Wireless technology is booming. Emerging applications such as connectivity and autonomous systems are forcing design engineers to address the management of massive amounts of machine-type sensors, increasing complexity across systems and heightening the need and expectation for quality, reliability, and flexibility of sensor networks.
Consider the evolution of 2G to 3G to 4G – each provided consumers with its best effort to transmit over the air to users in its path. Today, with 5G and the advent of 6G, customized communication is key to ensure the best possible connectivity. The time is now for engineers to create a network to not only handle the millions of people with cell phones, but also consider other entities that are connecting – such as vehicles. To customize amidst the noise, they need to bring intelligence to the system. This is where artificial intelligence (AI) and machine learning (ML) shine, since the technologies can assess the environment, and use wireless sensing for automated person detection to provide customized communication.
It’s essential to realize that complexities have gone beyond typical engineering systems. As such, rule-based intelligence – or AI – is now seen as the only way to tackle the complexities stemming from 5G and 6G wireless connectivity and sensor design. As design engineers tackle next-gen wireless communications, it is essential to understand where things are and the medium of transmissions. There are three challenges in doing so:
- Complexity and scale: The sheer number of users are many, so the complexity must be managed.
- Non-linearity: Not only are there too many users – but they interact together, and handling becomes difficult.
- Performance: Consumers won’t be happy with basic service – so design engineers must consider how to maximize it for each user.
It’s important to note, the only technology that can handle these three challenges is AI. With the future all about the end-user, wireless sensing for automated person detection assesses the environment through the characteristics of wireless signals. Research has been conducted on how to use the sensing information to improve the performance of wireless systems – and findings have revealed a few areas to consider to optimize the network, including channel state information (CSI), variability and its effects on CSI, robustness to identify important and non-important information and handling scale.
5G has timidly tried to apply AI sensing techniques, and its focus was optimal performance. 6G is still being defined – it’s about 6 years away based on research and industry momentum – but it’s clear that using sensing in the channel seems to be an inseparable component of 6G. Hence, 6G is embracing sensing and AI in a serious way to address the considerations and challenges mentioned above.
MathWorks can support design engineers in the research and development (R&D) phase to accelerate processes with various toolboxes based on AI tools. The success of 6G will depend on the effectiveness of AI-based estimation models to enable design of 6G applications. An effective workflow will make this possible, and this includes: data generation, algorithm selection and AI monitoring, training and validation, and deployment.
With enough data, AI tools can give design engineers low complexity, high performance solutions to tackle non-linear problems – which are typically hard to solve. AI models can therefore result in satisfactory overall system performance to simultaneously help engineers manage performance, non-linearity, and complexity. AI and machine learning techniques can also help design engineers meet the high expectations of delivering stringent quality of service (QoS) metrics, including increasing throughput and capacity, and decreasing packet error rates and power usage.
Ultimately, AI is coming to wireless communications with a vengeance, and any engineer in wireless must be equipped with AI techniques to tackle the increased complexities and enable top-end performance for 6G solutions.
Original article can be seen at: