Multiple and Gyro-Free Inertial Datasets.

Area of Science:

• Robotics and Autonomous Systems
• Navigation and Sensor Fusion
• Data Science and Machine Learning

Background:

• Inertial Navigation Systems (INS) are crucial for determining position, velocity, and orientation using accelerometers and gyroscopes.
• Existing INS research increasingly integrates data-driven methods for enhanced accuracy and efficiency.
• A significant gap exists in publicly available datasets for specialized architectures like gyro-free INS (GFINS) and multiple inertial measurement units (MIMU).

Purpose of the Study:

• To address the lack of GFINS and MIMU datasets.
• To provide a comprehensive resource for researchers studying novel INS architectures.
• To facilitate the development and evaluation of advanced navigation algorithms.

Main Methods:

• Designed and recorded novel GFINS and MIMU datasets.
• Utilized 54 inertial sensors organized into nine inertial measurement units.
• Configured sensors in three distinct arrangements mounted on a mobile robot, passenger car, and turntable.

Main Results:

• Collected 45 hours of inertial data with corresponding ground truth trajectories.
• The dataset enables the definition and evaluation of various MIMU and GFINS configurations.
• Data is made freely accessible via a figshare repository to encourage further research.

Conclusions:

• The newly created GFINS and MIMU datasets fill a critical gap in the research community.
• This resource will accelerate innovation in data-driven INS, particularly for gyro-free and multi-sensor systems.
• Availability of this dataset promotes reproducible research and the development of next-generation autonomous platforms.