Project Updates: New Name, Looking Ahead
It’s been an exciting few weeks working out the smaller issues with my automation system, adding new features, and looking ahead towards the Spring Show presentation. I’ve decided on a new name for the project: HouseMate
I even started experimenting with a logo design:
More to come in the branding department.
On the technical side of things, I’ve made some significant improvements to my automation system that so far have accomplished 3 things:
1. Improve the stability of the system
2. Integrate new hardware for new functionality
3. Begin support for mood-based scenes
Starting with stability, I’ve updated the home assistant software to it’s newest version, which thanks to the open source community that supports it, is even more efficient and robust than ever. I’ve also updated the separate component “Homebridge” (which emulates products compatible with Apple’s Homekit platform) to better control the colors of my bulbs. Previously I was handing Siri light adjustments directly within Homebridge, which would cause sync issues with Home Assistant. Now everything syncs with the main server, and glitches that I had previously now are avoided.
I finally fixed my DIY blinds opener, which needed a timing adjustment to avoid the motor from turning the blinds’ wand too far, causing the motor to get stuck. The opener runs off an ESP8266, which is a cheap Arduino alterative with built in WiFi. To adjust the timings I had to re-flash its firmware using the Arduino IDE a few times with different time values until the blinds opened and closed without hiccup. Now that they are in working condition, I was able to create a morning routine that opens the blinds when my Alexa alarm goes off in the morning (in addition to turning my beside lamp on).
Just to give a sense of how a sequence like this works, here’s a brief outline of what the system does each morning:
Set Alexa Alarm -> Alarm goes off, sends command to IFTTT.com -> IFTTT triggers Home Assistant Lamp and Blinds via HTTP Post command -> Blinds communicate with HA via MQTT server/message
As you can see, there are many pieces to get something like this working, but when it works, it’s a quite magical experience.
On the hardware side, I’ve implemented two new devices: A second motion sensor near the front door, and a device called “Leeo” which listens for smoke/CO2 alarms.
Having a second motion sensor setup ads another layer of automated action to the system. I already have a sensor setup on the top of the stairwell, which triggers stair lighting and 1st floor lighting at night. However, when walking up the stairs, the lights would only turn on when the sensor saw someone halfway up the stairs. Now, a sensor above and below ensures that the lights are always on when needed. Since our stairs are near the front door, the lower sensor doubles as a door sensor, which is great for security when everyone is away. I plan on adding extra rules to use this lower sensor to take greater advantage of it, such as triggering lights when arriving home, or even triggering an audio-based greeting tailored specifically to the person who just arrived home (more on that in a future post).
The “Leeo” sensor I picked up from Amazon is a 3 in one device. It’s first function is that of a connected microphone, which specifically listens for the frequency of my smoke alarm. If the alarm were to go off without anyone home, I immediately would receive a virtual phone call and recording of the alarm. Hopefully I’ll never need that feature, but adding connectivity to older hardware such as a smoke alarm makes it all the more useful.
The device also doubles as a temperature/humidity detector, and an RGB nightlight. Using integration with IFTTT.com, I was able to have the device send it’s temperature and humidity readings to Home Assistant. Having in-home climate readings as a part of my server allow for awesome automations, like turning a fan on when it gets too hot, or alerting everyone via SMS to change the thermostat at certain temperatures. Last year living in an apartment, I was able to automate a window fan in this way to turn on and off to keep the room temperature at a certain level.
Looking forward towards the Spring Show, I’ve begun my research into color theory and music/color combinations that I might be able to implement into the system. By the end of this week I plan to have half of the technical dependencies in place to support this idea, including google calendar and fitbit integration. I brainstormed a bit, and came up with a scenario that I’m working towards making a reality:
1. Person arrives home. System knows which person it is and when they walked through the door
2. Person is greeted with custom audio greeting, and sent a phone notification
3. Based on event and step data up until that moment, system recommends 1–2 mood choices that can be chosen via interactive phone notification
4. System adjusts to match preset scene.
A workflow as such might take longer to implement in reality, but ideally for the show my goal is to have a demo of the different scenes, triggered by manual entry of example events/steps etc. Stay tuned for more information on how this comes into fruition!
