These days, being environmentally friendly is becoming an important aspect of most technology. Another aspect is having technology helping us decrease our energy consumption by making us aware of how much energy we consume at any given moment.
Chetty, Tran, and Grinter presented a study at the UbiComp’08 conference of 15 households in a US metropolitan area, studying their current practices when it comes to resource consumption. Interestingly, the study finds that infrastructure for consuming electricity, gas, and water have become a part of the everyday life to an extend that the consumption is mostly invisible to the people using it. However, participants in the study would like to manage their resources, but sometimes feel that they are already doing enough, and often have no incentive to do more. The researchers believe that this can be changed by visualising resource consumption.
Their studies show that if people could be made aware of their current usage patterns, the state of devices and appliances in the home, and perhaps benchmark that to other homes, this would help them conserve water and energy. Furthermore, ubiquitous computing could help by making resource consumption visible, not only for an entire home, but also by make the impact of individuals visible in real-time.
Sources:
Chetty, M., Tran, D., and Grinter, R. E. 2008. Getting to green: understanding resource consumption in the home.
The World Wide Web is about unlimited access to information using the internet. Similarly, the idea behind the World Wide Grid is unlimited access to resources. The resources are typically storage space and processing power. However, while several grid projects exist, so far they have been difficult to use and since they rely on different technology, users would have to choose which grid to use.
The gEclipse project is about easing access to grids for users and developers. The project has developed an extension to the Eclipse development environment allowing easy access. The goal is that every user should be able to access the resouces using just a few mouse clicks. The environment supports the middleware to access different available grids, both research oriented and commercial.
Easy access to resources is especially important for pervasive computing environments. Resource limited devices might send jobs to the grid, accessing the available resources to perform tasks they are not capable of carrying out themselves. Tools to ease the development of such devices means that such scenarios have come closer.
Sources:
ICT Results - Foundations of the World Wide Grid
g-Eclipse Project
Although not that common in Denmark, in other parts of the word, central HVAC (heating, ventilation, and air conditioning) systems are in place in many homes. Patel, Reynolds, and Abowd, have looked into utilising this by measuring pressure variations in the air handler units, and presented the work at the Pervasive 2008 conference. Variations in pressure result from movement in the home, particularly transitions from room to room and when doors are opened and closed. The benefit of the system is that it works with a sensors mounted in a single place in the building, the air filter.
In the concrete experiment, five sensors placed on the air filter where used. Changes in the air pressure across all sensors are used to uniquely identify a particular event with an accuracy of up to 75-80% in preliminary experiments.
The system works regardless of the operation status of the HVAC, but requires a certain amount of changes in air pressure to sense events. Thus, only events like opening and closing doors, and people who fill most of a door frame walking through a door can be sensed. For that reason, it works reasonably well for detecting the movement of adults, while children moving through a door cannot be sensed. Nevertheless, it is an interesting experiment in utilising existing infrastructure for sensing events.
Sources:
Shwetak N. Patel, Matthew S. Reynolds, and Gregory D. Abowd: Detecting Human Movement by Differential Air Pressure Sensing in HVAC System Ductwork: An Exploration in Infrastructure Mediated Sensing, in Pervasive Computing 2008. (http://www.springerlink.com/content/m27744743kh37155/)
To provide context aware services in the home of the future, a prerequisite is to have access to information about the users context. One of the most useful types of context information is information about the users location. (See the komialt pixi-book for more types of context information). Given a user location, a home automation system can, e.g., save energy by turning off lighting when users leave a room or customize the music to the person entering a room.
Outside in the open the GPS system can be used to obtain location information but inside buildings the GPS signals gets distorted and cannot reliably be used for positioning. Therefore there have been a lot of research into indoor positioning systems.
Typically, indoor location systems require that an infrastructure supporting the system is installed. This can, e.g., be WiFi access points, DECT base stations, or Infrared beacons. The downside is that such an infrastructure can be expensive to purchase and install.
In PowerLine positioning, the existing infrastructure for distributing electrical current in a building is used for positioning. Patel et al. presented this idea in 2006 at the UbiComp conference. The system consists of two injectors that injects a signal into the infrastructure at two opposite ends of the building. These signals can be collected throughout the building using a portable device and via a fingerprintring algorithm, the signals are transformed into a location.
In this years UbiComp conference, Stuntebeck et al. presented a refined version of Patels method. In the new version, only a single injector is used inject up to 44 different frequencies and this improves the accuracy and the temporal stability of the system.
Sources:
Wideband powerline positioning for indoor localization, Stuntebeck et al., UbiComp 2008
PowerLine Positioning: A Practical Sub-Room-Level Indoor Location System for Domestic Use, Patel et al., UbiComp 2006
A Canadian company called Ecobee will soon launch a smart thermostat for managing heating and ventilation in the home of the future. According to the company, the thermostat will both enable savings and lessen environmental strain by controlling the heating and ventilation system intelligently.
The device will connect to the Internet through a WiFi connection and store usage information on a personalized web portal from where it is also possible to control the unit. The portal also includes functionality for programming vacation and giving alert messages on the display of the unit.
By connecting the thermostat with a smart energy expansion module, the use of electricity can be included in the system. This makes it possible for the user to continuously monitor the power consumption of the home.
The thermostat has a set of ZigBee expansion slots for integrating with other home automation devices.
According to the company, the device will pay for itself in energy savings within 12 to 18 months.
Sources:
EcoBee
Engadget
By many, light emitting diodes or LEDs are seen as the future in lighting. With low power consumption and high quality light, LEDs are suited for lighting in industry and homes. If adopted on a world-wide basis, the use of LEDs can reduce the global power consumption by as much as 10%.
Researchers at Boston University’s College of Engineering are currently trying to use LEDs for an additional purpose — wireless communication. If realized, the research will make it possible for light bulbs to act as access points to communication networks in homes and industry.
The basic principle of operation is to exploit that LEDs can be turned on and off very rapidly - so fast that it is imperceptible by the human eye. Data is encoded in the pattern of LED blinking. The researchers estimate that a bandwidth of 1 to 10 megabits per second can be achieved.
Using LEDs for communication instead of RF technology has a set of benefits:
- Security. Since light cannot penetrate walls, it is not possible to eavesdrop on signals outside a building.
- Interference. The signal is not affected by RF-equipment. Natural light sources will, however, slow down transmission speed
- Production cost. Researchers estimate production costs of adding communication capabilities to be below $2
While $2 is significant compared to the price of current light bulbs, it will not be as significant with LED lighting because the lifetime of LEDs are much longer than traditional bulbs.
The technology is also applicable to outdoor lighting - for example in the automotive industry. Brakelight on current car models are already using LEDs. Combined with this new technology, the brakelights could be used to convey additional information. E.g. what pressure is applied on the brake.
Sources:
Boston University Smart Lighting Center
Hotstocked
Cellular-news
Wikipedia on LED
A large body of research into Smart Homes and autonomous systems for the domestic domain use some sort of context awareness. For instance, a system governing a smart home may detect that a person enter a room, and consequently turn on the light or change settings to suit the tastes of the individual. The other way around, the light could automatically be turned off, possibly along with other devices, if no person is present in the room. Ideally, this saves power, but very few have looked into how precise use prediction has to be, for systems like this to actually save power. It might seem like a good idea, but not all schemes actually save power. If a system simply turns of flourescent light everytime a room is empty, this will actually cost more power if the room is used frequently. This is due to the extra cost of turning on a flourescent light, compared to the power usage when lighted. Similar patterns exist for other types of devices, e.g. computers.
Harris and Cahill have made an empirical study of the potential for power savings from knowing user location and predicting future behaviour. Among the findings are, that the potential is limited, and that sensing user presence to a degree where predictions become accurate consume more power than it saves, due to the extra sensors needed. Instead, the look into systems with fewer sensors and find that for people with highly predictable behvaiour, i.e. regular usage patterns for e.g. office computers, it is possible to come relatively close to optimal power management, while for people with less regular usage patterns, the power saving is much smaller, or result in unwanted delay for the user, for instance because a computer is in standby mode when it is needed. In their study, they specifically look into the display of a computer and the computer itself. Besides looking into various strategies, one of the more interesing conclusions are that the best strategy differ depending on how often a resource is in use.
Source
Colin Harris and Vinny Cahill, An Empirical Study of the Potential for Context-Aware Power Management
Home automation seems to be all the rage these days, as more and more companies provide remote control capabilities for various products for the modern home. Other companies provide full solutions for home automation, allowing users to control everything from lights to heat and blinds.
Recently, a number of applications have emerged for the Apple iPhone to remotely control home automation systems. Several producers have showcased applications allowing remote control of their home automation systems from a touch screen interface for the iPhone. Using the iPhone has several benefits; home owners do not need at separate remote for controlling the system, and some of the remotes allows for control of multiple locations using the iPhone data service, so the system can be accessed from anywhere. Some of the remotes can also control media centers, which saves another remote. The first applications are expected out by the end of the year.
Source
gizmodo.com
The danish design company Designnord recently launched a design concept for visualizing energy consumption in the home of the future. The concept consists of a semi-transparent screen displaying a flower and an Information Portal. When first installed the flower will be a small sprout. The well-being of the flower is determined by the energy consumption in the home. Overuse will result in a withered flower that will eventually die if the overuse continues over a prolonged period of time. The Information Portal, accessed through a computer, uses graphs and statistics to display energy consumption over time and projects future use.
The design concept was created for the 2009 United Nations Climate Change Conference to be held in Copenhagen and received a “Notable Entry”-award at the 2008 Greener Gadgets design competition.
Sources
Designnord
Greener Gadgets
Could Java fit on a mote and is it a good idea at all? Joe Polastre from Sentilla gave a great KomIalt seminar on infrastructure for pervasive computing.
Sentillas main product is a Java ME CLDC 1.1 implementation for TMotes (IIRC, the specs are hard to come by, to the details in this post are by memory). Some of the features of their product is that you can use Eclipse to program the motes and that they support extensibility via Phidgets.
Now, one good reason to put Java on this architecture is to show that you can, but what about:
- Memory usage? The mote has 10 KB of RAM and 48 KB of Flash. Furthermore, there are 1 MB secondary storage for paging programs in and out. The net result is that you have 3 KB of heap space for your Java objects, but on the other side you have support for, e.g., over-the-air update and sensor network communication
- Energy usage? The figures that Joe Polastre showed were 1 uA for sleeping, 1 mA for idle, and 4 mA for communicating. This is an order of magnitude less than SUN Spots. Furthermore, it can be argued that for typical sensor network applications it is the energy used while sleeping that is the dominating factor and in that case a virtual machine may not be a bad solution
- Programming experience? The toolkit and environment is definitely Java-centric. For many this means higher buildability than, say, C, by making use of class libraries and static typing. On the other hand, Java is interpreted which means a performance cost…
In any case, it is definitely an interesting technology with evaluation kits available, currently priced at $399.
