The goal was to improve the voice-recognition system for Honda. My role on the team was Human Machine Interface Engineer. I was in charge of user testing and creating the user interface specifications.
My 1st step was to understand the current state of the competition and voice-interface in general. I played around with the systems of other automotive manufacturers, Apple’s Siri, Google’s Assistant, Microsoft’s Cortana, as well as some in-home voice technologies. The goal was to see what’s out there & what’s possible.
This involved defining all of the questions that needed answers. Which voice-interface characteristics did users prefer? Which characteristics were easy to use and which caused usability problems? Of all of the various ways to do things, which were best?
Research was mostly conducted with hands-on tests, but distributed surveys were also used.
Data would typically involve subjective user opinion as well as completion time and error rate. I prepared the materials for the tests, I conducted the tests, and I processed the test data.
The specifications were presented as a rather large flowchart which map the system’s reaction to all the various user inputs and how that reaction would change depending on context.
I also specified what would be displayed to the user at the various steps within the interface. But project hasn’t been released publicly, so I can’t show anything.
Since the interface specifications that I made were a pretty large and complicated, I created a GUI for the benefit of internal stakeholders. Using Axure, this GUI was layered over the specifications so that people could scroll & click around in the specification freely while always understanding what was happening in the task flow.