Emerging IoT Device Connectivity Solutions

The backbone of any such IoT system is the internet connection. Without an internet connection, it operates like an ordinary device without the term “smart.” Depending on the type of IoT application, there are quite a number of connectivity solutions available in the market, but almost all solutions have their limitations, especially regarding the number of devices that can connect simultaneously. To address these issues, this article further discusses some emerging IoT connectivity technologies.

Currently, there are many wireless technologies with varying coverage from short to long. These include technologies like Wi-Fi and LTE, which almost everyone uses in daily life. Bluetooth is one of the earliest wireless technologies, initially an internal project of Nokia, later popular for very short-range (less than 100 meters) portable connections. However, Bluetooth consumes a lot of energy during operation, leading to the development of BLE (Bluetooth Low Energy). BLE is very popular in portable devices like smartwatches and other wearables.

Then came Wi-Fi, which is common in short-range network connections. Wi-Fi is standardized under the IEEE 802.11 standard and supports data rates much higher than Bluetooth (up to 600 Mbps). ZigBee is another popular technology widely used in industrial and home automation IoT applications. Compared to BLE, it can operate efficiently under larger network traffic loads. However, in terms of data rate and power consumption, BLE would be a better choice, and using BLE can expect longer battery life.

Finally, for remote connections, there is cellular communication technology, the latest being the fifth generation (5G). Currently, LTE (Long Term Evolution) is part of the 4G (Fourth Generation) standard, deployed globally, providing higher data rates than 3G, but still far behind true 4G connections. Today, with the arrival of 5G, data rates are expected to increase by about 100 times, latency to decrease by 10 times, and support more connected devices.

Existing wireless technologies can provide connectivity for most IoT applications, but there are still some shortcomings, especially when handling a large number of IoT devices. In the future, the number of connected IoT devices is expected to reach around 25 billion. To address the connectivity issues of such a large number of devices, many new technologies are being researched to solve the problems of existing technologies while retaining their good characteristics.

❶ CS-Based IoT Connectivity

This type of connection is based on the concept of Compressed Sensing (CS), where signal detection is performed by utilizing a few device activities. Unlike LTE and NB-IoT, CS-based connections use unlicensed random access. This allows IoT devices to directly upload payload data and transmit preambles. However, the high complexity of CS algorithms is one of its challenges, potentially leading to additional issues when handling a large number of devices. As the number of connected devices increases, it also needs to expand bandwidth. To address these issues, advanced technologies like NOMA and mMIMO have been developed.

❷ NOMA-Based IoT Connectivity

NOMA (Non-Orthogonal Multiple Access) is one of the most efficient unlicensed massive access methods, relying on the key principle of allowing overlapping of simultaneous frequency signals. This is achieved through the deployment of Power Domain Multiplexing (PDM), or Code Domain Multiplexing (CDM) and Successive Interference Cancellation (SIC) to perform individual decoding for each device. Its main advantage is allowing simultaneous access for multiple devices without expanding bandwidth. However, there are still some challenges in implementing NOMA, including suppressing error propagation during the SIC stage and optimizing factor graphs.

Simplified NOMA Connection

❸ Machine Learning-Assisted IoT Connectivity

Finally, machine learning-assisted connectivity can handle issues related to traffic control, resource allocation, and wireless link adaptation. Artificial intelligence and machine learning are two rapidly developing fields in technology, with rich solutions for every type of problem. With machine learning, issues present in dynamic wireless environments can be resolved by modeling and providing optimal resource allocation. The only major challenge here is the trade-off between the higher accuracy of machine learning models and the computational and power required to achieve such high accuracy.

❹ mMIMO-Based IoT Connectivity

mMIMO (Massive Multiple Input Multiple Output) is another emerging technology that can serve 5G by handling a massive increase in data traffic. Unlike NOMA, mMIMO processes wireless resources in the spatial domain, thus allowing a large number of Machine Type Communication (MTC) devices. In MTC, mMIMO can effectively resolve access conflicts, reduce access delays, and enhance random access capabilities. So far, there has been little research testing the true potential of distributed MIMO. It is necessary to design a transparent space with low cross-correlation.

mMIMO System Diagram

All the emerging technologies discussed look very promising as they can provide massive connectivity while ensuring reliable and low-latency communication between IoT devices. However, creating a “perfect solution” is very difficult, and these emerging technologies also have their own limitations and issues when applied. There are also technologies like satellite IoT that can provide massive remote connectivity and are under research.