Lean approach for affordable smart, connected products: A few tips and examples – Part 1

Planning to launch their first smart, connected product, enterprises may be afraid of the investments needed. However, there are ways to reduce costs of IoT development and implementation and still come out with the solution bringing business value for both vendors and customers.

The Internet of Things is actively growing in various industries, and more and more businesses adopt or consider the adoption of smart, connected products. However, for many business players, IoT sounds not only promising, but also expensive, and they may believe that cutting-edge products equipped with sensors and communication technology are somewhere beyond the investments they can afford, Alex Grizhnevich, process automation and IoT consultant, ScienceSoft.

Nevertheless, the dream of entering the IoT world may come true with the right approach, for example, with the so-called lean approach, which presupposes the systematic waste minimisation without sacrificing productivity (in other words, the implementation of only what is needed and the reduction of what doesn’t add value neither for a customer, nor for a vendor).

So, there’s no need to splash on a product in the very beginning or delay the launch until accumulating a huge budget – a solution with a well-built architecture, delivered even within a small budget, may be expanded with advanced features later.

In this article, we are happy to share our vision of affordable IoT development and implementation and show the ways to deploy effective smart, connected products within a limited budget.

Minimising the volume of analysed data

Although a smart, connected product can be theoretically equipped with a huge number of sensors taking numerous readings tens and even hundreds of times per second, not every IoT solution requires the big number of sensors and immense frequency of readings for effective work. For the sake of reducing IoT development and implementation costs, it’s possible to analyse only the data crucial for product’s performance.

Alex Grizhnevich

It doesn’t require many computing nodes, and in some cases a company may even resort to traditional data storing and processing tools. At the same time, the matter is to choose the optimal volume of the data needed to monitor and support a smart, connected product, but not sacrifice the quality of this product’s performance. Also, minimising the volume of analysed data can be partially addressed with the filtering of data at the gateways.

Example: Snow level monitoring – Sensors take data, for example, each 30 minutes (the snow level is highly unlikely to increase dramatically within these time periods) and the IoT system informs snow removal services about the necessity to clear the streets in certain city areas.

Starting with a solution with simple logic

Packing up an IoT solution with all the features imaginable will not contribute to its effectiveness when these features are underused or not needed at all (what is more, it may make a smart, connected product slow and ineffective). On the other side, it will be less expensive (and more secure in terms of ROI) to begin with adding only basic IoT functionality that solves real business problems to a smart, connected product.

There are plenty of examples of simple IoT solutions that work with rule-based logic and don’t even use machine learning. In the long-term perspective, more advanced features can be added (as well as new modules to a connected product architecture) when a company gets corresponding needs and resources.

Example: Smart waste containers – Smart waste management is an efficient alternative to traditional schedule-based trash collecting. The logic of this solution can be rather uncomplicated: sensors located on top of a waste container send the data that the container is full, and the IoT system sends a notification to the waste collection service. As in the previous example, there is no need to take sensor data too often.

The simple logic of this solution can be expanded with adding the machine learning component to a smart product architecture in the future. For example, it’s possible to build schedules of waste collection (predicting the time when the containers in certain areas are full) and develop the most convenient routes for the waste collecting machines to travel in the city.

In the second part of the article, we will continue the discussion and explore two more ways that enable enterprises to deploy effective smart, connected products within a limited budget.

The author of this blog is Alex Grizhnevich, process automation and IoT consultant, ScienceSoft

About the author

Alex Grizhnevich is a process automation and IoT consultant at ScienceSoft, an IT consulting and software development company headquartered in McKinney, Texas. His 17+ years’ experience in IT and OT includes programming industrial microcontrollers, developing web and desktop applications, databases and document management solutions for oil & gas and logistics. Holding the degree in automation and management of industrial processes, Alex is now focusing on IoT and machine learning on sensor data.

Comment on this article below or via Twitter: @IoTNow_OR @jcIoTnow

Recent Articles

Accelerating digital transformation for a greener future

Posted on: May 13, 2021

As countries and economies prepare to come out of the pandemic, the way we run businesses will never be the same again, says Bhushan Patil, SVP EMEA, Tech Mahindra.

Read more

Trust platform design suite speeds embedded security implementations

Posted on: May 13, 2021

In 2019, Microchip Technology released its Trust Platform for its CryptoAuthenticationfamily, bringing the pre-provisioned solutions for hardware-based secure elements to companies of all sizes that want an easy way to implement secure authentication.

Read more