Next Generation Domoticz - What lies ahead....

[thebeetleuk]

Hi All,

I have been on the home automation front for quite a while now (which reminds me I must update my signature with all the other things I have purchased over the last year!). Anyway I have been thinking about what might be coming next. I do love domoticz and I even have my family using it which as most of you will probably know, is the real test of a working solution. With countless switches, monitoring, heating control and so on... I'm starting to hit a bit of a roadblock. What next? So I can turn things on, turn them off, schedule them, write some scripts to try and be a bit smarter. My main problem seems to be I just can’t find the time to go and write the code needed to get the next step. I mean, every time I go into Maplin (on the rare occasion) I seem to see yet another brand of home automation. The only things is they are all doing the same stuff.

So what is that next step? Well to me it's having something that is starting to be much more intelligent, learns your habits, knows when you are getting up or coming home, do I really have to use a switch. Why can't it just know what I want?

I know this seems a bit far-fetched but this has to be where we are heading. If you think about Nest, it does try to say that it learns your habits. Now I can't say, as I don't have one, but seriously, how good can it be? Does it really know when I'm coming home, on holiday, out for the day?

So I'm thinking, what's the plan for Domoticz? What is the 1-2-5 year vision? From my own perspective we have lots and lots of good stuff here. The only thing is that I think we should be thinking of building on more core/flexible functions to the product that make it much better. Things that immediately spring to mind are heating control, presence detection, holiday functions, speech, voice recognition, facial recognition (I know lots of us have written scripts in these areas). As an example can't we build a generic presence detection function that takes whatever inputs you have (door sensors, phones -wifi, bluethooth, PIR, light, calendar, heat, and so on...) and then calculate if you are home, what room you are in? So instead of 50 scripts, we end up with one that we all use?

The second step would be to start combining these functions and letting the system start to learn? I'm no expert in this but I know it can be done. Again this comes back to the vision.

I really do want to make things better as I can’t stand the fact that one of these big companies will come along and do it first. Oh and it will cost a fortune!

If you have made it this far then I a must have said something of interest so thanks for reading

On so that's by brain dump over. I would be really interested if any of you have thoughts along these lines.

Martin

P.s. Thanks to everyone here who contributes to the home automation scene and a special thanks goes out to the people who run this forum as I know it must take time that not al of us have.

[asjmcguire]

The learning idea sounds great in theory - but the reason all the big companies are able to do this - is because they have the computational power needed to learn patterns for hundreds of thousands of users. However - the simpler parts of what you want are do-able. The Domoticz Android app for example has Geofences - so the app can tell Domoticz when you leave the Geofence and when you arrive back - you can use this to switch a virtual switch (mine is called v_AtHome) and using blockly you can can turn stuff on and off directly when this switch changes, you can also refer to the status of this switch in other blockly events - eg when the temperature drops below setpoint, check if v_AtHome is on before turning the heating on, otherwise don't turn it on. I also use door contacts and PIRs to switch a virtual switch (CafeSwitch) which uses Blockly to switch an actual device (Cafe Light) - so I enter a room, (the door contact and PIR both turn on the switch CafeSwitch) a light (Cafe Light) turns on - there is an off timer of 900 seconds - so with no further input the light will turn off again after 15 minutes, but Domoticz is clever in that each time the door contact or PIR is activated again it turns the virtual switch (which is already on of course) on again - which makes no change to anything else - but it restarts the internal timer Domoticz uses back to 0 so the 900 second countdown starts again, this way - provided there is movement in the room, the light will stay on and will only turn off 900 seconds after movement stops. It would however be nice to see the Android app have bluetooth support and then I can buy the Eddystone bluetooth things and stick them in rooms so that when the Android app detects the Eddystone UUID it can turn on a switch.

Note - the 900 second timer is attached to CafeSwitch not CafeLight - this way I can manually turn on CafeLight if I want it to be on during the day for example - the CafeSwitch is set in Blockly to ignore switching requests when the Daylight sensor says it is Daytime.

[ricorico94]

Hi,

I'm still a new user of Domoticz as I already started last November.. However, I became a fan and even my family finds some features quite great !
However, the best and also the most frustrating thing I face is the voice recognition.
I installed the S.A.R.A.H module on another device (Win10 PIPO with a KInect) to run this third party module and I really had a WAOUH effect on family and friends as I can now order my Domoticz to open blinds, switch on TV, start my robot-cleaning, heat the beds, etc. But it's also frustrating both due to unrecognized orders (I'm afraid the Windows based speech recognition of SARAH is less efficient than Google's.. and it becomes aeasily confused if TV is running due to ambiant noise from TV) and due to physical restriction (only one room where I can give orders). So, as in EUREKA movie, if Domoticz could really implement some reliable speech recognition (and possibility to extend with more microphone devices), it would sound great and really improve the WAF factor.

The other point could be in ergonomy: maybe I'll state here a wrong feeling as I'm still new, but I have the feeling that GUI becomes heavy. Meaning that GUI is very good for administering the platform, but when trying to visit the website from remote, the applet is long to download and often fails starting. Accessing the floorplans from my PIPO tablet forces to use both heavy browser (like firefox) and the web applet, so it's slow. I understood there's possibility to create custom webpage, but it's still using same applet, right ? So maybe one future direction of Domoticz could be to propose 2 environments, one for full administration, and one for 'light access'.

Regarding the "NEST" like "learning habits", I'm always suspicious. As in previous posts, they often base learning on predefined analysis of mass behaviors.. but I'm not sure (or don't want to admit) I live and behave like everyone else.. (at least , my heating scheme doesn't seem matching all mass consumption schedules). I believe there will be always two populations (at least). People looking for simplicity and happy with standard intelligence. And people more "geek" willing to take the pain of customizing for own purposes.

Another direction I had thought about was security.. Having a Domoticz possibly more oriented to security (meaning probably duplicating our systems into 2 distinct raspberries, one used for usual Domoticz, the second for security purposes, more isolated). And even maybe imagining how Domoticz could directly discuss with other Domoticz in surroundings (like a mesh/network) to keep capability to send alarms if power and internet are down. There are some technos now growing around (LoRa, Sigfox (with weaknesses, though), etc.) which allow long range signals..

Anyway, huge congratulations !

[asjmcguire]

You can set different access levels for different users, and thus restrict them from being able to do certain things - which removes from the web interface functionality that they don't have the access rights to use. You can also use various Android apps to control the system. As for voice control, rather than rely on Domoticz to do the heavy lifting, you could get Raspberry Pis and dot them about the house, there is absolutely bound to be scripts out there to do listen for voice commands and process them - and speaking to domoticz to actually do the control side of things is really really easy - it's just a HTTP call.

A quick look around this very forum gets us - Jasper Voice control which will run on a Raspberry Pi - http://jasperproject.github.io/documentation/ and this script http://domoticz.com/forum/viewtopic.php?f=21&t=10475 which will work with Jasper to control Domoticz.

[G4HDU]

Hi All,

My main problem seems to be I just can’t find the time to go and write the code needed to get the next step. I mean, every time I go into Maplin (on the rare occasion) I seem to see yet another brand of home automation. The only things is they are all doing the same stuff.

I know the feeling. I've just started to get into this home automation lark and the plethora of different systems available has been confusing to say the least. I have an rfxcom 433 MHz box which I have controlling some LightwaveRF sockets, some blind openers, some inexpensive Maplin plugin sockets all working happily together. Why LightwaveRF? Because the sockets look nice. A purely aesthetic reason and certainly not a technical one. One of the main reasons I went with Domoticz is its ability to integrate disparate systems; for serious stuff I will go down the z-wave path (it is less prone to interference from my 70cms (433 MHz) ham radio transmissions plus it is bidirectional.)

For the past 10 years I have been writing plugins for a php based cms. There is, from the users perspective, a simple method of installing and uninstalling these plugins which makes it very user friendly. A plugin author writes his scripts to conform to certain requirements so that they will integrate with other plugins and the main cms (in this case e107). Is there a demand for such a thing for Domoticz? For me, as a newbie, I would like to look through a repository of devices added by third parties and add for example an esp8266 rgb led controller because I have just bought one. To go to an app store and install it would make my life simpler, OK I can code it myself but not everyone is capable. The ability to contribute to the e107 community by developing some 40+ plugins from lifeboat launch status to a video gallery gave me a happy feeling (OK a smug grin).

So, after all that, is this a likely scenario for Domoticz?