Features > IoT
What We Need to Implement IoT – Talk Service-Oriented IoT (3) –

We talked about the definition and market prospects of IoT (Internet of Things) as well as its implementation in the last two posting. Today, let’s take a look at some current IoT implementation technology trends.

What’s Necessary to Implement Service-Oriented IoT

It’s easy to think of devices like cameras that are connected to the internet when one hears the term ‘IoT implementation’. As pointed out in the first posting, however, IoT is considered more valuable when it goes beyond just connecting devices to the internet and starts providing intelligent services by analyzing the data they collect.

Let’s see how IoT is making the transportation card system more meaningful: What people first notice is that they can just put the card on the reader connected to the internet instead of preparing the correct change. A bigger change many people don’t see is that the transportation card system analyzes the data from the terminals on buses in order to announce estimated arrival times, adjust intervals, show the best transfer spots, and even change bus routes best suited for the given passenger traffic pattern.

iot_01

The definition of IoT by LG CNS (Source: LG CNS Entrue World 2015)

In other words, there are four technological factors required in order to implement IoT: a smart device, an advanced network, cloud computing, and big data. Let me share some more specifics in the section below.

Four Technological Factors for IoT Implementation

Smart Devices

A smart device is a device with various sensors and a network connection. Smart watches are a good example of this. LG Watch Urbane has a gyroscope, an acceleration meter, a compass, an air pressure sensor, a heart monitor, and even a GPS sensor.

These sensors create meaningful data to figure out the user’s location as well as his or her health condition. It can also send this data to smartphones or for other services through Wi-Fi, Bluetooth, or NFC.

Smart glasses or wearable devices in the form of bracelets, automated vehicles and remote controlled robot vacuums as well as other electronics like washing machines and smartphones all belong to this ‘smart device’ category. You may think that only expensive or high performance devices up to the level of a PC can be considered a smart device, but all terminals that have sensing and communication functions basically fall into this category.

iot_02

Various smart devices (Source: Entrue world 2014)

A device that notifies drivers of available spots in a parking lot, for example, can also be considered a smart device as it has an infrared light sensor and communication equipment even though it can’t quite understand the situation or make a decision by itself.

Looking into the smart device a little closer, you’ll notice there are various sensors and communication modules, electronic modules mostly in the form of batteries, chips that act as the brain, OS as well as embedded software platforms and applications for remote control and monitoring functions. There’s also a system that combines communication, chips, OS, software platforms, and development environments together so the device can be developed in a simpler manner, and this is called an ‘IoT device platform’

One of the most commonly emphasized technologies in the field of smart devices is battery and power management. Because many IoT devices commonly are exposed to situations where they can’t be connected to a power source at all times, their battery capacity is important. The problem has been that batteries with better capacity were also larger and heavier and weaken the portability of the device.

Even for non-portable devices, installation space and construction costs for power equipment have been crucial issues. This is why the technology that saves power consumption to keep the device running longer is so vital.

Such low power consumption technology is being led by chip providers like ARM, Intel, and Qualcomm. ARM’s Big Little Processing is especially noteworthy. It has two cores with complementary functions so its optimal core can be used for the right performance that the task requires in order to consume energy more efficiently.

Complicated tasks with large amounts of data like games and GPS usually utilize a high-performance core, while simpler tasks such as web browsing and e-mailing use the low-power core so that the total power consumption can be optimized.[1]

These days there are also efficiency technologies that use software instead of chips. Let’s take a look at Android M’s Doze function.

Doze is a function that enhances power efficiency by monitoring the user’s pattern with motion sensors on the phone or tablet and turning off the background processes that haven’t been used for a while during standby mode. Google tested Dose on the Nexus 9 tablet and showed that this function increased the battery life by 50%. Technologies related to smart devices are being developed as we speak.

Advanced networks

A network is a communication technology that connects smart devices with smartphones or other smart devices and services. Network here, however, stands for an ‘advanced network’. How is the IoT network advancing?

The advancement of low-power technology is also happening in the field of networks. The low-power wireless communication technology standards such as Bluetooth 4.0 and Zigbee are some examples of these. Bluetooth 4.0 improved energy consumption by 50% compared to Bluetooth 3.0. The battery for the wearable band[2] with location and other alarm services using Bluetooth 4.0 lasts over a year.

LTE is not an exception, either. It is now evolving into LTE-M with low power technology which enables 10 years of communication with a single battery.

The advancement is under way for communication and bandwidth as well. 5G, the new wireless telecommunication standard, is aiming to reach 100 to 1,000 times faster transmission speeds compared to LTE. The IoT environment with its explosive amount of data now requires 5G to send and process data without lagging.

Let’s say you’re sending a remote-controlled robot to the site of a disaster. The robot has to send enormous amounts of data including video and audio as well as movement information from legs and arms to you without delay. You can then experience what the robot’s going through in real-time as if you’re there, and give it the proper commands according to the data you’ve received.

Cloud computing

Cloud computing came into being in order to enjoy high-quality IT infrastructure effectively while saving money. Cloud computing in the era of IoT, though, is more meaningful since it plays its role as the backbone of the service.

Once smart devices are connected to a cloud, the limit in storage space and time they used to have can finally be overcome so that users can enjoy services without interruption. Besides, big data analytics, which uses sensing data that IoT’s intelligent services revolve around also require cloud computing. The biggest trend in cloud computing technology for IoT is that more companies are investing in services and development clouds in order to create an IoT ecosystem. Xively, Pachube, Everything, Thingspeak, Thingworx, and Mobius are some of the platforms worth trying.

iot_03

Key map for IoT platform Mobius from Korea Electronics Technology Institute (Source: http://iotmobius.com)

Big data analytics

Big data analytics is one of the most important issues in the IoT environment. I mentioned that the focus of IoT will turn towards services from devices in the first posting, and big data analytics is the field with the biggest added value among all four technological factors.

Systematic analytics is becoming even more important because the amount of sensing data (including audio/video files, images and measurements from countless sensors) is increasing drastically and cannot be analyzed or understood manually.

One of the most eye-catching parts of big data analytics technology for IoT is the increasing interest in and application of machine learning. Machine learning is the modeling technology which enables machines to do modeling instead of humans for data analytics.

This means that computers will find the pattern and the classification system for data analytics, and come up with a meaningful result such as predictions for certain issues. So far, this technology has been used mainly for reading letters from image files, email spam filtering, and shopping suggestions.

Big data analytics in IoT, however, is used to find correlations within massive amounts of sensing data and to predict their future. GE’s Predictivity Solutions, for example, collects and analyzes 50 million types of data from 10 million sensors. This is to prevent downtime and clients’ assets like machines and equipment from shutting down all of a sudden.

TyssenKrupp Elevator also applied machine learning to their IoT data analytics. TyssenKrupp attached sensors to their elevators which collect data on speed, motor temperature, and door malfunctions and then send it to the cloud server.

Based on this data, they create a prediction model and see when a possible problem may occur. This way, they can perform regular checkups and fix problems before something happens. This means they can secure stability of that product while lowering the number of technicians dispatched.

IoT Implementation for Beginners

We’ve learned about the four factors of IoT implementation: smart devices, advanced networks, cloud computing, and big data analytics.

Some of you might have been a little bit confused by this posting, as these technologies for IoT implementation are quite complicated unlike the first and second articles of this series which says IoT has come close to our daily lives. For those of you who feel this way, I’d like to introduce some of the products and services that help even beginners implement IoT.

LittleBits, a hardware set for toys developed in the U.S., is an electronic circuit development kit that people without any knowledge of development can use as long as the concept of input and output is understood. This kit has various thumb-sized modules such as computing modules, motors, adjustable resisters to control the amount and speed of light and sound, as well as light sensors. Users can attach and detach these with magnets to build their own circuits.

ss

video: https://youtu.be/FPHj2H1C8JE

This kit with the most basic modules can actually create 150,000 different combinations, and you can also follow the projects shared on LittleBits online community, if you have difficulty building your own. Take a look at the video above for more details.

If you have some knowledge in development, you can also use Raspberry Pi or Arduino board for smart device hardware and Mobius Platform from Korea Electronics Technology Institute for software and cloud computing.

How can individuals, then, get all four technological factors of IoT implementation and combine them together?

iot_04

LG CNS IoT Platform

LG CNS suggests LG CNS IoT Reference Architecture which includes all technological factors, solutions and standards to implement various functions of an IoT system for these individuals. LG CNS also has its own IoT platform with IoT security and big data solutions.

Today, we learned about the necessary factors of IoT implementation as the topic of this series on IoT. In the next posting, we’ll talk about the current trends in the IoT field.

Written by Youkyoung Lim, Research & Development Advisory at LG CNS Information Technology Research Center

[1] Bumjin Yoon. 2013. “Mobile revolution, the question is ‘low-power’.” www.elec4.co.kr [back to the article]

[2] Source: http://www.beluvv.com/ [back to the article]

Post navigation

'Features > IoT' Category Post
  • IoT
  • Cloud
  • Big Data
  • Security
  • Data Center
  • e-Government
  • Transportation
  • Energy
  • Manufacturing
  • Finance