laitimes

Year of the Tiger Q&A| voice assistant, when will it really understand you?

The morning of the seventh day of the New Year may be even more terrifying than the three sisters and six aunts who constantly greet you about your marriage during the holidays, because you may have forgotten to get up and set your alarm clock. If you continue the habits of the previous seven days and sleep until you wake up naturally, then you will find that you have received various messages on your phone to urge you to work.

But in any case, the work of the New Year has begun, and I wish you all a smooth start!

As for the fact that I forgot to get up and set the alarm clock, I can actually share my morning with you.

Can you leave the voice assistant now?

Every night after lying in bed, I would let the little love speaker on the bedside table set the alarm for the next day. If I had to wake up early the next day or had something particularly important, I would set two alarms. Even if there is nothing to do in the morning, I will set an alarm clock for about the same time.

First, I can avoid sleeping directly until noon, and second, I have formed the habit of thinking about what time the next day is and what is going on every night. So forget to set the alarm clock, in my this is not there, it is recommended that you can go to buy a smart speaker as an alarm clock to try.

Year of the Tiger Q&A| voice assistant, when will it really understand you?

And compared with the mobile phone alarm clock, the wake up ringtone played by the little love stereo can be done every day without repeating, so I do not have a dislike for a music or ringtone.

Since the person is awake, it is time to go to the washing session.

Walking over to the washstand, I'll let the small stereo in the living room broadcast the weather and news for today. Although most of the time the sound of the water flowing will be less than the sound of the sound, but I can still get some useful information.

Finally it's time to go out.

Get in the car, press the start button, and it's time for some music to accompany me on my next hour-long journey of traffic jams. At this time, I will play songs or select a station through the iDrive7 voice assistant or CarPlay-connected Siri.

In this way, I basically use three or four voice assistants every morning. Of course, sometimes a fifth type is added, which is the voice assistant of the TV.

Although sometimes it can be annoying, after all, there are so many voice assistants, and each has its own wake-up word. But after habituation, their existence can indeed improve the convenience of life, and many things can be solved with words.

For me, it seems like I can't do without a voice assistant.

Therefore, the protagonist of this Year of the Tiger Q&A is those voice assistants who have unconsciously entered thousands of households, especially the on-board voice assistants installed in the smart cockpit.

Looking back on 2021, I experienced almost 30 to 50 kinds of in-car voice assistants. As for the performance, it is certainly uneven, some can indeed improve the car experience, and some may make you more and more upset. And the experience of the voice assistant is not too directly related to the positioning and pricing of the car itself.

Year of the Tiger Q&A| voice assistant, when will it really understand you?

My experience is that if the voice assistant chooses a Chinese technology company such as iFLYTEK, Baidu, Tencent, Alibaba and so on, then the experience will definitely be above the qualified line. If the car machine system copied the overseas version of the model, just a Chinese voice package, it may be catastrophic to use.

However, since it is a Q&A, we have chosen a company that is the most representative company in the field of in-vehicle voice assistants, which is iFLYTEK at present.

Year of the Tiger Q&A| voice assistant, when will it really understand you?

Although in-vehicle voice assistants only occupy a small part of iFLYTEK's business segment, the loading volume of more than 7 million vehicles in 2021 is already very telling. At present, iFLYTEK is the leader of in-vehicle voice assistants in the Chinese market.

So as an industry leader, in 2021, iFLYTEK's in-vehicle voice assistant has made breakthroughs in what aspects, and what problems still exist?

What are the improvements for the car voice assistant?

For iFLYTEK's in-vehicle voice assistant, 2021 may be a process of continuous improvement. Especially for domestic consumers, iFLYTEK has done to continue to improve the understanding and thinking ability of in-vehicle voice assistants.

As you can recall, the functions we asked for in-car voice assistants in the past may be to guide a navigation, find a song or make a phone call, which is basically limited to the function of the car machine. But now we also need it to control more functions, such as air conditioning, seats, car windows, and so on.

Year of the Tiger Q&A| voice assistant, when will it really understand you?

And many times you don't need to give a complete command, such as "adjust the air conditioning temperature to 25 ° C", just say that you are "a little cold", and the voice assistant can understand what you mean. With the increase of voice assistants' control over the car body, more scenarios that require voice assistants to understand and think will become more and more numerous.

So in 2021, what iFLYTEK has done is to make the car assistant better understand your words.

In addition, for overseas consumers, iFLYTEK has also launched an international version of the voice assistant. As more and more Chinese brands go abroad, the smart cockpits on these models also need to adapt to the habits of overseas consumers.

As a number of world champions in the field of artificial intelligence, iFLYTEK naturally has the ability to adapt the car voice assistant to different overseas markets, so that consumers in more countries and regions can experience a more intelligent car voice assistant.

What breakthroughs will there be in 2022?

As for 2022, the biggest breakthrough of iFLYTEK is not actually in the voice assistant, but in the cockpit multi-mode perception system with richer interaction.

Functionally, the multi-mode perception system is to have a certain perception of the special behavior or emotion of the main and co-drivers through the camera, microphone and other devices in the car, so that the vehicle can achieve some special functions. The simple understanding is that the intelligent cockpit with multimodal perception system has an active and humanized "emotional" expression and behavior for the people in the car.

If I remember correctly, in mid-2021, I participated in the test drive of the GAC Trumpchi GS4 PLUS and experienced a similar function on GAC Trumpchi's ADiGO 4.0.

Year of the Tiger Q&A| voice assistant, when will it really understand you?

For example, ADiGO 4.0 currently implements features where if you're on a phone call, the vehicle automatically lowers the media volume, allowing you to get a better calling environment. If you are smoking, the vehicle will also automatically turn on the air conditioning outreach, and open the corresponding side of the window to ensure the air quality in the car.

If you show sad emotions in your interaction with your voice assistant, the car machine will choose happy music when playing music.

Although it is not clear whether this function on ADiGO 4.0 comes from iFLYTEK technology, in 2022, in new models such as WM M7, Volvo 90 series and Polar Star 3, more intelligent multi-mode perception systems will be officially mass-produced.

So this year, the interaction methods of the smart cockpit will not only be richer, but also more natural.

The unsolved difficulties of the car voice assistant?

For iFLYTEK, although accurate identification and execution is no longer a problem, how to accurately read the instructions of the driver and occupants in a complex carriage and then interact in a more natural and realistic way is still a difficult problem.

Because for the voice assistant at this stage, you must ensure a relatively quiet cabin environment when using it, and then you can interact with the voice. But in daily life, you are driving in the main driver, and the co-driver is chatting with the rear seat, and the use of voice assistants at this time will indeed cause a lot of trouble.

Year of the Tiger Q&A| voice assistant, when will it really understand you?

It also reminds me of the TNT that Lao Luo released a few years ago, and it looks really good to write in voice, but who can guarantee that no one else in the office is talking? Of course, if you are the boss, you can sit in your own private office and use TNT, not afraid of being disturbed.

But as a boss, do you need to write all day long?

Of course, with the advent of multi-mode perception systems, it can also optimize the recognition effect of voice assistants under interference. Through the camera's perception of lip and eye movements, the voice assistant can correctly identify the interlocutor, or even the content of the conversation, reducing the probability of misidentification.

In addition, iFLYTEK has also launched a new generation of speech synthesis engine, which can be closer to the real person during voice interaction and behave more naturally, thus creating a more realistic interactive experience.

Will virtual digital people become the car voice assistants of the future?

Throughout 2021, with the explosive meta-universe, virtual digital people have also entered everyone's vision, even life.

In a sense, I feel that virtual digital people are the most reasonable state of voice interaction. So in the future, can virtual digital people become car voice assistants?

iFLYTEK's answer to me is that virtual digital interactions will become a trend, not necessarily humanoid, but also may be elves or elemental balls. In fact, what form is not important, the key is the ability and emotional design of interacting with people.

Year of the Tiger Q&A| voice assistant, when will it really understand you?

At present, the vast majority of smart cockpits with voice images basically choose simple graphics or cartoon images, of course, there are independent screens like NOMI to increase the experience of voice interaction. From an experiential point of view, the existence of these voice images can indeed bring a more affinity interactive experience.

Therefore, with the development of virtual digital interaction, these already existing voice images may be given more interactive functions, resulting in a better voice interaction experience.

Year of the Tiger Q&A| voice assistant, when will it really understand you?

In addition, through the design of some concept cars in the past two years, the future cockpit will definitely develop in the direction of simplification. So I wonder if I can project the image of a virtual digital person in any area of the carriage through holographic projection, and become an on-board voice assistant in the new era?

Write at the end

In fact, when I first started using the car voice assistant, my first thought was, "When will they understand what I'm saying?" "By 2021, this problem has basically been solved, and the voice assistants of the vast majority of new cars have done it now, and of course they can do better."

But when will they really understand us? This may be the next problem for the car voice assistant.

So how long do you think it will take to see the rapid development of artificial intelligence technology today?

Read on