Infrastructure delivers autonomous edge solutions
- By Stephanie Kanowitz
- Aug 10, 2021
The Texas Military Department (TMD), Camp Mabry, in Austin is pilot testing new technology that fosters the development of autonomous solutions at the edge.
The Public Infrastructure Network Node (PINN), built by distributed edge computing platform provider EDJX in partnership with the Autonomy Institute, is the first open standard to incorporate 5G, edge computing, radar, lidar, enhanced GPS and intelligent transportation systems into one system. The nodes, which look like streetlight poles, are designed to deliver edge sensors and low-latency computing capabilities needed to support autonomy and the internet of things.
PINNs “will enable TMD to accelerate emergency and disaster response for intelligent infrastructure [and] accelerate humanitarian assistance from disaster relief and domestic operations,” Lt. Col. Alex Goldberg, Texas lead for at the Defense Innovation Unit Joint Innovation Office, wrote in an email to GCN. Specifically, the nodes will allow the department to deploy autonomous vehicles such as helicopters and drones instead of putting humans at risk during search and rescue missions. “We could engage drones to determine people on rooftops that need rescue or roads that are blocked because of down power lines.”
Cell towers are often separated by miles, but advanced sensor networks require denser technology. The PINNs a variety of sensors in a single location, which can be deployed across sites. The 400-acre Camp Mabry will have up to 34 PINNs placed 1,000 feet apart.
To determine where they’ll go, EDJX and the Autonomy Institute are working with Esri, NVIDIA and Bentley Systems to create digital twins of the environment to allow for the testing and propagation of signals such as lidar, radar, 5G and Wi-Fi, said Jeffrey DeCoux, chairman of the Autonomy Institute and CEO of Atrius Industries, an autonomous infrastructure firm.
“Historically, the military has operated very much in a disconnected state, [focusing] on the bullets, the guns, the tanks,” DeCoux said. “Now, because of Industry 4.0 and the networked world that we’re in, it’s very much about networking those things together.”
PINNs capture sensor data, which is then written to EDJX’s EdjNet platform, which allow for that data to be processed by artificial intelligence algorithms, machine learning, augmented and virtual reality as well as predictive analytics.
“The data is written once to a PINN and the availability, or the access to that data, is made across however many nodes make up the plurality of the network so that when code that a developer writes needs that data, it makes the request of the network rather than a specific PINN or a specific location,” EDJX CEO and Co-founder John Cowan said. “That’s what allows us to achieve the ultra-low latency that’s needed for a lot of these internet of things-style solutions.”
Goldberg said that he expects to see several benefits from the use of PINNs, including driving research and development into autonomous systems and edge networks that the Defense Department could not accomplish without private-sector help.
Current computing and network architectures can’t support the low latency that connected and autonomous systems require, he added. “Ground and low-altitude autonomous systems require sensor fusion at fixed positions to address non-cooperative traffic,” Goldberg said. “This includes high-profile systems like self-driving cars, flying taxis and helper robots. It will also include a background of devices that largely remain unnoticed as they perform routine activities.”
Another benefit of the pilot is that it will show society at large the advantage of using artificial narrow intelligence. “DOD in many ways can be the first customer,” Goldberg said. “And if it’s good enough for the DOD, from cybersecurity to logistics security, to all of these other elements, then civil society is ready to adopt them.”
He said he expects the PINNs to be installed next month.
Cowan and DeCoux said their goal is to have tens of thousands of PINNs deployed by mid-2022 in cities, on highways, across military installations and in rural communities.
PINNs will be funded through private industry and public-sector collaborators, and they are supported by technology firms such as Hewlett Packard Enterprise, which is contributing technology and expertise to operationalize PINNs.
Stephanie Kanowitz is a freelance writer based in northern Virginia.