posted on Friday, 17th March 2017 by Geny Caloisi
This year has started with a loud and clear message: voice is the next big thing in home automation control. At CES and ISE we have seen an explosion of companies showing voice operated solutions, creating interfaces, and opening up the possibilities that voice command can bring to our daily lives. At ISE 60 per cent of manufacturers showed integrations with Alexa, Amazon Echo’s sophisticated voice control platform.
Recently, CEDIA teamed up with Amazon to organize a fascinating talk addressing the use of voice commands to manage and operate devices within the home and how custom installers can do their own programming using Amazon Smart Home Skill Kit. Naturally, Inside CI was there...
Dean Bryen, solutions architect at Amazon and Alexa and Echo evangelist, opened his ‘Mastering the Skills of Voice Control in the Home’ presentation quoting AI visionary Tim Tuttle: “Soon devices won’t have keyboards.”
Hands free, voice operated systems are not new. The first time the term ‘voice technology intelligence’ emerged, was in the 1950’s. In 2007, a CNN business article reported that voice command was over a billion-dollar industry and that companies like Google and Apple were trying to create speech recognition features. Apple introduced Siri in 2010, and Amazon Echo and Google Home followed suit.
Two main factors are making voice control more appealing: the increased accuracy of Automatic Speech Recognition (ASR) and advanced machine learning. Both of these have vastly improved in the last four to five years, according to Bryen.
“The difference between 95 per cent and 99 per cent accuracy when it comes to voice is amazing,’ said Bryen and added, “In 2014, voice search traffic was negligible. Today it exceeds 10 per cent of all search traffic. Siri, Google and Cortana exceeded 50bl voice searches per month. By 2018 30 per cent of interaction with smart machines will be through conversation. By 2020, over 200 billion searches per month will be done with voice.”
Voice commands are by nature conversational. People using a voice command will have their own accents, ways of calling things, and, there might also be a wide range of interlocutors of different ages that will want to use it. Installers will require intuitive innovation and a sense of playfulness to create Voice user Interfaces (VUI) that, literally, resonate with their clients.
Conversational use of voice commands will mean that, instead of asking for something to be switched on or off, a user will say “Good morning”, or “I’m Home,” or “I’m babysitting,” or “Good night” and this will be linked to a set of actions such as drawing the curtains and starts the coffee, turns on the lights and the TV, etc, creating the desired scene.
Bryen admits that Amazon’s Echo still faces some challenges, such as security issues, hierarchies differentiation for users, and location and contextual awareness.
“When Alexa is in a room, it would be easy to get a bit of ‘spy paranoia,’ admits Bryen, “but Amazon has earner the trust of its customer before and it will do it again. In 1995, Amazon asked its customers to add their credit card details to its platform and, although some people might not have felt at ease to start with, today is normal. Alexa will only save information on the cloud after its name is mentioned. Voice commands, using Natural Language Use to recognise the intent, are sorted in the cloud. All consequent machine learning is anonymized.”
The process between the voicing of the command and it being carried out can take 0.5 seconds and it only requires a 16GB broadband capacity, so any home in the UK could have Alexa.
Bryen painted a very favourable picture for integrators and the possibilities that VUI (voice user interface) customisation can bring. Studies have shown that 50 per cent of consumers say they plan to buy at least one smart home product in the next year. Although 60 per cent of consumers self- install smart home devices, the majority of device owners said that they would prefer professional assistance.
Integrators that are able to develop a good VUI will also reap the benefits of being able to provide a better customer service and maintenance. Updates on the cloud make the added value of the service provided by the installer a reality as well as it opens up for hardware upgrades and better margins.
How can it be done? By using Amazon’s Smart Home Skill API.
The skill kit requires a bit of programming knowledge but it’s not too dissimilar to HTML language. To carry out your own VUI you’ll need: an Amazon Developer Account, an AWS Account and knowledge of Java, Node.js, or Python.
Skills are made up of two components Skill configuration in the Amazon Developer Portal and Your skill code, hosted in AWS Lambda and the developer’s own HTTPS endpoint.
The language used is called JSON and a line of code will look something like this:
Header { "header": { }, "payload": { } } Payload
Very simplified, the elements on the programming narration will include: invocation names or launch request (i.e. ‘Alexa’), actions, custom slots (which identify an argument within an intent), and utterances. The system seemed simple enough to try it out!
Matt Nimmons, Managing Director of CEDIA EMEA concluded, “Voice control is one of the greatest technological advancements our industry has seen in recent years and is fast becoming common place in the home. It is imperative for home technology professionals to be knowledgeable on how to integrate voice control into the home and how to use and promote this technology to ensure it is a profitable business venture. As one of the leading voice control platforms in today’s market, we are delighted that Amazon has chosen to host this event, exclusive for CEDIA, in order to help professionals get the most out of this ongoing phenomenon.”
Geny Caloisi is an accomplished technology journalist who has worked in a variety of AV industry publications.
Inside CI © 2011 | Web Development by Chocolate Grape