SIBIC Technology Sharing: Demand and Challenges of Car Voice Interaction

For a driver, the most important aspect of an automobile's intelligence is that it is possible to perform some operations other than driving while driving, and this behavior is mostly dependent on the development of voice technology. At present, companies engaged in speech semantics in China have also noticed the prospects of the automotive field. They have started to implement a speech interaction program in this area. One of them is Si Chichi.


Yesterday, Sci-Chi made a product experience salon in Shenzhen. In their words, it was a “naked technology sharing session”. At the conference, they showed them the 3.0 version of their latest car dialogue operating system AIOS. Through this technology sharing, you also learned about the importance of voice interactive systems for the automotive industry.

The concept of VUI voice interactive interface

In the driving environment of a car, due to safety restrictions, there are not many additional operations the driver can perform while driving. In traditional cars, in addition to driving operations, operations such as answering calls and entertainment are integrated. In the steering wheel, this allows the driver to complete these things without leaving the steering wheel, but with the rise of the concept of "smart car", social networking, smart navigation and other functions have also appeared in the car. At the beginning, these functions were blocked. Into the car, smart rearview mirror and other equipment, but most need the driver to manually operate, which brings a lot of inconvenience to people in the driving state, on this issue, it was found that voice operations become just needed.

When the voice operation just emerged, the algorithm was very simple, the recognition rate was low, and only simple operations could be completed. This can't be called “intelligence” at all. At the site of SBCI Technical Analysis Salon, the product controller of SBCI. Zhang Yan mentioned the concept of a VUI at the scene, that is, the Voice User Interface. The concept is to put the two-dimensional screen operation interface into an operation organization made up of voice systems.

In the automotive field, user operations from the initial interfaceless to the current GUI (Graphical User Interface) is moving in the direction of VUI, which is also the trend, with the most effective voice guidance to help users complete the operation, this will not only ensure the driving Security, significant savings in labor costs, and the ability to free users from tedious, cumbersome driving activities. In addition to the VUI concept, this VUI concept can also be used in voice interactions with smart homes, smart robots, and other fields.

VUI's needs and challenges in the automotive industry

Since it is an interactive interface, it is necessary to ensure that the operation is accurate, convenient, and efficient. When people issue voice commands, they must ensure that the machine can quickly feedback and execute. This is particularly important in the on-board system, because the driving process, the intelligence The whole operation process of the on-board system needs voice to realize. At present, the demand is probably in the following points.

Quick interaction in navigation:

Navigation is an absolute core function for in-vehicle systems. However, most navigations have not yet achieved voice control, or only simple functions such as “voice search purposes”. In the VUI, the driver said: “I want to To go to XXX places, the machine should immediately react and plan the route, tell the driver distance, estimated time and other information. Throughout the interaction, the driver can also describe the appeal, such as "evading congestion," "shortest distance," etc. After saying the appeal, the machine will adjust the route plan accordingly.

Interrupted by cross-cutting areas:

For the “talk with the machine”, the most annoying place is undoubtedly the machine is stupid and embarrassed. For example, if you ask it a thing, sometimes it will say a lot of words, in fact, in the process. In the middle, the driver may hear half to understand what the machine is saying, but usually, if the machine is not finished, it is not possible to perform the next task, listening to an icy machine with a big noise The paragraph is indeed very impatient, especially for old drivers who have road rage and are not friendly.

Smart Rearview Mirror with Sibite Voice Solution

Sibichi made improvements on this point. In the new version of the dialogue operating system, an interruption function was provided. When communicating with the machine, the driver can interrupt the machine at any time and let him perform other tasks. This interruption can also be found in navigation, music, Voice, WeChat and other functions can be flexibly converted without returning to the main interface. There is no need to wake up again and direct voice control allows the machine to do what you want to do temporarily.

Synthetic tone switching:

In terms of machine voice, there is actually a small detail worth looking at. People who have used Siri should understand that robots speak most of a word and a word. Incoherent machine sounds make people want to sleep. In addition to standard speech, they do some celebrity synthesis sounds, such as Lin Zhiling and Guo Degang. , Can add a lot of interest, think of the activities of the Chi-scene, they also demonstrated for everyone the work done in this area.

Speech on-site synthesis

Social networking features:

The proportion of social operations in the automotive industry is indeed not small, and the social way of calling and texting is certainly not enough for modern online social networking. It is very dangerous for drivers to use social networking applications with smart phones while driving. Things, it is particularly important to integrate social networking applications into the on-board system and use voice to control all of them. Lei Feng network (search "Lei Feng network" public concern) in the Si Bichi technology sharing meeting, saw them integrated the WeChat to the vehicle system in the demonstration, the driver can use the voice to wake up the WeChat, and explain to Whoever sends, sends what, sends speech or text (text is converted to text, does not require the user to type manually), even when chatting to want to see, as long as the other party to send you a specific location, the system will automatically transfer to the map Navigation is very convenient.

In addition to the above-mentioned just-needed and fresh functions, practical basic functions such as voice interactions and traffic inquiries for entertainment such as music stations need to be implemented in the VUI. In other words, the car voice operation is more intelligent and the VUI has to be completely like a GUI. Mature.

Voice interaction is indeed a tool that can enhance people's operating experience. Before the advent of autonomous vehicles, voice interaction is absolutely an artifact that can improve the operating efficiency and driver safety factor in the car, but it is all in the entire voice system. Under the premise of high precision and high efficiency, this area is currently only in a relatively mature stage, and there is still much room for improvement. It is believed that in all areas of the future, especially where the vehicle needs voice assistance, VUI will certainly replace manual operation to some extent.

Multi-Pin Series Glass Sealed Connectors

Multi-Pin Series Glass Sealed Connectors,Gold Plate Metal Glass Enclosure,Gtms Multipins Enclosure Terminal,Multi-Pins Feedthrough For Sensor

Shenzhen Capitol Micro-Electronics Co.,LTD , https://www.capitolgtms.com