for all other ongoing manual work processes to continue. This automation works even as the facility is rearranged, and as work processes are added, deleted or changed. The company’s IMR-enabled robots manage a set of internal, statistical, detailed 3D maps of the facility and work queues of requested and future jobs. At the same time they provide real-time management interfaces, and maintain activity and audit logs which capture usage information.
 
This is going to be very interesting to watch…