Issue link: https://www.e-digitaleditions.com/i/1414667
October/November/December 2021 I 17 Voice Control for Touch-Free Meetings Voice control has recently become a legitimate and attractive alternative to interacting with high-touch surfaces. Amazon Alexa, Google Assistant, Apple Siri and other virtual assistants commonplace in the home have business variants that enable touch-free voice-controlled meetings and collaboration sessions. Voice control is finding its way into scheduling apps, AV systems, building control systems, and UC platforms. By con- versing interactively with a virtual assistant, users can perform a wide range of operations that would normally entail touching keyboards or control panels, including: • Search for and reserve meeting rooms • Start, end, and extend meetings • Auto-connect with remote attendees through contact lists • Turn on the AV system and control any AV function normally accessed through a control panel • Set room lighting, window shades, and room temperature Voice assistants rely on cloud-based artificial intelli- gence (AI). The diagram in Figure 13 shows implemen- tation of an Amazon Alexa voice control system and is typical of most voice assistants. Alexa Skill operates in the Amazon Web Services (AWS) cloud and works with a client app running in the user's computer or AV system control processor. The user speaks into a mic a command using a trigger word. In this example, the mic is contained in an Amazon Echo device. Alexa Skill translates the speech to a JavaScript Object Notation (JSON) data string that is passed to skill application code, called a lambda function. The lambda function parses the JSON data into the parts needed to trigger the desired hardware actions, based on a lambda function list of spoken words that a user could use to signal intent. The parsed commands are delivered from the cloud to the local client app via Amazon Simple Queue Service (SQS), a managed message queuing service. The client tells the control processor to output the required hard- ware command to the applicable device (e.g., turn-on the projector). The device is commanded to the required state via hardware driver code communicating via the control processor port appropriate to the device being commanded (e.g., RS-232, USB, Ethernet). The entire process is reversed to translate status from the commanded device into verbal confirmation that the command was executed. All voice assistant providers charge on a per-access basis for the cloud compute services that make them work. But with liberal free usage tiers (e.g., AWS allows one million free requests per month), most users will never make enough requests to incur charges. Security safeguards are of more concern to some potential users. Virtual assistants are always listening and can accumulate information that is not relevant to their purpose. The good news is that the virtual assistant services provide users with tools to control their privacy. They can mute the mics connected to the service, periodically erase history in their accounts, and tighten account settings. Echo Dot Lambda Function Amazon SQS Control Processor Hardware Drivers Device Being Controlled Local Client App (running on user computer or AV Control Processor) JSON Command Voice Response Response Message Response Message Status Response Destination intent:device type intent:device type Command intent:device type FIGURE 13: Amazon Echo, acting as the interface for Alexa, processes verbal requests into commands for hardware devices and generates verbal confirmations from hardware device status messages.