Those of us who have been following development of a Conversational Interface for the Internet of Things (IoT) are very familiar with the myriad of challenges confronting product or service developers. We all start with a vision, rooted in the old Star Trek TV series and perpetuated in subsequent movies and sequels. Whenever, wherever a person needs to invoke the aid of a computer or computer-based resource, that system stands “at ready,” apparently listening for a trigger word to get started on its merry way.

My personal favorite is when Scottie or some other denizen of the holodeck has the ready-to-please system replicating his or her favorite cocktail whenever it is asked. “Computer,” he may say and then continue, “I’ll have another.” It’s just that simple. As anyone who has invoked speech recognition in an automobile by pressing a button on the steering wheel well knows, this is all  too futuristic with today’s technology. That’s where 3iLogic-Designs enters the picture. Today this speech processing specialist with offices in Fremont, CA, and R&D in Gugaon, India, is formally introducing SimSim, automated speech processing resource designed to be integrated with the low-cost microcontroller or application specific processors that power today’s consumer electronic devices.

The implications of SimSim are profound. If it works as described, it can be that ubiquitous, always-on resource with a relatively large vocabulary of “take action” words that work at low power and without a connection to the cloud. 3iLogic-Designs takes a hardware-based approach, calling SimSim “the world’s first speech recognition hardware engine in the form of an IP core. SimSim features a high-performance, ultra-low-power, embedded design that resides on an extremely small footprint and is fully self-contained.” It is also capable of running in a speaker-dependent mode. The company is in the process of lining up relationships or deals with chip makers whose core products are the microcontrollers (MCUs) that act as the brains of all the “smart” appliances, mobile phones and tablets, consumer electronics, and automotive or home automation peripherals.

SimSim speech recognition capabilities seem to be different from other offerings because of its small power requirements and the small size of the program itself. Management claims that it is also able to conduct speaker recognition and can, thus, be a tool for personalization. So that brings us ever-closer to Star Trek’s world, where people are especially comfortable talking to their computer-powered devices, whether they are connected to the cloud or not.

Categories: Conversational Intelligence, Intelligent Assistants, Articles, Mobile + Location

Tags: Speech recognition, Voice-Control, Connected Car, embedded, connected home, hands-free