Video data for the cognitive mapping process of NeuroBayesSLAM system

Area of Science:

• Robotics
• Neuroscience
• Computer Vision

Background:

• Simultaneous Localization and Mapping (SLAM) is crucial for mobile robot navigation in unknown environments.
• Existing SLAM systems often lack the sophisticated spatial awareness seen in biological systems.
• Neurobiological principles offer a promising avenue for developing more intelligent robotic navigation.

Purpose of the Study:

• To provide raw video data demonstrating the NeuroBayesSLAM system's mapping and localization capabilities.
• To showcase the integration of neural activities (head-direction and grid cells) and visual data in a cognitive mapping process.
• To support research into transferring mammalian brain spatial memory mechanisms to robotic systems.

Main Methods:

• Utilized two public datasets: St Lucia suburb and iRat Australia.
• Recorded cognitive mapping processes using screen recording software on Ubuntu Linux.
• Captured neural activities, local view templates, and experience maps within the SLAM system.

Main Results:

• Presented four raw video files detailing the NeuroBayesSLAM system's performance.
• Demonstrated the system's ability to perform mapping and localization on diverse datasets.
• Visualized the internal states including neural correlates and the constructed experience map.

Conclusions:

• The provided data supports the development of biologically inspired SLAM systems.
• Highlights the potential of multisensory integration from mammalian brains for robot navigation.
• Enables further research into advanced cognitive mapping for both indoor and outdoor robotic applications.