Our Idea:

Ambient Intelligent Spaces for the many people

A framework for smart environments with freedom and privacy in mind.

As an open tool, Aiiot enables users to train their own intelligent environment. It consists of smart sensors and can be extended with personal, selfmade and existing IoT devices. Using an easy to understand node-interface, different flows and actions can be created — turning spaces and homes into sophisticated smart environments in just minutes.

The aim is to change the passive consumer, who must trust that their sensitive data will not be misused, to an active creator of his environment in a playful way.

Here we have probably forgotten an Image Description, please let us know.

End these cloudy times. It’s time for the sun!

Amazon and Google have found their way into our living rooms. The ubiquitous collection of data in our most private spheres is an unpleasant thought for many people — and yet Alexa & Co. offer a comfort you wouldn't want to miss in a modern home. Isn't there another way? Without data streaming to the outside and independent of the large data centers of intransparent corporations?

As Google shares its vision to create a centralized Internet of Things, we believe that a safe and non-dystopian future can only be established by user-owned concepts like autonomous peer-to-peer networks. Already today we can see the effects of huge cloud based services, that claim to improve our lives but instead enable and advance surveillance capitalism. With the coming age of ubiquitous computing and connected devices that are deeply integrated into our personal lives, we cannot give up our freedom into the hands of a few mega-corporations.

Here we have probably forgotten an Image Description, please let us know.
Many intelligent products today send data to the cloud and are dependent on it.

Another way of doing things.

Aiiot is a framework that enables users to train their own intelligent environment. Instead of streaming private data to huge data-collection centers around the world, everything stays inside your home. You are able to connect all your favorite Internet of Things devices with smart, machine-learning powered sensors. Shape your spaces to fit your needs perfectly.

Let us introduce our lovely Mates.

The foundation of Aiiot are the Mates. Each mate can help you make your environment smarter: Maybe you want to see when your cat is outside the window? Just use a cameraMate to recognize is. Learning a new thing never was easier: Just take a few examples of situations you want to use as input (e.g. specific sounds, outside weather or whenever your fridge is getting empty). The on-device machine-learning algorithm will learn from your input and predict what is happening.

Here we have probably forgotten an Image Description, please let us know.
The different layers of a Mate. Having everything on-board, makes a connection to a cloud unneccessary.

Training on device. No cloud needed.

Discretion is your mates first nature. To keep primary data (e.g. images or voice-snippets) safe, every piece of data never leaves the device.

[0,1,0,0,0]? Your data stays on the device.

Only predictions (e.g "there is a cat in front of your window" in form of [0,1,0,0]) gets communicated to your local homestation.

Pull the plug or make them (really) sleep.

Sometimes there are situations where you want to be 100% sure you have your privacy. Every Mate can have its sense disabled with a physical switch — no digital backdoors.

Open Source and Open Relationships.

The communication runs via MQTT so that self-built Mates and own sensors can be added to the network. In addition, aiiot has a NodeRED layer to allow even more connections. We remain open for new friends.

CameraMate

The cameraMate is able to see and interpret many situations: Maybe you want to check if you left your door open? Or turn on the reading light when you sit down with a book? Just show the cameraMate a few examples!

Here we have probably forgotten an Image Description, please let us know.

AudioMate

The audioMate can trigger actions by just listening to your environment. While listening, no data gets permanently stored — it just tells you when it recognizes the sounds you taught it. You have music playing in the living room? Maybe change the lights so something more relaxed.

Here we have probably forgotten an Image Description, please let us know.

RemoteMate

Sustainability also means being able to integrate existing systems. You can teach the RemoteMate functions of infrared or MHz remote controls so that it can perform them. The RemoteMate makes older devices look a bit smarter.

Here we have probably forgotten an Image Description, please let us know.

Every Mate you'll need

Camera and Audio Input is just the starting point. Aiiot can be extended with a huge amount of sensors and you can also easily connect and build your own Mates. Aiiot is NodeRED & MQTT compatible and opens the door for many more sensors and use cases.

Here we have probably forgotten an Image Description, please let us know.
Here we have probably forgotten an Image Description, please let us know.
The system behind Aiiot: Real-World data gets connected to IoT Devices with your own flows.

Your home. Your Homestation.

Instead of having to use multiple apps and web interfaces, Aiiot uses a central homestation that connects the mates to your IoT Devices. It runs on a local device, that can be, but doesn't have to be, connected to the internet.

Here we have probably forgotten an Image Description, please let us know.

Easy-Peasy. You control the flow.

Building a smart environment could not be easier. Just connect the triggers (things that your mates or sensors are trained on) to any connected device. You can even use basic logic to design complex flows.

Status Quo: Proof of Concept. Check.

We already build functional prototypes of the homestation, the cameraMate and the audioMate. Now it is time to get back to the concept board. Aiiot claims to be for the many people — and being able to be used easily by anyone is our most basic requirement. There are still many things to be decided and improved on.

Here we have probably forgotten an Image Description, please let us know.
Machine Learning can already be used on very basic hardware.

Figuring out what is really needed.

The next step is to invested more time in inclusive and diverse design and usecases. We also need to unterstand more and test what people want to build and which applications are really needed.

Trustable and Small Tech.

We try to build the product around the guidelines of the Trustable Technology Mark and the Small Technology Foundation. But we’re still at the beginning and can try even harder to make Aiiot a model product for these two approaches to a better future.


The beginning of a journey

We still have a long road ahead of us. As of now, we got several working proofs of concept. In the near future, we want to release the first stable open source version, as well as fine-tuning the design concept.

Here we have probably forgotten an Image Description, please let us know.

Plug & Play Solution

To also appeal to people who don't have time or believe they don't have DIY capabilities, we want to offer a Plug & Play product solution that simply works and doesn't need any further explanation. We explore different product designs to give as many different users as possible the opportunity to create their own intelligent environments.

Maker & DIY Solution

We want Aiiot to be an open system, that can be extended to your own need. That's why we are looking forward to release a first open source build of the system. In the future, we want to work on open source hardware that can be printed or build by the users themselves.

One does not simply design privacy.

With the IoT we may get a chance to build it better than the actual Internet — owned and controlled by the many, not the few. Our vision is getting an inclusive, meaningful and personal framework out there to homes of lovely humans… so they don’t need to buy surveillance products from Google, Amazon, Facebook and Co. We have no significant funding and even less time, but we have an idea how the future should feel like and we want to experience it.

Here we have probably forgotten an Image Description, please let us know.

About this Project.

This project was developed by Philipp Kaltofen and Maximilian Brandl as part of a free elective on Machine Learning in the course of studies Interactive Media Design at Darmstadt University of Applied Sciences. We are continuing to extend and improve on it in our free time.

A special thanks goes to Prof. Claudius Coenen, Prof. Andrea Krajewski, Prof. Garrit Schaap, Prof. Tsunemitsu Tanaka who are always there for us with help and advice.

Here we have probably forgotten an Image Description, please let us know.

Get in touch:

hello@aiiot.space