Technology Trends
The pace of technolgy change seems to increase every year, and none more so than within the electronics sector where advances in semiconductor technology, increasing processing speeds and data rates seem to happen on an almost weekly basis.
If you ask a child today to use their hands as a phone and pretend to answer a call, what do you think they will do?
Most probably, they will put their palm next to their ears like you would do with smartphones without any mechanical buttons. For anyone who was born before the 2000s, it might be more natural to use a thumb and pinky finger to do the same action. This shows how the electronics market is evolving at an extremely high pace with new devices coming out every year with more features, smaller in size and more powerful than its predecessors.
In this article, we would like to draw your attention to the three technology trends that we think will transform the electronics industry in the up and coming years.
Edge AI
Artificial Intelligence (AI) has often been a buzzword in the news recently, but what does it mean and what you can do with it? The term generally refers to the ability of machines to perform tasks that are inherently associated with humans, such as “problem-solving” and “learning”.
Some of the examples that are already available include smart assistance with tools like Siri and Alexa, conversational bots for customer service, customised recommendations based on your favourite shows or songs on Netflix and Spotify, and so on. Until recently, almost all computations were performed within the data centres in the cloud to satisfy the extensive computing power requirements of AI algorithms.
However, the issues that come with cloud computing; including the latency, scalability and cost as well as data security and privacy concerns, have sparked the evolution of AI at the edge of the network, or so-called edge AI. The benefits of implementing edge AI devices come from the fact that the computations are performed locally without sending a large amount of data to a remote location.
Semiconductor manufacturers are currently investing in the development of edge AI chips that have sufficient processing power to perform intensive AI computations. Deloitte estimates that more than 750 million edge AI chips will be sold in 2020. STMicroelectronics, for example, offers software and hardware solutions for implementing AI on the widely used STM32 family of microcontrollers, including STM32Cube.AI tool for mapping and running pre-trained Artificial Neural Network (ANN) algorithms.
5G
After four generations of cellular mobile communication technologies (from 1G to 4G), we have unavoidably reached the limit of speed and reliability of the connection that we can achieve on our devices.
With the number of internet users increasing exponentially, the existing network is struggling to support the increased mobile data traffic. 5G, the fifth-generation cellular mobile communication network, promises the latency of less than a few milliseconds and ten times the speed of today’s 4G networks. To achieve these parameters, wireless engineers are working on the development of several brand-new technologies, including millimetre wave broadcasting, small cells, massive MIMO and beamforming.
However, device manufacturers are faced with significant constraints, including cost, space and power consumption, when it comes to designing electronics components compatible with 5G. The operation of antennas, RF circuitry, as well as thermal management solutions, must be optimised to be able to completely leverage the benefits of 5G. Manufacturers such as Molex, TE Connectivity, Cypress Semiconductor, are just a few examples of companies driving the innovation in this area.
MEMS Sensors for Predictive Maintenance
With the adoption of Industry 4.0 and the Industrial Internet of Things (IIoT) initiatives, the manufacturing sector is increasingly investing in tools and solutions to achieve improvements in the efficiency of their production processes. Implementing predictive maintenance on the factory floor is perhaps one of the most prominent use cases for IIoT.
Predictive maintenance techniques can help to identify any deterioration in the equipment by continuously monitoring its “health” and hence avoid any unpredicted failures and downtime costs. With the IIoT in mind, sensors for predictive maintenance are required to not only withstand harsh operating conditions of the factory floor, but more emphasis is being placed on utilising wireless, ultra-low-power sensing technologies.
This puts the spotlight on micro-electro-mechanical systems (MEMS) sensors that can offer compact, low-power, yet extremely precise measurement capabilities for IoT-enabled predictive maintenance solutions. Although the popularity for MEMS sensors originated from the world of personal electronics, such as their use in smartphones and smartwatches for applications such as for tracking steps and distance with in-built accelerometers, the current advancements in MEMS technology can provide a more affordable and scalable alternative for existing sensing technologies within the industrial environment.