A Position Representation Method Based on the Localization Mechanism of Rat Hippocampus

The current methods for autonomous positioning cannot realize the integration of perception, composition and behavior. In order to solve this problem, inspired by the navigation and localization mechanism of the rat brain, this paper links the spatial information with the discharge characteristics of brain navigation cells, and proposes a position representation method based on the localization mechanism of rat hippocampus. The supervised learning mechanism is introduced to establish the spatial location memory model. We use the radial basis function (RBF) network to link the extracted landmarks with the place cells map. The location of the agent in space can be determined by the firing cells of the current location. The simulation experiment results show that the method in this paper is feasible to represent the position of the agent. The average positioning error between the estimated position and the real position is 2.35m, and the positioning error gradually decreases as the exploration time increases. Compared with the traditional methods, our method can directly judge the position and take the next action by using the place cells map after identifying the landmarks, which realizes the integration of the perception, composition and behavior to some extent.