×
Home About Us Blog Careers Contact Us

The emergence of Internet of Things, Big-data and Machine Learning Technologies

Driving Healthcare Innovations

By Mohammadu Ramsath on May 1st, 2019

Do you know what the internet is? Probably yes, if you are a technical person. Even if you are not a technical person, you have an idea on the technical aspect of the internet to a certain extent and its usefulness to you, to the businesses and to the world. Internet became an essential need and an integral part of our daily life. Everybody uses the internet nowadays for various purposes. But do you know what Internet of Things is?

An interesting point to highlight is the internet connects the people (people’s computers/mobile devices) and the computers (mostly servers) owned by another party. Similarly, for you to understand in simple terms, Internet of Things is the network connecting all your electronic/non-electronics devices as well as sensors and actuators to the internet. Now you may have a very brief idea about Internet of Things. Let’s dive more deeply into some use cases, technical aspects and benefits of Internet of Things.

 

What is Internet of Things?

Though Internet of Things is a new buzz word in technology it’s not completely a new concept.  As acronym stands we already had the internet of computers (computers connected to the internet), and now we have internet of things (things are connected to the internet) with the advancement in the embedded computers and chips technologies.

Many tech enthusiast and software developers have a misconception that “Internet of Things” is all about hardware and electronics.  It’s not only about hardware devices, but it’s a complete ecosystem where one machine or device can communicate to another device or software system without any human interaction and perform all or part of its activities. For the machine or device, you can think of any, which has some embedded electronic parts to communicate to sensor network or internet. Nowadays even non-electronic devices such as plants, farms can be connected to the internet with the help of sensors or embedded devices.

For example, take an automobile. It’s now equipped with very sophisticated electronic controls and sensor system. Let’s imagine there is also a connected (to internet) petrol station available. Then the simple IoT scenario is; when the fuel sensor detects the low fuel availability, then normally it alerts the driver. But if we have internet of things concept (sensors and communication to a network) implemented in the automobile, not only the alert to the driver can be generated, rather it can immediately send the request on the network for next available filling station and send the fuel fill request along with automobile identification details. Then this request or order already becomes available in the filling station, and the navigation unit of the automobile automatically shows the route to the filling station, as a result of the feedback or notification is received from the fuel sensor unit. Once the driver follows the route and reaches the filling station, the automated (and connected) filling station has the order and automobile identification beforehand. Using possible mechanisms it processes the activities like scanning the automobile to verify the identification, filling the oil to the automobile automatically and sending the notification and billing to the automobile’s control unit. If the automobile control unit has further IoT concept implemented, there is a possibility that even the payments can be made through the control unit by communicating to a payment gateway using the card or any other payment mode information entered by the user.

In this explained scenario there can be many practical issues that may arise to you. I would like to hear if you have any, but let me share my thought as well.

In the above scenario, there can be a possibility that the driver has to make other decisions while driving. For instance, if the traffic on the road is high, then he may decide to go to another filling station than the nearest. This is where we need analytics, Big-data technologies, Machine Learning and AI models coupled with IoT solutions to better serve the users with a maximum possible interconnection of the entire ecosystem.

This is just one simple scenario that the IoT concept can be implemented. There can be many more you can think of now. The IoT ecosystem spans across almost all kinds of industries such as farming and agriculture, industrial and machinery (IIoT-Industrial Internet of Things), Healthcare (IOMT – Internet of Medical Things), automobiles, electronics, transportation, manufacturing units, utilities, etc.

Read more about best IoT Trends: www.forbes.com/sites/bernardmarr/2019/02/04/5-internet-of-things-trends-everyone-should-know-about/#412d6f0f4b1f

It has become a rather useful feature for the healthcare sector because of the reduced cost associated risks all at the same time.

For example, let’s consider glucose or pressure monitoring. If a physician has the access to the blood pressure and glucose level data of a patient throughout the day with the pressure monitoring and transmitting device, the doctor has more insight into patient condition than obtaining pressure and glucose values of just two or three instances per day, which involves human resources and physical resources. Not only the immediate conditions but also some other useful inferences such as a graph of the variation of blood sugar vs blood pressure of a patient can be obtained and made visible to the physician.

Another significant benefit of IoT in healthcare is the continuous monitoring and alerting system of any patient who is in critical conditions. It can improve the speed of information transfer to a doctor than having a nurse or some other hospital staff pass the message.

Overall it reduces the need for a physician to visit and check patients more frequently. Instead, the IoT ecosystem keeps collecting patient vital signs and other data and transmitting to relevant recipients. This enhances the quality of care and reduces cost.

Medical devices integrated with IoT ecosystem eliminates the human mistakes in data entry and actively monitor the patient test data so that it can act faster than a physician or lab staff, for any critical condition detected in test results.

Also, with the IoT ecosystem, the telemedicine becomes more effective for the patients who cannot mobilize much or in a scenario where the physician and the patient are physically separated.

Following diagram illustrates a routine use case to be addressed in Healthcare IoT solution.

Read more about mobile healthcare with Medical IoT – www.openaccessgovernment.org/mobile-healthcare/62545/

 

What are the Technologies Used?

So far I have explained what “Internet of Things” is, how it works and how it’s interconnected to other modern technologies.  Now let’s look at and bit deep dive into IoT technologies which I feel you can benefit from.

The major components of IoT technologies are the devices and communication protocols since it is mostly the device-to-device communication what enables the IoT possible.  The next most important component of the IoT ecosystem is the connecting (software) platform.

So when you think of devices, we can categorize them into three;

  1. Sensors – Only collect data and transmit them to a (cloud) platform.
  • Ex – Temperature monitoring, Pressure Monitoring, Glucose level monitoring sensors etc.
  1. Actuators – Only receive commands/ notification from any other device/system and act upon it.
  • Ex – Motors, Robotic Arms etc.
  1. The devices acting both as Sensor and as Actuators – These devices both accept the commands and sense the data, and send (bi-directional communication)
  • Ex – Lab Analyzers

 

The IoT devices mostly consist of energy source, power management modules, RF (Radio Frequency) modules, and sensing modules. RF modules are responsible for device-to-device or device-to-system communication. The major known RF technologies used in the IoT are as follows.

 

We can see that high-rate long-range systems are well covered by current 3G and 4G cellular systems. Short-range high-rate is covered by WIFI, and short-range low-rate by Zigbee and Bluetooth systems. There are also many other wireless and wired protocols such as NFC and RFID which also enable the IoT devices’ communication.

 

What are we doing?

As mentioned above, after the devices/systems and the communication mode, the next most important thing in the IoT ecosystem is the connecting platform. You may have lots of IoT devices which collect data and perform some functions for you, but if you don’t have the data collecting and connecting platform then all your devices become a dump.

So we have been focusing on developing an IoT platform for anyone to provision new medical/non-medical devices and send and receive data to provisioned applications or systems. A question may arise to you, that while we already have many IoT platforms developed by tech giants and open source communities, do we need to re-invent the wheel?  Yes, you are correct! We do not need to reinvent the wheel, thus we do not develop an IoT platform from scratch. Rather we use some of the open source IoT platforms such as Device Hive and Kaa as our base platform and introduce some other custom features and analytics to them.

Analytics on the collected data is the most value adding a feature to any of the IoT platforms. Usually, there can be two types of analytics we can perform, one is the real-time analytics, other is the batch analytics.
Batch analytics is performed on the persisted data while real-time analytics is performed on streaming data in real-time.

Last but not least, this is the point where the Big-data, Machine Learning and AI (Artificial Intelligence) concepts are involved in the IoT ecosystem. When we collect data from sensor networks and devices, it produces huge volumes of data, especially in healthcare. So, to handle them and persist them, and to do analytics on top of them, we need Big-data technologies. Similarly, for us to infer some useful information from the history of data, we may need ML (Machine Learning) and AI (Artificial Intelligence) models to be created and run. You can read more on these concepts in-detail with some technical insights on our other blog posts sooner or later.

Share now