Vocollect Voice Technology Overview - DatasmithSA/Pigeon-Voice-Training GitHub Wiki

In order to understand the process of implementing the Pigeon Voice Server in a voice solution it is important to get a brief background on Vocollect's voice technology and their product offerings.

Vocollect are the providers of both hardware and software products to enable the implementations of voice solutions. The solutions are brought to market and implemented by Vocollect partners i.e. VARs (Valued Added Resellers). The partners develop the customised voice solutions for end customers (warehouses) by providing professional services.

Hardware: Vocollect provide their own hardware which are mobile devices as well as headsets and accessories (chargers, belts etc.). Additionally, they also provide their software voice engines to 3rd party OEMs (Original Equipment Manufacturers), such as Honeywell, Zebra and a few others. Thereby allowing the Vocollect voice recognition and TTS (Text-To-Speech) technology to run on a multitude of devices. Below is an example of Vocollect's A700 series device and their SRX2 wireless Bluetooth headset.

/images/A700.png /images/SRX2.png

Software: Vocollect provide a multitude of software products:

  • Voice engines: Vocollect voice engines provide the only voice recognition technology in world that allow for speaker dependant voice recognition i.e. an operator records their words prior to usage and the voice recognition is performed against their own recorded voice. Hence eliminating the typical issue of people with varying accents not being understood by the voice recognition engine.
    • Voice Client: the first generation voice engine provided by Vocollect that encapsulates a voice recognition engine as well as a TTS (Text-To-Speech) engine. It also provides a framework for sending and receiving messages between the device and a server application.
    • Voice Catalyst: the second generation voice engine that improves on its voice recognition quality with adaptive voice recognition as well as as provides many other great features.
  • Device & Operator Management Software:
    • Voice Console: a website that allows administrators and warehouse supervisors to manage all devices, operators and device voice applications across multiple warehouses.
  • Software Development Tools:
    • Task Builder: the first generation IDE (Integrated Development Environment) for developing client voice applications to be executed on voice devices. The compiled voice applications are referred to as Tasks. The Tasks are targeted to be executed with Voice Client. However, Voice Catalyst has been engineered to be backward compatible, thus allowing the execution of the newer Voice Applications, as well as the Tasks developed with Task Builder.
    • Voice Artisan: the second generation IDE that allows for advanced voice applications to be developed and provides improved development tools, such as emulators, debugger, writing unit tests many other great features. The compiled application from Voice Artisan is referred to as a Voice Application and is compatible only with Voice Catalyst.
    • Voice Direct ERP for SAP: a client Voice Application and Task, as well as a server-side translation engine that allows for the development of voice applications that integrate into SAP via ITS (Internet Transaction Server). It is essentially a screen scraper that takes existing HTML pages and translates them into voice prompts. A fair amount of the development is required and performance is slower due to the HTTP request overheads. Furthermore, it requires for the HTML screens to already exist on SAP, and any unexpected changes in the HTML screens can cause the voice system to break.
    • Voice Express: fairly similar to Voice Direct ERP for SAP, but instead handles the translation of Telnet (VT220, 3270, 5250) screens to voice. A fair amount of the development is required and performance is slower due to the Telnet overheads. Furthermore, it requires for the Telnet screens to already exist on a WMS, and any unexpected changes in the Telnet screens can cause the voice system to break.
    • VIO (Voice Interface Objects): a thin client Voice Application application and server-side SDK for developing server-driven voice solutions. Due to the constant device-to-server communication required in such an implementation, its performance is considerably slower than a thick client solution that can be developed with a middleware server solution such as with the Pigeon Voice Server.
  • Voice applications:
    • Voice Direct: a finished client Voice Application and Task provided by Vocollect that caters for all typical warehouse processes. The source code of Voice Direct to provided to Vocollect VARs and and may be customised as per a customer's requirements. However, customising Voice Direct can be fairly difficult. Furthermore it often requires end customers to make modifications to their WMS and or warehouse processes to work the way the Voice Direct processes have been implemented. Lastly, Voice Direct requires for the WMS to have an existing API for Voice Direct to be integrated out-of-the-box. Therefore, many end customers and Vocollect partners prefer to develop simple custom tailored Voice Applications/Task from scratch. N.B. However, should customers specifically request a Voice Direct solution, the Pigeon Voice Server can be used as the request handler for LUT/ODR messages being sent from the Voice Direct Task/Voice Application on the device.