Smart watches need intelligent systems

With analysts projecting that 15 million wearable tech devices will be sold worldwide in 2013 and sales increasing to 75 million devices by 2017, the hype around smart watches and other wearable tech is unavoidable. One common requirement for these devices is a system that manages the information flow and human communication between user and device - what Nuance refers to as "intelligent systems."
Wearables like smart watches need intelligent systems to enable a meaningful human-device interaction

The hype around smart watches and other wearable tech is unavoidable these days. That hype has already turned into a big business. Technology market tracker Strategy Analytics projects that 15 million wearable tech devices will be sold worldwide in 2013 and sales will increase to 75 million devices by 2017. The category is broad, including watches, glasses and fitness monitors. Functionality ranges from simple sensors that gather data to intelligent peripherals that are paired with smartphones – all the way to powerful, standalone systems.

One common requirement for these devices is a system that manages the information flow and human communication between user and device. At Nuance, we call these things “intelligent systems” (see below) because of their ability to communicate on human terms, understand language, manage conversations and interpret contextual information to provide the right information in the appropriate format.

Without the support of intelligent systems, wearable tech could quickly devolve to the metaphorical VCR flashing 12:00. With the right application of intelligent systems a smart watch delivers on its promise – an ever-present connection to a broad world of media and information.

The size and format of smart watches provide a level of convenience that consumers desire, but also present a series of challenges that need to be elegantly overcome to deliver on the potential of this brand of wearable tech.

First, let’s address the screen size. With a small screen, you can’t rely on a user interaction model that was developed for a smartphone or a personal computer. One important requirement for these devices will be a natural language interface that provides the ability for the user to get access to deep functionality and rich content without memorizing a specific set of commands or keystrokes (remember the VCR). Natural language can be applied not only to voice commands, but also gestures and other input modalities to offer a more flexible interpretation of the user intent.

Wearable tech will also benefit significantly from the application of contextual awareness to set itself automatically into the appropriate operating mode. For example, the device should be able to conclude that you are driving and set itself into a hands-free mode (much like Dragon Mobile Assistant does today). Other smart modes might include sleep mode (don’t bother me unless it is time to wake up or there is an emergency), exercise mode (turn the face of the watch into a gesture-based media selector when you determine that I am on the treadmill), meeting mode (don’t bother me unless you have an update on my fantasy football team) or even TV mode (make my watch a remote control when you can tell I am watching TV at home).

The ability to offer these sophisticated capabilities brings up another challenge with wearable tech – the limitations in processing power and related battery life challenges associated with a device that can fit on your wrist. Today, intelligent systems require the processing power and data access associated with high-end server farms. This means that wearable tech needs reliable cloud connectivity to deliver on its promise.

Audio processing is another important factor in performance. Ideally, you will be able to issue a voice command in the general direction of your smart watch – a feature that will require advanced audio handling to suppress any ambient noise. Voice biometrics can be applied to make sure your annoying friend doesn’t shout “smart watch: erase my hard drive” or something equally destructive. Intelligent systems will also be able to determine whether you are in a noisy environment and turn off some voice features to eliminate false recognitions.

Not all wearable tech will have all of these features. In some cases, there will be trade-offs that control cost, size, power and other design impact. Some devices may require Bluetooth audio connections, for example, to work within the design constraints.

Overall, the promise of wearable technology is…promising. Not only for the Tony Stark wannabes, but also for the rest of us who want safe, convenient access to information, media and entertainment. To be sure, the promise relies on the appropriate application of intelligent systems. I don’t want my smart watch flashing 12:00.

What are intelligent systems?

Intelligent systems are large-scale software systems that mimic elements of the human brain – they operate like the central nervous system when paired with devices and sensors that act like the peripheral nervous system of the human body. Intelligent systems are designed to interpret human forms of communication, including speech and language, visual cues and touch. These systems not only recognize human communication, but they also derive meaning and understand the intent of the users. Like the human brain, intelligent systems also consider context when deriving meaning. Context can include basic location (your point on the map) and sophisticated location (you are home, in the car or in a meeting). Another important capability provided by intelligent systems is dialog management – the ability to carry on a conversation with a user and respond appropriately to their initial requests and follow-up responses.

Intelligent systems are used in many applications including personal assistants for consumers and enterprise applications for customer service and healthcare information management.

Intelligent systems are behind a wide variety of offerings from Nuance including consumer products like Dragon Mobile Assistant, enterprise customer service solutions like Nina Mobile and even large healthcare systems like Nuance Clintegrity 360.

Tags: , , , ,

Peter Mahoney

About Peter Mahoney

Peter Mahoney is the senior vice president and general manager of Dragon for Nuance’s Healthcare division. He is responsible for all aspects of the Dragon speech recognition business, encompassing the healthcare, enterprise, and consumer markets. Peter joined the company in 2004 as the vice president of worldwide marketing for the speech division. Prior to Nuance, Peter held leadership roles in marketing, product management and operations at Performaworks, ATG, Engage and PictureTel after a successful career in sales and alliances at IBM. Peter also serves as the Chairman, Board of Directors for Easter Seals Massachusetts, a non-profit, community-based health agency dedicated to helping children and adults with disabilities attain greater independence. Peter earned a B.S. from Boston College with a double major in Physics and Computer Science.