无环境信息下多尺度网格细胞群空间表征模型

针对无环境信息条件下网格细胞空间表征问题, 提出一种多尺度网格细胞群的空间表征模型。通过减少吸引子网络模型网格细胞的偏好朝向, 使网格细胞具有特定的感应方向, 仅对特定方向上的速度产生响应。基于改进的吸引子网络模型, 构建多尺度网格细胞群, 并对空间进行表征。仿真结果表明, 改进后的网格细胞模型只响应相应方向的速度, 准确整合速度信息, 表征部分未探索区域, 且适用于无环境信息条件下一维、二维空间。在100 m×100 m的区域, 表征结果的误差不超过0.2 m, 证明了该模型能够扩大网格细胞的表征范围, 提高表征精度与表征能力。

Spatial representation model based on multi-scale grid cell population without environment information

The spatial representation model based on multi-scale grid cell population is proposed to solve the problem of using grid cells for spatial representation without environmental information.By reducing the preference orientation,the grid cells have a specific inductive direction and only respond to the velocity in a specific direction.We construct a multi-scale grid cells model based on the improved model to represent the environment.The results show that the improved grid cell model only responds to the velocity in the corresponding direction,integrates the velocity information accurately,represents part of the unexplored region,and is applicable to one-dimensional and two-dimensional space without environmental information.In the area of 100 m×100 m,the results’error does not exceed 0.2 m.It is proved that the improved model expands the scope of characterization of grid cells and improves the representation’s accuracy and ability.