Researchers have developed FloorLocator, a breakthrough in indoor navigation technology, which combines the high efficiency of Spiking Neural Networks (SNNs) with the advanced learning capabilities of Graph Neural Networks (GNNs). This innovative approach ensures remarkable accuracy and scalability for floor localization, crucial for enhancing emergency responses, indoor positioning, and personalized recommendation systems.
Indoor positioning is transforming with applications demanding precise location tracking. Traditional methods, including fingerprinting and sensor-based techniques, though widely used, face significant drawbacks such as the need for extensive training data, poor scalability, and reliance on additional sensor information. Recent advancements have sought to leverage deep learning, yet issues such as low scalability and high computational costs remain unaddressed.
In a recent study (DOI: 10.1186/s43020-024-00127-8) published in the journal Satellite Navigation, on 11 March 2024, researchers from Chongqing University have unveiled “FloorLocator”, a system that revolutionizes indoor navigation with unprecedented accuracy and efficiency.
FloorLocator sets a new benchmark in indoor navigation, significantly outshining traditional technologies with superior accuracy, scalability, and computational efficiency. This innovative system integrates Spiking Neural Networks (SNNs) with Graph Neural Networks (GNNs), marrying SNNs’ computational efficiency with GNNs’ advanced pattern recognition. SNNs bring unparalleled computational efficiency to the table, while GNNs excel in sophisticated pattern recognition. This blend not only boosts floor localization performance but also deviates from the data-intensive, inflexible approaches of the past. FloorLocator reimagines floor localization as a graph-based learning challenge, mapping Access Points (APs) to a dynamic graph for effortless adaptation to new settings a feat unmatched by current technologies. Achieving at least 10% higher accuracy in complex, multi-floor buildings than the latest methods, FloorLocator’s success is attributed to the strategic integration of SNNs for efficient computation and GNNs for adaptive learning, revolutionizing indoor navigation.
Dr. Xianlei Long, lead researcher, emphasized, “FloorLocator is not just an advancement in technology; it’s a leap towards creating more resilient, efficient, and accurate indoor navigation systems. By utilizing a graph-based learning approach, it can easily scale to new environments without the burden of high computational costs and extensive data collection. “
FloorLocator surpasses current technologies in accuracy, scalability, and efficiency. This approach enables dynamic adaptation to new environments and sets a new standard in the field, offering vast applications from enhancing emergency responses to improving indoor positioning and personalized recommendations, establishing it as a key solution for future indoor.
Credit: Satellite Navigation
Researchers have developed FloorLocator, a breakthrough in indoor navigation technology, which combines the high efficiency of Spiking Neural Networks (SNNs) with the advanced learning capabilities of Graph Neural Networks (GNNs). This innovative approach ensures remarkable accuracy and scalability for floor localization, crucial for enhancing emergency responses, indoor positioning, and personalized recommendation systems.
Indoor positioning is transforming with applications demanding precise location tracking. Traditional methods, including fingerprinting and sensor-based techniques, though widely used, face significant drawbacks such as the need for extensive training data, poor scalability, and reliance on additional sensor information. Recent advancements have sought to leverage deep learning, yet issues such as low scalability and high computational costs remain unaddressed.
In a recent study (DOI: 10.1186/s43020-024-00127-8) published in the journal Satellite Navigation, on 11 March 2024, researchers from Chongqing University have unveiled “FloorLocator”, a system that revolutionizes indoor navigation with unprecedented accuracy and efficiency.
FloorLocator sets a new benchmark in indoor navigation, significantly outshining traditional technologies with superior accuracy, scalability, and computational efficiency. This innovative system integrates Spiking Neural Networks (SNNs) with Graph Neural Networks (GNNs), marrying SNNs’ computational efficiency with GNNs’ advanced pattern recognition. SNNs bring unparalleled computational efficiency to the table, while GNNs excel in sophisticated pattern recognition. This blend not only boosts floor localization performance but also deviates from the data-intensive, inflexible approaches of the past. FloorLocator reimagines floor localization as a graph-based learning challenge, mapping Access Points (APs) to a dynamic graph for effortless adaptation to new settings a feat unmatched by current technologies. Achieving at least 10% higher accuracy in complex, multi-floor buildings than the latest methods, FloorLocator’s success is attributed to the strategic integration of SNNs for efficient computation and GNNs for adaptive learning, revolutionizing indoor navigation.
Dr. Xianlei Long, lead researcher, emphasized, “FloorLocator is not just an advancement in technology; it’s a leap towards creating more resilient, efficient, and accurate indoor navigation systems. By utilizing a graph-based learning approach, it can easily scale to new environments without the burden of high computational costs and extensive data collection. “
FloorLocator surpasses current technologies in accuracy, scalability, and efficiency. This approach enables dynamic adaptation to new environments and sets a new standard in the field, offering vast applications from enhancing emergency responses to improving indoor positioning and personalized recommendations, establishing it as a key solution for future indoor.
###
References
DOI
10.1186/s43020-024-00127-8
Original Source URL
https://doi.org/10.1186/s43020-024-00127-8
Funding information
This work is supported by the National Natural Science Foundation of China (No. 42174050, 62172066, 62172064, 62322601), National Science Foundation for Excellent Young Scholars (No. 62322601), Open Research Projects of Zhejiang Lab (No. K2022NB0AB07), Venture & Innovation Support Program for Chongqing Overseas Returnees (No. cx2021047), Chongqing Startup Project for Doctorate Scholars (No. CSTB2022BSXM-JSX005), Excellent Youth Foundation of Chongqing (No. CSTB2023NSCQJQX0025) , China Postdoctoral Science Foundation (No. 2023M740402), and Fundamental Research Funds for the Central Universities (No. 2023CDJXY-038, 2023CDJXY-039).
About Satellite Navigation
Satellite Navigation (E-ISSN: 2662-1363; ISSN: 2662-9291) is the official journal of Aerospace Information Research Institute, Chinese Academy of Sciences. The aims is to report innovative ideas, new results or progress on the theoretical techniques and applications of satellite navigation. The journal welcomes original articles, reviews and commentaries.
Journal
Satellite Navigation
DOI
10.1186/s43020-024-00127-8
Subject of Research
Not applicable
Article Title
Accurate and efficient floor localization with scalable spiking graph neural networks
Article Publication Date
11-Mar-2024
COI Statement
The authors declare that they have no competing interests.
