.png){kind=icon}
.png){kind=icon}
Return
Map construction method for granary inspection robots
.png){kind=icon}
PRE
AI
DOI：10.1080/15397734.2025.2505202.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This article studies the use of inspection robots to construct environmental maps of domestic bungalow granaries, employing mobile robots to build different granary maps under varying conditions. The 2D maps of the granary are provided to the robots for inspection tasks, helping them reach designated locations and replacing manual inspection methods. During the construction of 2D maps, when new obstacles are discovered or the robot's pose significantly deviates, the construction switches to 3D maps for more detailed perception and analysis of the environment, ensuring the safety of grain storage. High-precision 2D maps are crucial for the robots to navigate effectively to specified locations during inspections. To address the issues of particle degradation and loss of particle diversity in the traditional FastSLAM algorithm, which lead to reduced accuracy in robot localization and map construction, an improved FastSLAM algorithm based on the hunter-prey optimization (HPO) is proposed. The HPO algorithm, improved with chaotic strategy and Levy flight strategy, optimizes FastSLAM by enhancing particle prediction accuracy. Simulation studies conducted on the MATLAB platform show that the IHPO-FastSLAM algorithm has higher pose accuracy and landmark estimation accuracy compared to the traditional FastSLAM algorithm. Finally, the improved algorithm was tested in a constructed granary environment, and comparison results of the constructed maps demonstrate that the improved algorithm has higher mapping accuracy. This research contributes to the application of mobile robots in granaries, advancing automation and intelligence in the grain storage industry.
Keywords:
Constructing maps
FastSLAM algorithm
granary inspection
hunter-prey optimization algorithm
mobile robot
Journal
IF:
2.9
Papers:
704
Citations:
3.8K