VIO-DualProNet: Visual-Inertial Odometry with Learning Based Process Noise Covariance DOI Creative Commons
Dan Solodar, Itzik Klein

arXiv (Cornell University), Год журнала: 2023, Номер unknown

Опубликована: Янв. 1, 2023

Visual-inertial odometry (VIO) is a vital technique used in robotics, augmented reality, and autonomous vehicles. It combines visual inertial measurements to accurately estimate position orientation. Existing VIO methods assume fixed noise covariance for the uncertainty. However, determining real-time variance of sensors presents significant challenge as uncertainty changes throughout operation leading suboptimal performance reduced accuracy. To circumvent this, we propose VIO-DualProNet, novel approach that utilizes deep learning dynamically real-time. By designing training neural network predict using only sensor measurements, integrating it into VINS-Mono algorithm, demonstrate substantial improvement accuracy robustness, enhancing potentially benefiting other VIO-based systems precise localization mapping across diverse conditions.

Язык: Английский

Adaptive Kalman-Informed Transformer DOI Creative Commons
Nadav Cohen, Itzik Klein

Engineering Applications of Artificial Intelligence, Год журнала: 2025, Номер 146, С. 110221 - 110221

Опубликована: Фев. 12, 2025

Процитировано

1
VIO-DualProNet: Visual-inertial odometry with learning based process noise covariance DOI
Dan Solodar, Itzik Klein

Engineering Applications of Artificial Intelligence, Год журнала: 2024, Номер 133, С. 108466 - 108466

Опубликована: Апрель 27, 2024

Язык: Английский

Процитировано

5
A survey on Ultra Wide Band based localization for mobile autonomous machines DOI Creative Commons
Ning Xu, Mingyang Guan, Changyun Wen

и другие.

Journal of Automation and Intelligence, Год журнала: 2025, Номер unknown

Опубликована: Фев. 1, 2025

Язык: Английский

Процитировано

0
Underwater localization system for marine seismic airgun arrays validated through robotics DOI Creative Commons
Ulises Tronco Jurado, Peter Wilson, Philippe Blondel

и другие.

International Journal of Intelligent Robotics and Applications, Год журнала: 2025, Номер unknown

Опубликована: Март 6, 2025

Язык: Английский

Процитировано

0
DCNet: A data-driven framework for DVL calibration DOI Creative Commons

Zeev Yampolsky,

Itzik Klein

Applied Ocean Research, Год журнала: 2025, Номер 158, С. 104525 - 104525

Опубликована: Март 24, 2025

Язык: Английский

Процитировано

0
A review on Control momentum Gyroscopic Stabilization for intelligent balance Assistance in Electric Two-wheeler DOI Creative Commons

Prithvi Raj Pedapati,

C. Ramesh Kumar

Results in Engineering, Год журнала: 2025, Номер unknown, С. 105069 - 105069

Опубликована: Апрель 1, 2025

Язык: Английский

Процитировано

0
Snake-inspired mobile robot positioning with hybrid learning DOI Creative Commons
Aviad Etzion, Nadav Cohen, Ofer Levi

и другие.

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Май 4, 2025

Mobile robots are used in various fields, from deliveries to search and rescue applications. Different types of sensors mounted on the robot provide accurate navigation and, thus, allow successful completion its task. In real-world scenarios, due environmental constraints, frequently relies only inertial sensors. Therefore, noises other error terms associated with readings, solution drifts time. To mitigate drift, we propose MoRPINet framework consisting a neural network regress robot's travelled distance. this end, require mobile maneuver snake-like slithering motion encourage nonlinear behavior. was evaluated using dataset 290 minutes recordings during field experiments showed an improvement 33% positioning over state-of-the-art methods for pure navigation.

Язык: Английский

Процитировано

0
Robotic Sequencing for Intelligent Mission Management DOI Open Access

Ioannis Giachos,

Christina Paschaliori,

Evangelos C. Papakitsos

и другие.

WSEAS TRANSACTIONS ON INFORMATION SCIENCE AND APPLICATIONS, Год журнала: 2025, Номер 22, С. 440 - 449

Опубликована: Май 28, 2025

In this paper, an algorithm is presented that enables autonomous movement of a robotic vehicle based on intelligent Human-Robot Interface (iHRI). This developing system with dialogue and understanding capabilities in limited Greek vocabulary. additional feature work transforms the into research tool will participate missions for actions at local points within open area, either land or sea. These might include, example, sampling soil, water, air. The capable navigating through area by recognizing various minimal voice commands from human operator. Furthermore, system, (iHRI) it possesses, calculating order to place. correct considered be visiting nearest point first. process called sequencing. not followed during motion planning, only those where command specifies time parameter. exactly capability provided iHRI.

Язык: Английский

Процитировано

0
GHNet: Learning GNSS Heading From Velocity Measurements DOI
Nitzan Dahan, Itzik Klein

IEEE Sensors Journal, Год журнала: 2024, Номер 24(4), С. 5195 - 5202

Опубликована: Янв. 8, 2024

By utilizing global navigation satellite system (GNSS) position and velocity measurements, the fusion between GNSS inertial (INS) provides accurate robust information. When considering land vehicles, like autonomous ground off-road or mobile robots, a GNSS-based heading angle measurement can be obtained used in parallel to bind drift. Yet, at low vehicle speeds (less than 2 m/s) such model-based (MB) fails provide satisfactory performance. This article proposes GHNet, deep learning framework capable of accurately regressing for vehicles operating speeds. GHNet utilizes only current measurement, from single receiver, regression task. It is shallow network its ability reduce noise capture nonlinear behavior. We demonstrate that outperforms MB approach simulation experimental datasets. applied any type as passenger cars, robots

Язык: Английский

Процитировано

1
VIO-DualProNet: Visual-Inertial Odometry with Learning Based Process Noise Covariance DOI Creative Commons
Dan Solodar, Itzik Klein

arXiv (Cornell University), Год журнала: 2023, Номер unknown

Опубликована: Янв. 1, 2023

Visual-inertial odometry (VIO) is a vital technique used in robotics, augmented reality, and autonomous vehicles. It combines visual inertial measurements to accurately estimate position orientation. Existing VIO methods assume fixed noise covariance for the uncertainty. However, determining real-time variance of sensors presents significant challenge as uncertainty changes throughout operation leading suboptimal performance reduced accuracy. To circumvent this, we propose VIO-DualProNet, novel approach that utilizes deep learning dynamically real-time. By designing training neural network predict using only sensor measurements, integrating it into VINS-Mono algorithm, demonstrate substantial improvement accuracy robustness, enhancing potentially benefiting other VIO-based systems precise localization mapping across diverse conditions.

Язык: Английский

Процитировано

0