Just like the cortex governs human capabilities, the capabilities of robots depend on Artificial Intelligence (AI). With path-breaking work in the field of AI, advanced robots aren’t a distant future, they are becoming a part of our present with companies like NVIDIA (NVDA) speeding up the process towards their adoption. Here’s a look at what NVIDIA is doing these days.
Right now, industrial robots are the most popular form in use. The International Federation of Robotics (IFR) has reported an increase of 31% in global sales of industrial robots in 2017. Some of the reasons given for the adoption of robots are: improvement in efficiency, cost effectiveness, and flexibility to change. The sale of industrial robots reached a record of 387,000 units in 2017 from 294,300 in 2016. During 2017, China installed around 138,000 industrial robots, while Japan was at 46,000 units while the U.S. recorded a sale of 33,000 industrial robots.
It is estimated that more than 3 million industrial robots will be in use in factories around the world by 2020. In a survey conducted by automatica, 70% of the respondents believe that human-robot teams will improve manufacturing by combining human talents with the strengths of robotics. The industrial robotics market is expected to be worth $71.72 billion by 2023.
While industrial robots are becoming popular, the other category of robots, service robots, which will be used in homes, hospitals, or logistics, are lagging behind. While industrial robots which can be easily trained to repeat a particular task, the adoption of service robots requires them to learn to respond appropriately each time to a particular situation. NVIDIA has some solutions for that.
Robots have begun to do more than just perform simple tasks. This is where machine learning and deep learning—which lie underneath the crust of AI—come to play. In 1959, Arthur Samuel defined machine learning as a “field of study that gives computers the ability to learn without being explicitly programmed” while deep learning is the “science of training large artificial neural networks.”
NVIDIA defines the relation between AI, machine learning and deep learning like, “The easiest way to think of their relationship is to visualize them as concentric circles with AI—the idea that came first—the largest, then machine learning—which blossomed later, and finally deep learning—which is driving today’s AI explosion—fitting inside both.”
NVIDIA’s Push to Robotics. NVIDIA’s Jetson has been the complete platform for developers and researchers to realize AI applications on various Unmanned Ariel Vehicles (UAV), Unmanned Ground Vehicles (UGV) and other kind of robotic hardware. NVIDIA Jetson uses the power of deep learning to drive this exciting new era of smart embedded robotics.
Over time, NVIDIA has come up with advanced versions of Jetson, such as Jetson TX1 and Jetson TX2. Launched in June 2018, NVIDIA Jetson Xavier is the latest and most advanced version of the series. Jetson Xavier is ten times more energy efficient while its performance is twenty times better of its previous version. With more than 9 billion transistors, Jetson Xavier delivers over 30 TOPS (trillion operations per second) — more processing capability than a powerful workstation in an embedded module under 30W. Jetson Xavier has six kinds of processors which support dozens of algorithms simultaneously in real-time for sensor processing, odometry, localization and mapping, vision and perception, and path planning.
Jetson Xavier is the heart of NVIDIA Isaac—"a platform to power the next generation of autonomous machines, bringing artificial intelligence capabilities to robots.” According to NVIDIA founder and CEO Jensen Huang, “NVIDIA Isaac includes new hardware, software and a robot simulator for training AI-powered robots in a virtual-world.”
NVIDIA is making robots ready for the real world. Training robots in real environment has safety concerns in addition to being labor and cost intensive. These concerns are put to rest with training in a simulated environment. However, inadequacy of simulators to capture reality ‘perfectly’ creates hiccups in transferring simulated experience in the real world.
NVIDIA’s Isaac is a big advancement since it can help bridge the “reality gap.” NVIDIA Isaac enables the creation of accurate real scenarios with utmost details which accelerates the training process while reducing times spent on training and testing, all at reduced costs.
NVIDIA’s has been actively working to push advancements in robotics. Back in 2016, NVIDIA and FANUC Corporation started working together to increase robotics productivity and bring enhanced capabilities to automated factories worldwide. In 2017, Avitas Systems, a GE Venture, partnered with NVIDIA to optimize the use of robotics for inspection and better defects detection on industrial assets.
NVIDIA researchers have recently developed a deep learning-based system wherein a robot learns by observing human actions. The work will go a long way in enabling humans to work alongside robots seamlessly.
The company has been increasing its spending on research and development year-on-year. During FY 2016, FY 2017 and FY 2018, its spending on research and development was $1.33 billion, $1.46 billion and $1.79 billion respectively. Overall, as some unsolved challenges begin to get solved, robotics will witness a leap in years ahead.
The NVIDIA Jetson Xavier developer kit, which includes the Isaac robotics software will be available starting in August from distributors worldwide.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
