I have Network Identification enabled because of COURSE I want to give Sense as many clues as possible to detect things in my home.
I have a large number of Z-Wave switches and outlets in my house, some of which are controlling large loads (like a home exhaust fan). It seems to me that if Sense was able to listen in on the on/off signals sent on Z-Wave network, it might be able to associate those with load increases/decreases.
If Sense sold a Z-Wave add-on—that sat on the Z-Wave network and bridged information over Ethernet or Wi-Fi to the Sense monitor—I would pay money (up to $50) to improve Sense’s identification of loads.
I would absolutely hop on an integration to my current z-wave network (which is run/managed by HomeSeer). I don’t care about integrating with Homekit or Google Home even though they are probably more market friendly from a PR view.
I suspect that almost everyone who’d use this zwave doohickey is running some kind of expandable home automation system - home-assistant.io in my case. Instead of custom hardware, what if we could just publish home automation events at the sense? A unicast UDP packet with the zwave node id & on or off would probably be useful.
This wouldn’t be helpful for aggregate machine learning like whatever network traffic is coming from a tesla charger, but it’d be a very strong signal for just my house. For instance, my dehumidifier gets switched on when all the windows are shut, and sense fails to identify its first cycle (from powered off to on) and does fine afterward (from idle to on).
If Sense could talk to Home Assistant (or other hub/controller), there’s all sorts of info that could help Sense tag/name devices. It wouldn’t need to be Z-wave capable (or Zigbee or MQTT or RF or…), just compatible with the hub. I’d be glad to work with folks on this.
Same here with SmartThings as my automation hub. My understanding is that SmartThings API is fairly accessible. I can see the current watts used on all kinds of zwave/zigbee devices in my home. I get that Sense is getting data at a much faster rate but is sure seems like this would be incredibly helpful for learning already named devices and have a starting point for learning when those devices are on or off with a pretty good idea of the current wattage. A great way to add to the machine learning knowledge base.
I have the following SmartThings devices that return a wattage number. I can view them in the ActionTiles PC Web display or on my SmartThings phone APP. I would prefer that this data be integrated into the SENSE data so that I have a whole house view of my wattage in one APP.
SmartThings devices that I have installed in my home that monitor wattage:
A display of my ActionTiles (linked to the SmartThings Hub) with wattage information is shown below. Please consider pulling wattage data from these devices (through the SmartThings Hub) into the SENSE APP. Thanks.
I too am a SmartThings user with some devices that collect power usage data. I have a feeling this is a little too niche of a scenario for the Sense team to devote resources to, but I’ll add my datapoint nevertheless.
Samsung SmartThings is a pretty big platform. I have 50+ devices running on it. Most of them Light Bulbs actually as it supports Sengled light bulbs which happen to be pretty good lightbulbs for a good price (8 pack for $49.95). All of these have Wattage information, so could very well be used by Sense.
So, it’s possible to leverage sense’s hue integration via ha-bridge which emulates a hue hub. It’s not exactly easy to setup, but certainly possible. I currently have 60+ emulated bulbs from various zwave switches/ zigbee bulbs. This is all being driven from home assistant -> ha-bridge -> sense.
You have to use an old version of sense app as well as an old version of habridge bridge to pair, but once paired everything can be upgraded.
This has some downsides. One: you can really only emulate a 10watt bulbs, but considering most of my lights just so happen to have 10 watt bulbs behind them, this worked out pretty well. Two: once an emulated bulb is added to sense, you can never get rid of it…
Of course native and supported integration would be best.
My house uses Insteon devices. The majority of which communicate over RF and powerline. I wonder if Sense could read these communications? We power our irrigation pump using an Insteon switch. Our landscape lighting runs off Insteon switched as well. I imagine the address for the Insteon device is embedded in the communication and could be used by me to link power usage with specific devices
Sense only has WiFi hardware in it today. It communicates with TP-Link and Wemo smart plugs and the Hue controller via broadcast packets and responses. Sense might be able to talk to other smart home hub/controllers on a home LAN, but that kind integration might not be as helpful as the current ones. Why ?
latency - unless on/off data is timestamped fairly accurately, there can be a substantial (2 sec) delay between a change in device, and the event reaching a device connected to the hub via WiFi (I see that with the Hue).
power data - most smarthome devices data includes only on and off events, not actual power readings. Not as useful as power readings that come from the current smartplugs plus the Hue.
data rate - even if your smarthome device supplies the hub with power readings, supplying them at the rate (2x per second) the current smartplugs are read, wouldn’t likely be possible. Z-Wave, Zigbee, etc. just isn’t up the the challenge. 72Mbs WiFi seems to be challenged when really with 20 or so smartplugs.
I’m not sure what most of these PLC frames look like, but I wonder if Sense’s 2MHz voltage sampling rate is sufficient to directly read certain PLC events from the line? That would remove any latency introduced in the various transcoding and processing events to relay the data to the hub through a third party device.
I too am an insteon hub user. I would think there would be a way for sense to monitor the commands being sent over the power lines to know when something it getting turned on/off. This may make it easier to identify devices that have similar patterns.
