To build truly intelligent ambient environments, organizations are adopting cloud-based platforms that can apply artificial intelligence (AI) algorithms across all kinds of IoT deployments—spanning from the edge to the cloud.
At the airport, you check in by showing your ID and boarding pass—just once—and then breeze through baggage, security and boarding without showing any documents again. After your flight, a self-driving car pulls up directly in front of you at the airport curb, arriving at just the right time to take you directly to your hotel. At the hotel, you are checked in automatically as you enter the front door, and a digital key to your hotel room is loaded to an app on your smartphone. As you approach your room, the door unlocks and the lighting and heating adjust themselves to fit your tastes—all without anyone lifting a single finger.
Making this hassle-free scenario possible is a range of technologies working collaboratively, including surveillance cameras, artificial intelligence (AI) face-recognition algorithms, location-aware apps and the internet of things (IoT). In combination, these and other technologies comprise the building blocks of “ambient intelligence,” i.e., environments that are aware and can interact with individuals. In the era of ambient intelligence, such technologies often are virtually invisible, integrating into the environment to simplify everyday tasks, enhance safety and allow individuals to become more productive.
Many of the capabilities described at the beginning of this article already exist, from a streamlined passenger identification system at Aruba Airport, to Waymo’s driverless transportation service for guests at a Phoenix hotel, to Hilton’s Digital Key, to Marriott’s smart-room project. However, to build truly intelligent ambient environments, organizations are adopting cloud-based platforms that can apply artificial intelligence (AI) algorithms across all kinds of IoT deployments—spanning from the edge to the cloud.
Cloud-based solutions, such as Microsoft’s Azure, are essential to enabling ambient intelligence. Working in combination with edge-computing solutions, these platforms allow customers to quickly develop applications and to manage and maintain devices. They also allow customers to take full advantage of the information gathered by ambient-intelligent systems, using cloud-based analytics and artificial intelligence to turn data into useful insights that can generate new revenue and business opportunities.
Inside ambient-intelligence products
At their heart, devices that support ambient intelligence are IoT devices, i.e., embedded systems equipped with sensors and wireless network connectivity. However, such devices also are defined by other characteristics, including context awareness, personalization, adaptiveness and anticipatory intelligence.
The successful implementation of such capabilities is dependent on AI technologies, including face recognition, object recognition, natural language processing and human action recognition. Some ambient devices are employing limited amounts of AI, instead focusing on hardware and operational technologies, like sensors and communications. However, such inadequate AI capabilities eventually will stymie the capabilities of ambient-intelligence systems.
To overcome deficiencies in intelligence, ambient intelligence devices can send AI algorithms to the cloud for processing. However, cloud processing presents drawbacks in terms of power consumption and performance. As a result, many new IoT devices are integrating more sophisticated electronics—such as graphics processing unit (GPU) microchips—that are capable of handling compute-intensive AI tasks. For example, GPU market leader Nvidia is offering the TX2 Developer Kit designed for building AI-enabled edge devices and software.
A truly flexible ambient intelligence platform should be able to support both cloud and edge AI deployments, providing the flexibility needed to work with any type of device, whether that device has built-in AI capabilities or not.
The following sections examine the solution companies are using for their ambient-intelligence projects.
Shell technology puts safety first at gas stations
Royal Dutch Shell is an energy giant, listed as the world’s fifth largest company in terms of revenue in 2017, according to the Fortune 500 ranking. The company operates 44,000 gas stations worldwide, serving 30 million retail customers.
For these stations, safety and security are paramount concerns, with occurrences such as thefts, smoking, car accidents and refueling mishaps potentially resulting in dangerous and costly problems. Surveillance cameras can collect video of such events. However, to make this video useful, it must be monitored and reviewed—a process that can consume massive quantities of time, money and manpower when conducted manually.
Shell is aiming to provide a faster and cheaper alternative using AI technologies. The company has developed a system that uses closed-circuit camera footage and IoT technology to automatically identify safety hazards and quickly alert staff to quickly respond to potential problems. Deep-learning algorithms located in the cloud and the edge monitor the video feeds and identify any events that could represent a safety or security issue.
This approach is much faster, less expensive and less labor intensive than manual review.
When building a pilot version of this system, Shell chose Microsoft’s Azure platform to serve as the basis. Azure is designed to support the building and management of applications and services. Microsoft’s Azure IoT Edge offering is a cloud service that can distribute and run algorithms on cross-platform IoT devices, allowing AI applications to run both in the cloud or offline.
Honeywell offers building-automation solution
Product and system vendor Honeywell is a leader in commercial building automation systems and products. Honeywell now is offering the Honeywell Vector Occupant App, designed to serve the needs of building managers and occupants by promoting greater efficiency.
The Vector Occupant App leverages IoT connectivity to deliver features including indoor location, mapping, routing, presence, proximity notifications and analytics. With the App, building occupants can unlock doors remotely, issue their opinions on room comfort levels and reserve meeting spaces from their smartphones—among other capabilities.
Honeywell chose Azure to power its app. Azure provides cloud-based big-data processing that allows the app to enhance the usefulness of its analytics.
View SmartProtect smartens up windows
View Inc. is a provider of smart-window products. The company’s SmartProtect solution can automatically and instantly detect any glass breakage. When breakage occurs, a warning is sent through Azure IoT to SmartProtect, informing customers about the time and location of the breakage.
View stated that SmartProtect is just the first of many Azure-based IoT solutions it plans to offer. The company aims to bring greater intelligence and personalization to buildings via smart window technology.
The ambient intelligence future
With the technological building blocks in place, the number of ambient technology deployments is likely to rise dramatically in the coming years. Organizations engaged in ambient intelligence projects need to adopt cloud platforms that can support both cloud and edge deployments to develop exciting new applications. Cloud-based solutions like Azure offer capabilities that are essential for deploying and managing IoT systems, as well as for conducting data analytics required in successful ambient intelligence deployments.
Original article found on eetimes.com