Shulou(Shulou.com)06/02 Report--

Since the popularity of Amazon Echo, more and more people like smart speakers.

Including the domestic Tmall spirit and Xiaomi love, there are also a lot of people.

But the ensuing privacy leak comes again, and the speakers are always listening. What if they accidentally spread the private conversation?

In May 2018, local media KIRO7 in Seattle said that Amazon Echo, placed indoors in a home in Portland, Oregon, recorded a user's conversation at home and sent the conversation to a contact in the user's phone book.

Although Amazon and Google have always stressed that their smart speakers do not collect user conversations, only to monitor specific wake-up words for standby, there is no guarantee that a round-the-clock microphone will not mistakenly record private information.

So the major manufacturers add hardware switches to the speakers one after another, but this back and forth switch is always forgotten. Not to mention that physical contact is not convenient when washing clothes, cooking and going to the toilet.

Is there a better way?

Let's take a look at Project Alias.

The purpose of this project is to add a "sound switch" to the smart speaker without physical contact.

Principle: keep playing "white noise" from the loudspeaker to prevent the microphone from listening. Until the user says a wake-up call, Project Alias will stop the constant noise and allow the user to interact with the smart speaker.

Currently, Project Alias is only available for Google Home and Amazon Echo smart speakers, but the project team has released the relevant design documents and code, according to which creators can tailor similar "small mushrooms" to other brands of smart speakers.

Download address of related design files:

Https://www.instructables.com/id/Project-Alias/

Related code (welcome to participate and maintain the project):

Https://github.com/bjoernkarmann/project_alias

Hardware inventory:

1. Raspberry pie 3A +

2. ReSpeaker dual microphone expansion card

3. Two 16mm small speakers

4. JST 2.0 connector or old Jack connector

5, 4 screws

6. Some wires

7. "mushroom" shell

Production process:

1. First of all, you should have a raspberry pie. After installing the Raspbian system, enable SSH and connect to the Internet.

2. Install ReSpeaker microphone driver, TensorFlow, Keras and other dependent software such as Flask, Flask-SocketIO, python_speech_features, etc.

3. After setting up all the software environment and deploying the program, open port 5050 of raspberry pie with a mobile browser, call it by name 4-6 times to train it, and this "little mushroom" will recognize your voice.

Project Alias was once called "parasite" by its founders, but I think it is more like a "bodyguard", adding a sense of security to our smart lives.

Friends who are interested in the products and technologies mentioned in this article can follow Weibo or Wechat's official account "IoT outpost" and contact me.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.