LPWA network is the engine of long-range Internet of Things. And is the only cost-efficient solution to allow long-range M2M communications today. On the one hand, wide coverage of wireless IoT networks requires well-managed infrastructure distributed through millions of square miles. On the other hand, we are experiencing dramatic growth of the number of smart devices connected. Until now the LPWA networks were not management or capacity constrained. However, one of the biggest challenges in LPWA networks today is the number of different smart devices and sensors transmitting and receiving data at the same time. This challenge need be managed in in order to keep the network under control.
Most of the long-range IoT applications are deployed in ISM unlicensed spectrum and therefore can interfere with other IoT networks within the dedicated bandwidth. When the number of nodes increases rapidly the network should adopt and relocate resources by itself in real-time. A human cannot control a billion of nodes connected in a wide-area – this is where Artificial Intelligence comes along!
The design of an efficient, dynamic and automated network management framework requires support from the field of artificial intelligence. To solve the problem of human-free IoT network management and effective capacity utilization we have launched the automated algorithm based on artificial intelligence. The proprietary algorithm named Albert uses trained neural networks for real-time system control which allows self-management or long-range IoT networks with billions of smart devices.
Albert enables intelligent, real-time network control and traffic management between IoT gateways and smart devices within the network. It uses neural algorithms and Bayesian methods to optimize the interaction of nodes and gateways on the network. Albert adopts a number of network parameters in given area to find the best balance between available spectrum, dynamic noise environment, channel utilization, taking into account the distance from nodes to the nearest gateway, the QoS and interval between transmissions application requirements and energy efficiency preferences.
“For businesses, in light of industry trend to increase the number of devices self-managing LPWA network will help to overcome interference, utilize higher spectrum, and extend battery lifetime. In the end, all enhancements considerably reduce the operating costs and improve the service quality.” — noted Aleksandra Akhmadishina, Deputy CEO.
WAVIoT gateway with SDR technology and NVIDIA CUDA graphics accelerator for spectrum processing can scan and process up to 2.5 MHz of the radio band. It gives you an ability of continuous monitoring of radio spectrum and allows to accumulate information about noise environment for a long period of time. Such information is used further by Albert for smart channel allocation and signal power adjustment to improve packet delivery ratio and interference avoidance.
If the signal from one NB-Fi device is received by several gateways – Albert will make a decision to reduce the signal power and increase the data rate of such devices so only the nearest gateway will receive the signal. It allows to shorten the transmission time and helps to save the battery power as water meter do not have an external power feed. On the other hand in case of high level of noise or occasional noise burst on gateway Albert will increase signal power to make possible reliable signal reception by two or more gateways. Furthermore, the applications with high QoS will have a higher priority in the optimization process.
Albert can automatically handle nodes signal power, data rate and optimizes the frequency and time slots distribution for the network including millions of smart devices. At the lab tests showing the result as the number of errors and unreceived data packets become less as network become smarter and adapts to the environmental conditions.
“Based on analysts’ predictions by 2023 we will connect to the internet additional 40 billion devices. 15 billion of those are estimated to be LPWAN. It is 6 million new LPWAN devices per day. And if these predictions are even remotely correct then network management will become one of the key issues. Albert is a significant step forward for future worldwide expansion of NB-Fi network,” — commented Marat Zaripov, co-founder, and CMO.
The current version Albert will go through the tests in a real life project in Sofia, Bulgaria. The intelligent wide-area network will handle more than 1 million of different smart city devices such as smart meters (water, electricity and gas) or smart lighting.
We are currently rolling out network infrastructure and preparing the system for launch within several weeks.
We believe such automation will dramatically minimize human involvement which produces effective and time-saving solutions for proper and dynamic supervision of these large and heterogeneous networks.
And Albert is the first AI to help us manage long-range IoT networks with billions of smart devices worldwide.