Edge computing needs Edge AI

20 August 2018

The tech analyst CBInsights recently published a report called “What is Edge Computing?”. It is an interesting report that clearly points out that edge computing will be critical if the digital transformation is going to be able to deliver business value over time in line with our high expectations. This is obviously driven by the strong IoT and mobility trend.

The general IT trend has sent the computing power to the cloud but now it needs to swing back again and put more focus on the edge.

Edge computing is all about putting the data processing power back to the edge rather than the cloud. However there are many challenges that we need to overcome:

  • Challenges to set up communication between IoT devices and the cloud
  • Communication costs and costs for wireless communication solutions
  • Getting access to computing power with guaranteed performance when you need it
  • Challenges with slow, interrupted or non-working wireless connections
  • The need for real time operations
  • The need to manage the troves of data that will be generated by IoT devices, connected cars and other digital platforms
  • Personal data integrity

Where is edge computing trend most important?

The report from CBInsights talks about the following areas where key applications for edge computing will be found.

  • Transportation
    • eg automotive vehicles
  • Healthcare
    • eg remote patient monitoring
  • Manufacturing
    • eg predictive maintenance
  • Agriculture & Smart Farms
    • eg monitoring remote sites and livestock
  • Energy & Grid Control
    • eg safety monitoring with oil and gas utilities

There are a large number of possible applications in each of these areas defined already and it’s just the beginning. The challenges to get edge computing to work is quite different depending on the applications.

Edge computing developments

All the major tech players like Amazon, Google and Microsoft are investing heavily in solutions for edge computing, and they are emerging as edge computing leaders.

But interest is not limited to these giants. As more connected devices emerge, many players within the rising ecosystem are working on software and technology that will enable edge computing to take off.

Many have put a lot of faith in LPWAN (Low Power Wide Area Networks) for IoT applications. Using LPWAN’s will make it easier and cheaper to connect IoT devices to the cloud and central computing power. But relying on LPWAN and cloud computing only, will not be enough for many applications, due to the fact that these networks are slow and offers limited bandwidth. Edge computing will be required in many cases.

Some people believe that in order to create real business value from LPWAN systems, it’s not enough with the most basic applications, such as tracking. To make real money, you need a system that can deliver high value content. 

Imagimob is a Swedish startup that has focused on developing artificial intelligence software that allows for local processing on hardware devices on the edge, Edge AI. Edge AI technology can deliver high value content from IoT edge devices.

What is Edge AI?

Edge AI means that AI software algorithms are processed locally on a hardware device. The algorithms are using data (sensor data or signals) that are created on the device. 

A device using Edge AI software does not need to be connected in order to work properly, it can process data and take decisions independently without a connection.

The recent Gartner Hype Cycle for Emerging Technologies (published August 16, 2018) is recognizing Edge AI as a "must-watch technology". They Hype Cycle is dispersing technologies along the phases of the cycle: innovation trigger, peak of inflated expectations, trough of disillusionment, slope of enlightenment and plateau of productivity.

Edge AI applications

Imagimob has been involved in a large number of projects where Edge AI has been the key component in the system solution. Running the Edge AI software on small, low-cost MCU’s has been critical in the system architecture. The software works in real time, and has been able to report certain conditions or behavior on the device based on sensor data.

Imagimob has been involved in fall detection systems for the elderly, intelligent clothes for safety applications, smart access systems, smart fitness systems, pet monitoring systems, intelligent steering wheels and self-predictive electric drives.

Imagimob has developed software and technology that is specifically designed for running AI on small devices. It consists of three major components:

  • Hardware and software for capturing sensor data
  • Software for training the AI model for different application scenarios
  • Software (binary) that runs the AI model on the IoT device

Conclusion

The report from CBInsights is very interesting and highlights the importance of the edge computing trend. We firmly believe that Edge AI will be a very important component in the edge computing architecture.

And as the final conclusion in the report says:

“And while some use cases prove the value of edge computing more clearly than others, the potential impact on our connected ecosystem as a whole could be game changing”.

If you want to know more about how Edge AI can help you solve your challenges with edge computing you can download our white paper about Edge AI here.