Every automation system requires 3 elements: sensors, controllers and actuators. A cybernetic house interconnects these in a way that makes both immediate and past readings available, enabling levels of awareness and adaptability that would not otherwise be available.
This picture is Marty with an adapted ultrasonic distance sensor, taken from a video of an appearance on Computer America. Yes, really that much hair loss.
There is, alas, no single approach to organizing our explanation of the overall automation set-up without confusing things. So, since one won't do, we'll use two. We can start here with a practical explanation of how we addressed the broad-strokes problems, then dive a little deeper into each of them in their own sections.
Somewhere between figuring out what we want to know and what we need to know comes the question, how are we going to know it? That gets us into the specific things we may need to measure (temperature, proximity, range, humidity, air pressure, door status, vehicle detection, vehicle identification, gas detection, smoke detection, ambient light, wind speed, soil moisture and so on) and a determination of the best ways to measure them. It brings us into secondary interpretations of primary sensors; for example, range detection may also help us determine speed and direction. It also allows us to make pragmatic choices; for example, with a garage that's 28 feet deep, neither the IR nor ultrasonic range sensors we want to use have enough range to reliably sense whether the garage door is open or whether there's a car in the garage space; only Lidar or narrow-beam radar can work.
Happily, we solved a major, potentially complicated challenge by inventing a ceiling awareness pod (CAP) that we can install everywhere, a Bluetooth Beacon system for detecting and identifying cars with running engines and several other tricky little solutions.
For more about how we deal with sensors, see:
Sensor Loci: The places they go and the tasks they are there to perform
Sensor Elements: Information about individual sensors and sensor groups
Sensor Use: Once you know where and what we will deploy, this explains how we can exploit their information
Some sensor data also comes from devices used for other purposes; for example, the surveillance cameras that ring the house can signal the presence of objects within their fields of view by monitoring the extent of pixel-shifting within an otherwise still image.
Our Products section has information about products we use in every part of this project, including information about specific sensors (as well as controllers and actuators).
No part of this design exercise has been as frustrating as the search for controllers and how to approach them. The home automation category includes hundreds of products but no one protocol there can handle all of the sensor and actuator needs of the house project. Most products are of the wall measles variety, meaning visible little white lumps all over the place. Most require battery babysitting. In many cases, their control languages are unsophisticated, unable to deal with conditions any better nuanced than if-then constructs. Some of them require connections to external servers in order to work. And most of them are too ornery to talk to each other. The Raspberry Pi shown here is part of the solution.
Stitching patches together became a mandate. You may be aware that some vendors use the not-inappropriate term fabric when referring to their sensing and control systems for commercial and some residential automation.
For our stitching, the thread is wired Ethernet, which allows us the fast response times necessary when taking frequent measurements from thousands of readings; none of the wireless protocols has broad enough bandwidth or low-enough latency to accomplish that.
Those Ethernet connections are aided by MQTT (a publisher-broker-subscriber machine-to-machine message handling service) and by a MySQL relational database server. With almost 100 Raspberry Pi 3 Model B single-board computers, we gain highly intelligent and capable controllers that are actually computers with a 64-bit 4-core 1.2GHz CPU in each. This provides a very intelligent endpoint capability for preprocessing and analysis of data, which reduces the amount of data that needs transport across the system. For example, only changes in values (from one perspective, news) need to be posted to the relational database.
For more about how we deal with controllers, see:
Controller Loci: Almost every controller in the system is a Raspberry Pi 3 Model B. Some controllers (like the relational database and the surveillance system NVR, for Network Video Recorder) need higher-horsepower computers. About one third of all nodes are located in a central control rack, about one third in the CAP ceiling pods and about one third deployed near more specialized points of use. .
Controller Elements: We occasionally have to attach a Raspberry Pi to a companion gateway or hub to a specific protocol; for example, a Raspberry Pi cannot directly produce the signal levels needed for a DALI lighting interface.
Controller Use: Some controllers have direct task responsibilities involving sensors and actuators; some have monitoring duties related to keeping an eye on the values being read system-wide; some monitoring duties involve supervisory roles and some have specialized roles.
Our ceiling awareness pods (the house will have 30 of these) each provide, among other things, an 8 by 8 array of analog temperature measurements with analysis of both background and warm-body temperatures and positions, interface to sensors in light switches and doors and selected power outlets, fan blade proximity sensors, infrared remote control signal emitters, an amplified small speaker and more.
Awareness, even with intelligence, can only result in automation when they control something that can effect changes in the real world; in the design of androids, that's called the manipulative imperative. In a house, actuators are how automation can turn lights or fans on or off, open or close a garage door or an electronic deadbolt, phone the fire department, push a display to a door panel, signal when you've pulled far enough into the garage and more.
The information you'll find here:
Actuator Loci: Like sensors, actuators are endpoint devices; like sensors, many actuators will essentially be extensions of a near-endpoint Raspberry Pi or an adjunctive control bus, like DALI for lighting. Every light, every ceiling fan, every deadbolt, the lawn irrigation system and many other household locations are eligible.
Actuator Elements: Traditional magneto-electrical relays can control things of almost any voltage or power range and remain reliable for a very long time, so we may call those into use in some places, and a high-current variation called a contactor can throw the switch on very high-power loads. Our overall favorite for controlling external devices is the high-side switch, a simple circuit with a small parts count involving mostly transistors and resistors.
Actuator Use: We choose to use actuators where we need to and want to use them, not just because we can use them. They can make sure, for example, that garage doors open and close only but always when appropriate. They can lock the doors against bad guys but make sure they're unlocked for first responders in the event of a fire. This section runs down the net effect of our networked effectors.
All of these devices require power, and we will provide the specific needs for each. A conduit-fed 4-wire DC power feed throughout the house provides stabilized (and battery-backup, surge-protected) 3,3, 5 and 12-Volt sources everywhere, and we hack into devices to hardwire them to our DC supplies so there is no battery babysitting required. Most devices are mounted and connected via standard electrical boxes installed in ceilings or, on occasion, walls.
© Copyright 2016, 2017 and 2018 Newstips, Lord Martin Winston and J2J Corporation; all rights reserved
This site needs a larger screen and is not optimized for phone or tablet viewing