Redefining the IoT Edge Experience at CES 2018

At a minimum, a successful IoT solution includes sensing, connectivity, processing and security. While each of these elements is widely available from a number of vendors on its own, they often aren’t all available from a single source.

By now you’ve certainly heard of the Internet of Things (IoT) and all that it promises in terms of convenience and efficiency.

Historically, OEMs have met this challenge by adding resources to their in-house development teams. In the era of incremental product development, this strategy was effective, as the addition of engineers with similar skill sets allowed manufacturers to scale their efforts while maintaining cost and time to market goals.  However, what OEMs need today is a set of highly integrated, scalable technologies that support synergistic IoT product and system development without requiring their engineering team to do extensive upfront, base level capability development and integration.

Embracing the Power of the Edge

IoT deployment is actually a collection of many complex systems integrated together to work in a coherent manner.  The vast majority of these systems operate largely at the network edge.

Technology suppliers are taking note. At CES 2018, NXP Semiconductors is showcasing an Edge Compute Experience that includes seven linked kiosks in an IoT Edge Corridor, highlighting application areas such as machine vision, machine learning, smart retail, digital signage and the smart home. The kiosks are powered by a scalable suite of intelligent semiconductor solutions ranging from ultra low-power Kinetis microcontrollers (MCUs) to multicore i.MX 8 applications processors to Layerscape edge compute and communications processors. But the real achievement is that these substations are all connected to form a real-time local edge computing network that can function independently of the cloud.

Figure 1: NXP Edge Compute Experience Technical Overview

Upon entering, attendees will be greeted by a machine-learning-based facial recognition platform powered by Layerscape processors that correlates biometric identity with an NFC-enabled name badge. Similar systems are positioned throughout the IoT Edge Corridor that allow the network to track booth activities at various levels of granularity, such as what kiosks certain individuals frequent, feedback provided by visitors, which kiosks are the most popular, and the total number of booth guests.

Guests will encounter several other powerful examples of IoT edge compute. Among these is the embedded Amazon Alexa capability at each kiosk that processes voice commands locally on an i.MX 6 processor and/or on the cloud, while another station demonstrates the Au-Zone DeepView computer vision neural network running on i.MX 8 to identify types of food stored in appliances. Capabilities such as intelligent sensor data interpretation, remote device management and secure device provisioning are a constant throughout.

All of these elements are connected using QorIQ Layerscape communications processors that incorporate high-speed packet processing accelerators alongside multiple Arm® Cortex-A72 cores. This versatility allows Layerscape devices to function as IoT edge gateways that reduce latencies and privacy issues associated with cloud connectivity by locally running the advanced facial recognition training algorithms that are typically run in the cloud. At the end of the IoT Edge Corridor, these capabilities are exhibited on a dashboard that displays the booth data, as well as personalized attendee information.

Go from Prototype to Product at CES 2018

Layerscape-based gateways also facilitate cloud communications, as demonstrated in the IoT Edge Corridor through attendees’ ability to send a picture of their choice to the Google Cloud platform and process it with DeepArt.io AI algorithms, it is then returned to the edge to be selected and printed off as a souvenir piece of art.

Examples of customized artwork

The emergence of viable, self-contained IoT edge networks reinforces the need for product OEMs to focus on their core competencies. Getting to market quickly and economically is critical, however the ability to iterate on top of those platforms later is the key to success.

This approach requires a foundation of highly integrated processing and communications platforms that abstract the rigorous engineering of connectivity, security, machine learning and other disciplines that most product OEMs don’t have time to leader. It also demands robust, commercial-grade software support that accelerates product development lifecycles.

These elements are present across NXP’s portfolio of intelligent edge compute solutions, and enhanced through application-specific development kits that allow OEMs to realize their own opportunities at the IoT edge today.

Don’t believe it? I challenge you to visit NXP’s IoT Edge Corridor at CES 2018 booth CP 25 and experience the edge for yourself.