Mar 19, 2020 Industrial

AIoT Has a Nice Ring To It

By Matt Jones

Matt-Jones-headshot-2

The Artificial Intelligence of Things (AIoT) is a relatively new term for the evolution of a domain Wind River has been playing in for a very long time. If we think of many of the first applications of Artificial Intelligence in connected devices, it is adding autonomy to previously human controlled systems. Think advanced autonomous drones, automated driver assistance features in vehicles, or even autonomous factory robots. These autonomous systems tend to still interact with humans, and as such they are safety critical. More so, they’re connected and have associated security risks. Importantly, you can have a secure device that doesn’t deliver safety critical functionality, but you cannot have an insecure safety critical device.

These connected autonomous systems are incredibly complex, and require an intelligent systems platform from the device edge to the infrastructure edge to the cloud; and in the device they require real-time operating systems (RTOS) with guaranteed performance, coupled with AI/ML algorithms that are mostly associated with Linux. The use cases and requirements span the complete system. The systems may require containerized applications running in the cloud and on edge devices. The systems may require AI/ML frameworks that span the RTOS and Linux on the device. Wind River can offer the complete package with its comprehensive software portfolio.

On the Linux front, Wind River Linux, based on Yocto, is the market leading embedded Linux distribution. Wind River is among the top contributors to the project. To achieve this, we build Linux over 3000 times a day, run over 12,000 tests every night, and a complete set of over 140,000 tests for every release – and we now release it publicly every three weeks with the Continuous Delivery Model. Wind River Linux also supports containers, enabling applications from the cloud to run on devices.

On the real-time front, VxWorks is the market-leading safety critical RTOS, with over 500 certified programs and over 2 billion devices deployed.

Wind River Linux and VxWorks, with the Wind River Hypervisor, enable the device; but we also support a range of AI/ML frameworks. If you visit Wind River Labs, you’ll see that we have examples of ROS2 (Robotic Operating System) and Tensorflow running across the embedded platform.

To complement this, we have the Wind River Cloud Platform, based on the StarlingX open source project we sponsor. It takes a range of cloud technologies, including Kubernetes and Docker, and optimizes them for our low latency Linux distribution. This platform allows the scalability of cloud technology to the edge of the network, such as in 5G cell towers, and even beyond.

With our differentiated portfolio, Wind River has a complete intelligent systems platform – enabling our customers to design, develop, simulate and deploy their connected intelligent and autonomous systems faster than ever before. Check out our Byte Size Discoveries video series to learn more about how Wind River technology is powering everything from mission-critical flying cars and drones to medical devices and robots.

Previous Digital Twins are hard. Maybe that’s good.
Next Surviving the Frequency of Open Source Vulnerabilities