English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We present BioSLAM, a lifelong SLAM framework for learning various newappearances incrementally and maintaining accurate place recognition forpreviously visited areas. Unlike humans, artificial neural networks suffer fromcatastrophic forgetting and may forget the previously visited areas whentrained with new arrivals. For humans, researchers discover that there exists amemory replay mechanism in the brain to keep the neuron active for previousevents. Inspired by this discovery, BioSLAM designs a gated generative replayto control the robot's learning behavior based on the feedback rewards.Specifically, BioSLAM provides a novel dual-memory mechanism for maintenance:1) a dynamic memory to efficiently learn new observations and 2) a staticmemory to balance new-old knowledge. When combined with a visual-/LiDAR- basedSLAM system, the complete processing pipeline can help the agent incrementallyupdate the place recognition ability, robust to the increasing complexity oflong-term place recognition. We demonstrate BioSLAM in two incremental SLAMscenarios. In the first scenario, a LiDAR-based agent continuously travelsthrough a city-scale environment with a 120km trajectory and encountersdifferent types of 3D geometries (open streets, residential areas, commercialbuildings). We show that BioSLAM can incrementally update the agent's placerecognition ability and outperform the state-of-the-art incremental approach,Generative Replay, by 24%. In the second scenario, a LiDAR-vision-based agentrepeatedly travels through a campus-scale area on a 4.5km trajectory. BioSLAMcan guarantee the place recognition accuracy to outperform 15\% over thestate-of-the-art approaches under different appearances. To our knowledge,BioSLAM is the first memory-enhanced lifelong SLAM system to help incrementalplace recognition in long-term navigation tasks.

BioSLAM: A Bio-inspired Lifelong Memory System for General Place Recognition