Voice-recognition technology is also helping make it easier for drivers to get to their desired points of interest. As accuracy has improved, voice systems are making a transition from strict commands to more natural-language instructions. This dynamic, point-of-interest functionality lets drivers start by saying “I’m hungry” to start a search of restaurants. The system will automatically begin scanning for nearby options. If desired, drivers can focus in on fast foods, sushi, coffee or other options.
At the same time, voice-recognition specialists are making it easier for drivers to simply ask for an address. Speakers can now say the city, street name and number as a single stream. That makes it much simpler and quicker than pausing after each segment to wait for a system acknowledgement. An Automatic Speech Recognition (ASR) system converts the spoken input into an address. There is no need to train speech-recognition systems anymore, they work completely independent of the user. The navigation software loads the needed speech data on demand. This improves performance and saves memory. The necessary speech data is compiled already within the map data. No extra external data or online connection is required.
While these changes may seem pretty straightforward and desirable to users, they’re quite difficult to implement. Systems must understand various accents even when the vehicle is in a noisy environment, such as heavy rain or on a rough road.
Regardless of background noise levels, the vehicle’s microphone must pick up requests so the system can begin searching. These searches can be complex, especially when the navigation system must search for cities and streets that may have names that sound quite similar. It takes a combination of powerful computers and very sophisticated algorithms to perform all these tasks in a timeframe that’s