?
Carrier Networking Mobile Edge Computing
4 minute read

AICUDA Vaidio: Converged Edge Platform ISV Spotlight

In today's world of camera availability everywhere from buildings to street lights, technology has enabled businesses to provide artificial intelligence (AI) analytics using computer vision technology to interpret information seen in images and videos.

As the remote workforce continues to expand, mobile edge computing (MEC) has become increasingly important in enabling the device applications remote employees need to do their jobs.  

Low latency, high bandwidth scenarios have grown, which has led to an increased demand for GPUs at the edge to provide processing of data for video feeds, gaming, mixed reality applications and other solutions leveraging the NVIDIA Metropolis application framework

The WWT Converged Edge Platform (CEP) provides an edge-based computing solution to allow consumers to meet the demands of low latency, high bandwidth processing of edge applications and virtualized services. Solutions like private 4G/5G networks can be created within the Converged Edge Platform, while hosting GPU and non-GPU intensive applications for a variety of use cases. 

Our Converged Edge Platform reference architecture starts with generic x86 compute architecture, and then layers the VMware Telco Cloud portfolio on top as the infrastructure for hosting and managing edge workloads delivered as containers and virtual machines. The MEC runtime environment using MobiledgeX creates a single cloud console and SDK for developers and operators to manage container clusters and application workloads. 

One use case recently validated in our Advanced Technology Center (ATC) using the Converged Edge Platform is from NVIDIA’s Metropolis partner ecosystem — AICUDA’s Vaidio, which provides video search and AI analytics for security, health, safety and operations. 

Live video streams and archived video files process massive amounts of data to deliver real-time notifications using AI-engines. In order to accomplish this at the edge, compute processing with physical GPUs handle the mathematical algorithms that filter videos based on the AI engine being utilized. 

Instead of deploying compute clusters at each site to support thousands of cameras, the Converged Edge Platform can be centrally located to run the MobiledgeX cloudlet and AICUDA application, which lowers the cost to the consumer using bandwidth thinning along with meeting the lower latency requirements for operating at the edge.

AI analytics using computer vision technology provides the means to support a variety of organizations.

  • Traffic management, such as monitoring racing events or street traffic.
  • Retail environments, such as tracking high-trafficked areas of a store or reporting on inventory using robotics or handheld IoT devices.
  • Factories, such as watching for incidents, tracking product assembly, or monitoring equipment for heat, smoke, pressure or failures.
  • Public safety and health. For example, tracking social distancing, monitoring spills to avoid accidents, keeping an eye on mask wearing or detecting other health issues.
  • Warehouse and logistics, such as inventory counting, and shipping and receiving.
  • Stadiums and casinos, like people counting, or monitoring for public safety(objects left behind, accidents, etc.).
  • Access control, such as door sensor integration, facial recognition and compliance entry.
  • Transportation hubs with object detection or airplane maintenance.
Intelligent video use cases.
Intelligent video use cases.

Here’s a quick summary containing some of Vaidio’s artificial intelligence analytics:

  • Crowd management
  • Equipment status detection
  • People identification, including facial recognition
  • Incident management
  • Personal protection equipment (PPE) detection
  • Infrastructure and security management

Like the Converged Edge Platform, AICUDA is designed to support a variety of industry verticals, including manufacturing, healthcare, public sector, service provider, telecommunications and more. 

At the bottom of this article are some sample captures from Vaidio using Cisco Meraki RTSP cameras setup in the Advanced Technology Center, as well as archived video footage. Here you can see AI analytics being applied to detect various use cases we defined during the validation process. The edge servers processing the analytics within this cloud-based application included NVIDIA's V100 Tensor Core GPU.

World Wide Technology provides services for onboarding independent software vendors (ISV) like AICUDA as part of an engagement that leverages our VMware Telco Cloud products in the Advanced Technology Center. This engagement is known as the WWT Edge Innovation Studio.

For more information, check out our Converged Edge Platform hub or get hands-on access with our Converged Edge Platform Lab in the ATC

Intrusion Detection
Intrusion detection
Vaidio with Video Search Analytics
Vaidio with Video Search Analytics
Make and Model Recognition with Vehicle Counting
Make and Model Recognition with Vehicle Counting
Person and PPE Detection
Person and PPE Detection