Computex 2018: Building smart robots gets easier with this new tech

• The new computer houses 9 billion transistors.
  
• This computer can perform over 30 trillion operations per second.
  
• NVIDIA’s new computer comes with a software development package to support Artificial Intelligence (AI) software development.
  

On Sunday at Computex 2018 in Taiwan, Jetson Huang, the co-founder and CEO of NVIDIA, showcased two of its newest products developed to create, program and train intelligent robots. With the two new products Jetson Xavier and NVIDIA Isaac, making an intelligent robot won't just be easier, but it will be cheaper as well.

Jetson Xavier is the world's first computer which is specifically designed for robotics, according to NVIDIA. The new chipset houses 9 billion transistors, that have a capacity to deliver over 30 teraflops (trillion operations) per second. And all this performance is delivered by consuming only a third of the energy that a light bulb uses.

This computer also houses half-a-dozen processors including a Volta Tensor Core GPU, core ARM64 CPU, dual NVDL deep learning accelerators, an image processor, a vision processor and a video processor.

“This level of performance is essential for a robot to take input from sensors, locate itself, perceive its environment, recognize and predict motion of nearby objects, reason about what action to perform and articulate itself safely,” says NVIDIA’s press release.

To support and make all this processing power accessible to the users and to extend the performance boundaries of such an incredible piece of hardware, NVIDIA also introduced the NVIDIA Isaac.

NVIDIA Isaac is an advanced software which has three components:

• Isaac SDK: Just like Android comes with the Android SDK, NVIDIA has introduced its own software development kit which will empower developers with a collection of tools to create their own AI software.
• Isaac IMX: Isaac Intelligent Machine Acceleration applications is a collection of NVIDIA-developed robotics algorithm software. NVIDIA also claims that this would save developers “months of development time and effort”.
• Isaac SIM: It is a simulated training environment where developers can train their AI models.

NVIDIA is planning to introduce the Jetson Xavier computers with a price tag of $1299, which is almost one-tenth the price of workstations with comparable processing power.

If this computer lives up to its promises it could revolutionise the AI industry and make AI development more affordable for students and professionals.