Today’s connected cars offer access to an enormous amount of content services, but when experiencing the “connected services menu” for the first time it can be frustrating for the driver. Finding the cheapest station offering Diesel, searching for a Heavy Metal radio station or posting an update on Facebook while in the car requires clicking through endless menus: starting with finding the app menu option, searching for the right app, launching the App and finally entering the search query with a challenging onscreen keyboard.
In some OEM programs, part of the process described above has already been streamlined and speech-enabled. That’s certainly a step in the right direction and I know this kind of user interface quite well from my own connected car. The ultimate next step is certainly to have the content available by simply touching the voice button on the steering wheel which uses Natural Language and features a ‘flat’ menu. “Find me a cheap gas station with Diesel,” “Search for Heavy Metal radio stations,” or “Post to Facebook ‘my connected car is nearly perfect.’” It’s no longer necessary for the driver to find the right app: the speech system understands the intent of query, dispatches the query to the right content source and then presents a multi-modal response to the driver.
What do we need to get there?
For an optimal in-car user experience while searching for content in the car it is key to effectively combine Automatic Speech Recognition (ASR), Natural Language Understanding (NLU), content retrieval and conversational capabilities. A close integration with content service providers (CSPs) enabling access to the data and web API of CSPs is also very important. Nuance has been very engaged in building up a strong content partner ecosystem and embracing leading brands that support automotive-relevant use cases. This helps ensure that the use cases drivers care most about are being provided in a user experience they expect. In addition to the right ecosystem, it’s crucial to combine content and Natural Language Understanding for an optimal result.
Another common issue that needs correcting is that users very often perform searches using expressions for categories, genres or even music titles that do not precisely match the data fields provided by the content service providers. And what’s more, data fields are not even consistent across different CSPs. This is something the driver shouldn’t have to worry about, though – it should instead be accounted for in their holistic in-car user experience. For example, Nuance’s NLU technology transfers those expressions into a format that is optimized for the search capability of the respective CSP.
The representation of the results can be a challenge for some users too though, and certainly affects the user experience. Depending on the particular CSP, the results list for a search can be completely different. It can contain just one exact match or, if partial matches are allowed, a configurable list with more than one hit is also possible. Streamlining the results format and delivery will help the drivers who are viewing the results of their queries. In addition, the search result can and will change over time as new content is added. Dragon Drive takes resolves these issues and provides a stable multi-modal presentation of search results.
At Nuance we make content an integrated and optimized part of the Nuance Dragon Drive solution. The solution guarantees the highest Natural Language standards, availability of the complete service portfolio independent from the driver’s smartphone type and a cost-effective operation of the whole system which includes a full lifecycle management of the content system.
To learn more on this topic, you can view our below whitepaper.