Under the microscope: LINGUATRONIC voice control: "Hey Mercedes"
One of the strengths of MBUX is its intelligent voice control with natural language comprehension, which is activated by the words "Hey Mercedes". The new LINGUATRONIC (included in MBUX with extended functions) supports many infotainment functions (e.g. destination input, phone calls, music selection, writing and hearing messages, weather forecast), as well as numerous convenience functions such as climate control/heating/lighting.
Conventional voice control systems in cars call for certain fixed commands from their users. Thanks to natural speech recognition, MBUX's LINGUATRONIC, on the other hand, obeys virtually every command, recognises and understands nearly all sentences from the fields of infotainment and vehicle operation. For example, "Will the sun be shining tomorrow in Amsterdam?" is now understood equally as well as "Will the weather be fine in Amsterdam tomorrow?".
The intelligent language assistance is activated either via a button on the steering wheel or with the command "Hey Mercedes". It's not the human who has to adapt to the machine, but the other way round. Indirect speech is also recognised, for instance if the user says "I am cold" instead of the clear command "Temperature to 24 degrees" in order to operate the climate control.
The voice control is also capable of learning. On the one hand it tunes into the user and their voice and also understands non-native speakers better; on the other hand the software models on the server learn new buzzwords or changing use of language with time. The system also no longer answers stereotypically, but varies in the dialogue output too.
The fundamental way the language assistant operates: the voice entries are freed from background noises, compressed and transmitted. The voice control is a hybrid system. In other words, it uses both onboard software and the cloud to understand speech as well as possible and respond to the user's requests.
Both the computer in the vehicle and the server evaluate the data and send a reply. The system decides which reply is the most likely, then within a few seconds the reply/reaction follows. This means that in contrast to many other assistants, the language assistant also answers even when there is no internet connection.