Study on Optimization of Mapping Method for Multi-Layer Cage Chicken House Environment

This study delves into the mapping method for the navigation system of a chicken coop disinfection robot. It systematically analyzes the problems of insufficient effective particle count, high particle repetition rate in environmental map information, and penetration phenomenon in traditional SLAM laser point cloud mapping technology in chicken coop environments containing multiple layers of chicken cages. To address these issues, this paper proposes an optimized mapping method based on an improved ICP algorithm, significantly improving the laser point clouds' registration performance. At the same time, by limiting the sampling of environmental map information particles within a specific range and optimizing the screening based on the predicted distribution of particle poses and the matching degree of the map, the diversity of particles and the accuracy of map information have been effectively improved. The field experiment results show that the maximum error of this method on the chicken coop environment map does not exceed 3.5 cm. The environmental characteristics of the chicken coop are maximally preserved, which verifies the effectiveness and robustness of this method and provides a scientific basis for the mapping method of the livestock and poultry breeding robot navigation system.