With costs coming down for sensor-based technologies,
vision and tactile feedback robotic systems are becoming increasingly more
intelligent, people-aware, and are converging on a new paradigm and set of
robotic applications.
This new direction is focused around three advances: proficiency
without extensive software coding at the outset of an implementation, the
ability for robots to work side-by-side with humans, and concepts like cloud
computing that have the potential to help create more sophisticated learning
machines.
Some developments are still futuristic, but already more and more
robots are finding their way into warehousing, order fulfillment, health care
applications, and may find a place in the future as "autonomous service robots"
in automotive assembly plants.
Learning Trend
"We conducted a survey within the packaging industry, and
the results came back very strongly that users of robotic automation are just
tired of programming," says Rush LaSelle, vice president of sales and marketing
for Adept Technology Inc. "It's probably the number one dissatisfaction in
implementing robotics. Users want more intuitive, easy-to-deploy technology."
LaSelle says that this trend, along with reduced costs and higher
functionality with sensory perception-based technologies, is a very strong
driver in the cloud computing thought process. The idea is to create learning
machines that can essentially take algorithms and become more adept and
proficient without a lot of extensive coding at the outset of an
implementation.
"The value proposition for robots moving forward is the ability
to search, validate and truly become the last mile for intelligent decision
making," says LaSelle. "In the same way
a surgeon's assistant is tasked with giving precisely the right instrument when
the doctor needs it, the learning element is all about optimizing the process
to predict what a surgeon might need next."
He says that, after the incision, for example, the surgeon may
need a clamp - and we all know that computers are much better and accurate at
processing both that type of information and large amounts of data. The ability
to learn, predict and accurately provide a surgeon with the right instrument is
very much data-driven. "So the last mile is taking what's known, anticipated
and expected, and converting it into accurate physical activity," LaSelle says.
Service Robots for Automotive Assembly
General Motors, the largest user of robots in the world,
with a fleet of 25,000 robots used in factories around the world, is also
stepping up its robotic investigations with a specific long-term target:
assembly operations where future robotic solutions will assist people in
building better and safer cars.
"We have been observing robotic advancements in Japan and
European countries for years, and believe a new field and era in robotics is
emerging for what we call service robots," says Dr. Roland Menassa, GM's
advanced robotics manager in the Manufacturing Systems Research Lab.
The focus at GM is on developing humanoid, two-arm-type robots,
which is behind a current five-year collaborative agreement with NASA's Johnson
Space Center that began in 2007. In April, NASA announced it will launch the
first human-like robot into space later this year. It will become a permanent
resident of the International Space Station. Robonaut 2, or R2, was developed
jointly by NASA and General Motors under their cooperative agreement to develop
a robotic assistant that can work alongside humans, whether they are astronauts
in space or workers at GM manufacturing plants on Earth.
The 300-lb R2 consists of a head and a torso with two arms and
two hands. R2 will launch on Space Shuttle Discovery as part of the STS-133
mission planned for September.
"We don't have a finished product at this point, and I don't want
people to have the misconception that GM will put humanoid robots into the
plants tomorrow. That is not the case," says Menassa. "The technology will
continue to evolve but along this journey we stand to benefit from the spin-off
technologies that have been developed for eventual use in other facets of the
manufacturing process."
From the GM perspective, walking robots are not attractive in a
manufacturing setting, so the focus has been on the upper torso, the dexterity
of the hands and degrees of freedom to do real work. There has to be a level of
intelligence so that you can interact with the robot in a human, intuitive
way. And the development needs to drive
a new set of safety technologies which GM intends to deploy in other
traditional areas.
"Our position is that the technological advancements achieved
with the collaboration between GM and NASA will enable us to offer customers
safer vehicles and build them in a safer way," Menassa says. "If you look at
the control and sensor advancements that we are developing with this robot, it
can help us drive the development of future vehicles and safety systems, as
well."
Over the very long-term future, measured in decades, the goal is
to have a robot that can use the same tools that humans can. And unlike
traditional robotics where there's a need to spend 3x to 10x, depending on the
application, to enable the robot in the process, this new technology does not
need traditional support equipment such as fencing, light screens and safety
mats.
"From an assembly perspective, we think these robots can enable a
flexibility solution that we cannot provide today," says Menassa. "And at the
end of the day, GM uses robotics or any automated solution to enable the
process. Process is what drives everything."
Autonomous Robotic Workcells for Surgical Instruments
Robotic Systems & Technology is bringing contemporary
automation and robotics into an area of the hospital system called sterile
supply with the development of autonomous, robotic workcells to manipulate,
track and process surgical instruments and supplies.
RST's latest product, PenelopeCS, is intended to revolutionize
the way hospitals manage inventory of surgical instruments. PenelopeCS
automates key functions in the hospital's sterile supply department where
used/dirty surgical instruments are cleaned, sorted, inspected and repacked
into containers to be sterilized for the next procedure.
"We are introducing a very smart robot - a real robot with eyes
and a brain into the workplace," says Dr. Michael Treat, a surgeon and founder
and leader of Robotic Systems & Technology Inc. "This represents a new
trend in robotics and a new approach for medical applications. There are robots
in the operating room but those are teleoperated devices and basically serve as
sophisticated power steering for the surgeon."
Treat says the expertise developed over the last six years has
centered on designing machine vision routines for surgical instruments and the
intelligence built into its software to recognize and organize the instruments.
But the challenge is both technical and organizational, configuring the machine
and software to fit nicely into the existing process. "It's an interactive
manufacturing process, and really a new kind of robotics in the sense that it's
more involved with people than assembly line robots," says Treat. "PenelopeCS
has a brain which is focused and limited, but there is independence and
independent activity that the software in the rule base generates. The goal is
to advance both robotics and hospital efficiency using a man-machine
partnership."
Intelligent, Mobile Robotic Control
From individual robots in hospitals to fleets of AGVs in
factories and warehouses, automation solution providers are increasingly
relying on autonomous mobile robots to handle navigation, positioning, tasking
and configuration.
Mobile Robots Inc. has been manufacturing automated guided
vehicles for 15 years, and has thousands of robots installed around the world.
The company is a tool and component supplier, not a maker of end-user solutions
and they supply design engineers with a base autonomous robot.
"Our robots are designed to do the most sophisticated, autonomous
behaviors, navigation and localization without having a roboticist or
programmer involved," says Jeanne Dietsch, CEO for Mobile Robots Inc. "They are
designed for engineers to add onto and integrate into warehouse management
systems, external scheduling and inventory control systems."
The company's Motivity technology is centralized around
autonomous navigation and localization but it is more than that. It uses
natural, feature-based autonomous navigation and localization which eliminate
the need to install lines or beacons used with other AGV solutions. Software
tools for setup and installation of the robot, along with networking and
interfacing tools plus a real-time control GUI, reportedly make it easy for the
installer to use ready-made robot behaviors and responses. "An engineer can
create a system by dragging and dropping objects that would take months for a
programmer to create," says Dietsch.
MobileEyes software tools provides a way to set up, implement and
monitor robots, and serves as a monitoring station or command and control
station for users, if desired. Mobile Planner enables customization of a
project, so application designers can create missions, tasks, behaviors and
personalities appropriate for their specific workplace, whether it's an industrial plant or a hospital patient's room.
The robots learn to recognize their location in a facility based
on natural features; for example, learning the facility layout by looking at
walls, parts of the factory or the ceiling in a dynamic space like a warehouse.
Because they don't require beacons or lines in the floor, it reduces the
upfront cost of the system. With older AGV systems, before you put in the first
AGV, there is normally a huge infrastructure investment and it isn't something
an engineer can try out in a facility beforehand to see how well it works.
Dietsch says that one example of where Motivity systems have
proven very useful is an intractable problem for tire factories. In the tire
manufacturing process, a manual process is typically used to carry tires from
the gantry where they are stored and cured for about two weeks to the press
where they are cooked. This process of carrying the tires had been done
manually using carts because of the distance from the gantries to the presses.
Though there are many forklift trucks and people moving within the facility,
robots deployed to handle this task have been able to deal with the traffic and
have driven millions of miles per year safely.
In the largest installation for Motivity, there are 38 robots
running simultaneously, and the system made a 25-percent throughput improvement
over the manual system. The robots are now waiting for the presses to clear,
rather than vice versa, and five global tire manufacturers are deploying these
systems.
The advantage of Motivity is that it enables flexible automation,
and users no longer have a robot that can only go from point A to point B.
Instead, the robot can move from A1, A2, A3, A4, A5 and A6 to B1, B2, B3 and B4
in any sequence. Locations can be randomized, unknown or dynamically guided.
"Instead of going onto a cart where it is distributed by someone to various locations where it is actually being
used, the material goes onto a robot, and the robot can take it the last mile.
The first mile and the last mile are always the most difficult in any delivery
situation, and our robot is excellent in those applications."