Sensor fusion and Edge AI are increasingly critical in various applications due to their synergistic ability to enhance data processing and decision-making at the source.

Sensor fusion combines data from multiple sensors to provide more accurate, reliable, and comprehensive insights, addressing the limitations of individual sensors. Edge AI, in turn, brings machine learning capabilities closer to the data source, enabling real-time processing and analysis without the latency associated with cloud computing, ensuring swift and efficient decision-making, and providing more end-user privacy.

This combination is pivotal in applications such as autonomous vehicles, smart cities, industrial automation, and healthcare, where quick, accurate decisions are essential.  The market for sensor fusion is dynamic and exciting. In 2021, smartphone sensor fusion accounted for the largest market share, as an Emergen Research report stated. Now automotive is expected to take the lead, according to a recent report from Market Data Forecast, primarily because of vehicle safety needs.

Industrial applications also follow this trend. Indirect proof of its growing importance is that TDK Corporation, one of the world’s leading electronic components and devices companies, just formed a new entity to bring Edge AI and sensor fusion to Industry 4.0. The mission for all sensor providers should be to help manufacturers struggling to process real-time data streams from their machines.

Considering these developments in all the fields mentioned, POLYN’s NASP has an advantage. It is built for Edge AI and provides the best possible sensor fusion methods based on deep neural networks. Deep model fusion is an emerging technique that merges the predictions of multiple deep learning models into one.

Multi-channel neural networks are trained to process several different inputs, such as how the human brain extracts insights using a wide range of sensory inputs (audio, visual, tactile, and more). A multi-channel neural network helps generate predictions from multi-modal data, where each data stream is important to the overall joint inference generated by the network.

The increasing role of sensor fusion and Edge AI was, in our view, the important theme of Sensors Converge 2024, the premier event recently held at the Santa Clara Convention Center.

Healthcare innovations received considerable attention at last year’s conference and were explicitly covered this year. Wearable sensors – offering real-time monitoring of vital signs, glucose levels, and even stress indicators – are set to empower patients and healthcare providers with accurate and timely data, enabling proactive health management. Remote monitoring solutions that leverage advanced sensors to track patients’ health conditions outside traditional clinical settings enhance patient care and reduce healthcare costs.

Sensor fusion enables a more sophisticated data interpretation, leading to the most accurate diagnosis and more intelligent decision-making. For example, POLYN’s NeuroSense processes heart rate data from PPG and body movement data from IMU sensors, making it possible to distinguish between a more rapid pulse that is due to running and one that is due to a heart problem. In general, measuring only electrical and movement signals falls short of capturing the full spectrum of symptoms; therefore, it must be supplemented with chemical sensor data, and here we see an area for future development.

The area of sensor application most discussed in every conference and the media is automotive. The coming years are poised for a dynamic transformation in automotive sensor technology driven by advancements in Advanced Driving Assistance System (ADAS), autonomous driving, electrification, and the widespread adoption of sensors across global car fleets. POLYN is proud to be at the frontier of these innovations with its VibroSense, which extracts useful data from tire vibrations. Road condition data monitoring, fused with data from LiDAR, RADAR, camera, acoustic, and ultrasonic sensors, supplying alerts to the vehicle ADAS and utilities responsible for road services will increase driving safety significantly.

One could feel the heartbeat of technological innovation at the announcement of the 2024 Best of Sensors Award winners. This prestigious award recognized the most innovative and impactful technologies introduced over the past year. Several winning solutions are associated with the concept of wireless low-power sensor node operation. Among them: LAYER®Vault by Dracula Technologies, the low-light OPV energy harvesting crafted with Organic Photovoltaic inkjet technology; EnerCera rechargeable battery by NGK Insulators; Novelda’s X7 Ultra-Low Power Presence Sensor; and Blecon’s solution enabling sensor IoT connectivity with Bluetooth Low Energy.

POLYN’s AI chip supports this concept, which is essential for Edge AI. It offers sensor data pre-processing at an ultra-low power consumption of only around a hundred microwatts.