It won’t be long before smart spaces become a major industry sector looking to integrate sensors and AI into everyday environments. As such, we will look at some potential examples of the future and what hardware you need to consider making such a project a reality.
What are smart spaces?
When it comes to the future of IoT-related technologies, the term “smart spaces” often gets thrown around. However, before we can investigate hardware and applications in smart spaces, we first need to define what a smart space is.
Simply put, any environment that has been digitized to the point at which it is both readable and writable can be thought of as a smart space. When we say readable and writable, what we mean is that key data about the environment can be read, processed, and then affected with the result. For example, indoor environments can be thought of as a smart space whereby the indoor air quality is measured for temperature and CO2 levels, the results are processed by a computer, and windows and AC units are then adjusted to improve air quality by replacing stale air, heating the space, or providing cooling. However, unlike a traditional environment control system, a smart control system would process the environment to react depending on multiple factors such as the number of people present, the current energy prices, and behavior patterns from users.
Fundamentally, a smart space is the act of digitizing an environment so that it can be monitored and controlled by intelligent systems that can respond in real time. Just about any environment can be turned into a smart space — roads, forests, and even city streets. If it can be digitized, then it can be turned into a smart space.
What will be the fundamental requirements for an embedded system in a smart space?
The fundamental component of any smart space (or any modern electronic circuit, for that matter) is an embedded microcontroller whose task is to read sensory data, process I/O, and communicate with remote servers to provide data and receive commands.
The second most important component in a smart space is communication: How does an embedded microcontroller communicate with a remote server for data? While some spaces (such as indoor industrial) may permit the use of wired Ethernet or RS-232, it is most likely that smart spaces will require the use of wireless communication. In this case, designers can choose between cellular, Wi-Fi, Bluetooth, and LoRaWAN.
Regardless of the wireless technology used, the very act of using wireless communication puts multiple pressures on a design, including power usage and security. While the power concern can be managed with infrequent data transmission and the use of low-energy microcontrollers, the security aspect of wireless communication introduces some major challenges.
For one, wireless security allows for attackers to gain entry of devices without needing to be physically next to the device. Second, smart spaces with multiple devices all using the same technology could be a tempting target for a hacker, as hacking one may provide an entry for all devices, and this would increase the reward for a successful hack.
Thus, an embedded device must not only be low-energy, but it must also integrate strong security features such as hardware encryption, true random-number generation, trust zones, and cryptographic accelerators.
To better understand smart spaces and the constraints that they face, let’s look at some examples of the challenges faced by a design and what an engineer would need to look for.
Example 1: smart bins
Smart bins are already starting to be developed, and they could help to create smart spaces in cities that can respond to usage in real time. For example, a bin with a sensor detecting its current usage would be able to arrange for pickups only when needed, and this would cut back on CO2 emissions from vehicles needing to collect the rubbish. Second, a smart bin would be able to track the bins’ usages against the current time, and this would give city planners insights into the traffic that an area experiences
As such a device would be incredibly simple (a simple PIR sensor would work here), almost any embedded microcontroller will have the processing capability to detect when rubbish is present and then signal some remote server for a bin collection when needed. Even though bins are in city areas, which have access to infrastructure, the fact that bins are visited by workers means that any maintenance needed on the sensor can be done with every collection, and this would include changing batteries. As batteries are easier to integrate than mains cabling with power converters, it makes sense for such a sensor to be battery-operated. Finally, the use of such sensors in a city area means that they can take advantage of cellular networks, as these are widely established in cities.
In this case, the most important factor to a design engineer would be the need for a low-power processor with the ability to communicate with a cellular modem or integrate its own cellular modem in the form of a system-on-module. Either way, low power would be the key to such a design.
Example 2: activity monitor
Whether it is an outdoor or indoor area, an activity monitor is a device that can observe the immediate activity of the surrounding environment. Such activity includes the number of people moving through the area, the length of time a loiter spends in the area, and even identification of individuals. In this example, such a device would primarily need a camera, and this already puts a processing demand on any microcontroller chosen. The use of facial recognition may also need to be done locally for privacy purposes (i.e., edge computing), and this will see the need for local AI processing.
Finally, the power concerns from such a device are somewhat irrelevant when considering that most camera systems use the main device’s power. This is to ensure that the camera has all the energy it needs to process video as well as ensure the reliability of the camera.
Example 3: smart lighting
Outdoor spaces that use lighting can take advantage of smart devices to make decisions on whether lighting is required. This has the advantage of saving energy when not in use, and the dimming of lights can also help with light pollution for residents nearby as well as the sky. Furthermore, the reduction in lighting increases the lifespan of lights, and this helps to reduce e-waste as well as reduce the overall lifetime cost of a lighting system.
To create such a system, sensors capable of detecting activity are needed but would not need to be operated to the same degree of accuracy or resolution as that found in the previous example (the activity monitor). In this case, the only point of concern is if a space is occupied, and more advanced data may look to see the speed and direction of an occupant (this could be used to create a prediction path on what lights to turn off and on).
It is also likely that such a sensor could do most of the processing on-chip and not need to use remote cloud services (of course, this could always be used to add functionality to the lighting system). Considering that smart lighting systems will have a stable power connection, there is little need for battery-operated sensor systems, and this means that low-power solutions are not essential.
Instead, an embedded device in this application should focus on having strong security for the simple reason that an attacker gaining access to the device would be able to control lighting at will, and this could be used for nefarious purposes (i.e., cover tracks, commit crimes, etc.). Strong security would also be needed, as such a system may also be able to identify individuals if a camera is used.
The future of smart spaces looks bright
We are still in the early days of smart spaces, and while many products on the market are being sold as “smart,” most of them are, in fact, are not smart (i.e., do not perform any real data processing nor react to events in real time). As engineers design and develop sensors and edge-computing devices for use in smart spaces, we need to make sure that the right kind of technology is chosen for the application, and to do this, we need to ensure that we fully understand what is required of a device.
In the case of smart spaces, privacy and security will be the two largest areas of concern, and thus, any embedded device in a smart space would need to have at the very least, multiple levels of hardware security, including encryption, key storage, and certificates. After that, energy consumption will be an important factor, as the likelihood that a smart space sensor will be battery-operated is high. From these primary concerns, the choice of communication technology would then be made, and it is at this point that an embedded smart space design can then finally be constructed.