A Perfect Fit: Wearables and Speech
The computer industry has always been driven by the synergistic competition between hardware and software. As our society conceives of new ways of using computers, code jockeys create ever more capable and complicated software to accomplish the task. Software performance demands in turn challenge hardware developers and designers. Hardware improvements continue the cycle by inciting even more demands from society. Recent improvements in speech recognition software continue this contest in the desktop world. For the wearable computer platform, however, these improvements herald a quantum leap in usability. For the first time in the nearly 40 year history of "wearable" computers, reality can meet society's expectations. This synergy bodes well for the growth of this once nascent segment. For those not familiar with wearable computers, these systems distinguish themselves from other portables in the areas of human interface and ergonomics. A true wearable is initially donned as one would put on a piece of apparel. Once activated, however, the system should interface with the user in a natural, unobtrusive manner. Humans ordinarily transmit information through the mouth with speech and receive information through the eyes and ears. An important aspect of this fact is that the tools that we use for this information transfer are located on the head. Thus, an ideal wearable computer would not only provide a seamless interface for aural and visual communication, but also remain accessible to the user's mouth, ears, and eyes throughout the range of daily activities. Early adopters and wearable system integrators are aggressively working to make wearables effective in the operational environment. The hardware manufacturers have done their part by designing and building affordable wearable computers that can support "hands-free operation." One example of this idea is the deployment of Xybernaut Mobile Assistant III wearable computers by the U.S. Customs Service. Customs officers on the Arizona-Mexico border search outbound vehicles and verify that these vehicles are not stolen. This task requires them to query three mainframe databases over the USCS's wide area network. In a prototype system designed by Customs and SENTEL Corporation, customs officers wear a Xybernaut computer with a head-mounted monitor, microphone and speaker. While connected to the USCS WAN, an officer reads aloud the license plate number and other identification data. With the Dragon Dictate speech engine, a database query is created and the response is presented to the officer in the monitor.
Speech in the Wind
Although the system shows promise, there are some real problems to overcome. Accurate speech recognition in a stiff wind is difficult. There are also hurdles associated with speaker accents and dialects. One especially challenging application is at NASA's Kennedy Space Center, where scientists are looking into how wearables can improve their quality assurance operations. They are experimenting and conducting usability tests with the ViA wearable computers from interactive solutions. By using voice commands, a video camera and wireless Internet access, an inspector could work with a scientist located anywhere in the world. It is obvious that such rigorous operational requirements need powerful speech recognition and synthesis software. What the speech technology community needs to grasp, however, is that wearable hardware provides the perfect platform for exploiting the capabilities of speech technology. A user does not need speech on a desktop. For a wearable system, however, this capability is imperative. This rare software and hardware synergy calls for focused cooperation between wearable hardware vendors, speech software developers, and those system integrators that provide wearable solutions to the end users.
The question now is how should we proceed to develop this new market segment? One task is to convincingly show significant benefits in using these technologies. There are many vertical markets where wearables provide such benefits, but market growth will not occur unless it can be convincingly demonstrated. Pilot projects are the tool for making these convincing demonstrations and the system integrator is positioned to make a pilot project work. Only the integrator, who is intimate with a particular vertical segment, has the experience and contacts needed to ensure success. It is also important to note that it's easier for government agencies to go the pilot or prototype route because they can justify an experimental project solely in process improvement or service enhancement terms. Commercial businesses by their nature must focus on monetary returns on investments. Without a significant ROI track record only the largest companies can even think about using advanced wearables and speech technology. Once significant service and process benefits are shown in the government sector. Although system integrators are key to developing successful pilot projects, they are not well positioned to identify pilot project opportunities. A prospective customer researching a new technology will inevitably first call the technology developer. Since the developer is first to hear from prospective early adopters, they can direct these prospects to appropriate system integrators for pilot project development. Strong alliances between developers and integrators are therefore essential for segment growth. Wearable hardware and speech software manufacturers should actively partner with integrators in targeted vertical markets. Don't forget to include government segment integrators in your plans. Those integrators will close the critical early sales needed for commercial verticals. Wearable computer market growth will depend on the capability of speech recognition software. Widespread deployment of speech-driven solutions depends on growth in the wearable computer market. This synergy is critical to both markets and must be supported by strong alliances between wearable hardware manufacturers, speech software developers, and wearable system integrators. This is the formula for rapid growth in both markets. By harnessing this synergy and focusing on customer satisfaction, the dreams of both wearable manufacturers and speech software developers will be realized.
Kevin L. Jackson is the Chief Technology Officer for SENTEL Corporation and may be reached at KJackson@Sentel.com or (703) 739-0084.