(Press-News.org) In autonomous driving, robotics, and augmented reality, accurate localization remains one of the most challenging problems. Traditional visual–inertial odometry systems often struggle with environmental variations, sensor noise, and multi-modal information fusion, limiting applications such as autonomous vehicles navigating complex urban environments and drones operating in GPS-denied areas.
In a study published in the journal iOptics, a research team from Chongqing University of Posts and Telecommunications proposed a modality fusion strategy—visual–inertial cross-modal interaction and selection mechanisms. This approach not only improves localization accuracy for robots under GNSS-denied conditions, but also enhances the robustness of the algorithm in complex environments.
“The inspiration for attention mechanisms comes from human visual and cognitive systems. When we look at an image or understand a sentence, we do not process all information equally, but instead focus on the most relevant parts,” explains lead author Associate Professor Changjun Gu. “When reading different information, we typically establish connections between these key pieces, forming the foundation of the widely used transformer architecture today. Therefore, we introduced attention mechanisms in our research.”
The main hurdle lies in effectively leveraging the complementary strengths of visual and inertial sensors while addressing their respective limitations. “Visual sensors provide rich environmental information but are sensitive to changes in lighting conditions and textureless environments, while inertial sensors deliver continuous motion measurements but suffer from accumulated drift over time. Hence, a key question remains: how can a system intelligently combine these modalities to achieve robust and accurate localization across diverse environments?” explains Gu.
To that end, the research team introduced two technologies. The first is global-local cross-attention module, which enables effective knowledge exchange between visual and inertial feature representations. Unlike conventional methods that simply concatenate features, this module allows the system to learn which aspects of each modality are most relevant for accurate localization. By preserving the complementary strengths of each sensor while enabling meaningful cross-modal communication, the system achieves remarkable robustness in challenging scenarios.
The second is dual-path dynamic fusion module. Rather than blindly fusing data, this module identifies and selects the most informative features according to current environmental conditions. In well-lit environments with rich visual textures, the system emphasizes visual cues; in challenging lighting or feature-sparse areas, it shifts focus to inertial measurements. This adaptive approach ensures optimal performance across diverse operating conditions.
Project leader Xinbo Gao, explains, “Up until now, deep learning–based visual–inertial localization methods have primarily relied on simple feature concatenation followed by a pose regression network, with few studies considering modality fusion and selection. Our approach demonstrates that cross-modal interaction and selection mechanisms can be effectively incorporated into visual–inertial localization, leading to improved accuracy.”
The team hopes this work will encourage researchers to explore visual–inertial modality fusion and selection strategies, thereby enhancing both localization accuracy and generalization ability.
###
Contact the author: Changjun Gu, College of Automation, Chongqing University of Posts and Telecommunications, Chongqing, 400065, China. Email: gucj@cqupt.edu.cn.
The publisher KeAi was established by Elsevier and China Science Publishing & Media Ltd to unfold quality research globally. In 2013, our focus shifted to open access publishing. We now proudly publish more than 200 world-class, open access, English language journals, spanning all scientific disciplines. Many of these are titles we publish in partnership with prestigious societies and academic institutions, such as the National Natural Science Foundation of China (NSFC).
END
Immune checkpoint inhibitors (ICIs), including antibodies against the programmed cell death protein 1/ligand 1 (PD-1/L1) effectively at boost the immune system; however, they may cause immune-related adverse events (irAEs). The predictive biomarkers and risk factors for irAEs have not been determined. Further, while activated T-cell infiltration of tissues is a hallmark of irAEs, it remains unclear if this is a prerequisite for irAEs development or merely a consequence.
To that end, a team of researchers from The ...
A recent study published in Engineering has unveiled a novel approach to generating functional organoids from human adult adipose tissue. This method, which bypasses traditional stem cell isolation and genetic manipulation, offers a more straightforward and scalable pathway for creating organoids that can be used in regenerative medicine and disease modeling.
The research, led by a team from the Shanghai Jiao Tong University School of Medicine and the Shanghai Institute for Plastic and Reconstructive Surgery, demonstrates the potential of human adult adipose tissue to differentiate into organoids representing all ...
Beijing’s metro system could soon carry express parcels alongside commuters during quiet periods, according to a new operational blueprint published in Engineering. Researchers from Beijing Jiaotong University and East China Jiaotong University have designed a joint optimization model that simultaneously plans train timetables, rolling-stock circulation, and cargo loading while allowing trains to skip stations with low demand. The approach, tested on the Yizhuang Line, cut operating costs by up to 25% compared with conventional fixed-composition, all-stop services.
The study addresses the rapid rise in urban freight volumes—already 15 million parcels a day in Beijing—by ...
Eye tumors are difficult to treat because drugs often struggle to reach tumor tissue without damaging healthy structures of the eye. A new review published in Biofunctional Materials summarizes how nanomaterials—extremely small, engineered particles—could improve both the diagnosis and treatment of eye cancers. The authors highlight recent advances, key challenges, and future directions for using nanotechnology to achieve safer and more precise eye tumor care.
Treating tumors inside the eye presents unique challenges. The eye is a small and highly sensitive organ, protected by biological barriers that limit drug penetration. While ...
Perovskite solar cells have garnered widespread attention as a low-cost, high-efficiency alternative to conventional silicon photovoltaics. However, defects in perovskite films impede charge transport, resulting in energy loss and compromised operational stability.
One solution to this problem is "passivation treatment"—a process that adds chemicals such as simple salts or organic molecules to the film. These small molecules or ions latch onto defects in the perovskite material, preventing the defects from interfering with electrical flow. Unfortunately, verifying the internal efficacyof various passivation ...
A research team led by Prof. Jihong Yu and Prof. Wenfu Yan (Jilin University, State Key Laboratory of Inorganic Synthesis and Preparative Chemistry) recently developed a potassium-based Cu-SSZ-39 catalyst (Cu-SSZ-39-K) with an increased concentration of Al in the 8-memberedrings. This design promotes the formation of more [Cu(OH)]+-Z species, resulting in nearly double the NOx conversion at 150-225°C and excellent retention of activity after hydrothermal ageing at 800°C. The work was published ...
In the realm of modern medicine, RNA-based therapies have emerged as a promising avenue, with significant advancements in metabolic diseases, oncology, and preventive vaccines. A recent article published in Engineering titled “The Future of AI-Driven RNA Drug Development” by Yilin Yan, Tianyu Wu, Honglin Li, Yang Tang, and Feng Qian, explores how artificial intelligence (AI) can revolutionize RNA drug development, addressing current limitations and offering new opportunities for innovation.
The article highlights ...
Platinum-ruthenium (PtRu) alloys are notable for their catalytic activity and durability in the hydrogen oxidation reaction (HOR). However, the mechanisms behind their superior performance compared to pure Pt are not fully understood. Research has identified many factors influencing HOR activity, including initial hydrogen binding energy, OH affinity driving the bifunctional mechanism, and factors like hydrogen bond (H-bond) network rigidity, connectivity, and interfacial water orientation. These complexities of the interfacial catalytic reaction present significant challenges in understanding the underlying mechanisms and achieving more precise enhancements ...
Full news release
Dalian, China-Sulfur, long feared as a “poison” that shuts down precious metal catalysts, can actually help them work better when used in just the right way, according to new research published in Chinese Journal of Catalysis.
A team led by Prof. Yunjie Ding at Dalian Institute of Chemical Physics, Chinese Academy of Sciences andProf. Xueqing Gong at Shanghai Jiao Tong University, has shown that a tiny, carefully tuned amount of sulfur can boost the speed and robustness of a key industrial reaction by up to twofold.
The reaction, called hydroformylation, adds carbon monoxide and hydrogen to simple molecules ...
Matías Gómez-Corrales, a recent biological sciences Ph.D. graduate from the University of Rhode Island, and his advisor, Associate Professor Carlos Prada, have published a paper in Nature Communications, revealing key mechanisms in speciation in corals and proposing a new hypothesis on the origin of species in the ocean.
Their new study examines how coral species form and contributes to long-standing questions in evolutionary biology about how marine biodiversity originates. The work builds on decades of ecological, reproductive, ...
