In 2021, the ACM International Conference on Mobile Computing and Networking (MobiCom), a prestigious international conference, published a paper on positioning inside tunnels via the use of magnetic forces, a study conducted by Professor Chen Jyh-cheng and his team of students. This was the second time that a full-length paper (with all of its authors being Taiwanese) was published by MobiCom since its establishment in 1995. Concurrently, the team came in first place at the MobiCom 2020 Student Research Competition (category: graduate school), marking the first time that Taiwanese succeeded in this internationally renowned competition.
Professor Chen indicated that self-driving cars will be the trend of the future. Today, the problem with self-driving systems is that satellite positioning fails or is unable to operate effectively in areas including tunnels and multi-level roads, preventing drivers from trusting the systems. Accordingly, this study used magnetic fields to achieve positioning in areas where satellite positioning fails. The study results (i.e., “MVP: Magnetic Vehicular Positioning System for GNSS-Denied Environments) were subsequently published by MobiCom 2021.
Positioning is an indispensable function in self-driving cars because they rely on it to identify their positions. However, the inability of cars to receive satellite positioning signals in areas such as tunnels, parking lots, and underpasses denies the cars from performing self-driving. Additionally, in multi-level roads, satellite positioning is unable to determine the exact positions of the cars (i.e., whether they are on elevated roads or ground roads), causing drivers to potentially receive wrong driving instructions and making self-driving unsafe. Professor Chen’s team introduced cars that used magnetometers to measure magnetic fields, and utilized an algorithm that it had developed to compare the measured magnetic fields with magnetic field maps. The results showed that the proposed method achieved positioning and offered a positioning accuracy of 5.14 meters without the use of satellite positioning. Currently, the research team has conducted large-scale road experiments in 56 tunnels and 23 bridges in two countries for 36 months, producing 5,943 pieces of data that provided correct positioning to verify the effectiveness of the system. Moreover, the system can be accessed using only smartphones and requires no high-priced positioning equipment, considerably reducing the costs of precise positioning when satellite positioning is unavailable. This study is the first-ever study to use magnetic forces to achieve precise vehicle positioning in areas where satellites fail, increasing completeness and safety of self-driving systems.
Since 1995, MobiCom has published less than 10 papers with authors from Taiwanese research institutions; most of said papers involved Taiwanese researchers working with foreign universities. The last time that a study was conducted solely by Taiwanese team members and published by MobiCom was 22 years ago in 1999. The present study features Ph.D student Wang Jia-cheng as its first author, who came in first at the ACM MobiCom 2020 Student Research Competition, marking the first time since 2005 (the year that the competition was introduced) that a student from a Taiwanese university came in the top 3.
Professor Chen’s team expressed their hopes for the study results to enable self-driving cars to have positioning capabilities in any environment, making self-driving car systems more complete and their passengers ride with a peace of mind.
Chinese version: https://www.nycu.edu.tw/news/3056/