Graph-based Maps Formation for Mobile Robots by Hidden Markov Models

摘要

The present paper proposes a probabilistic approach to recognizing the environment of a mobile robot and to generate a graph-based map based on the estimation of Hidden Markov Models (HMMs). This is because recognition of the environment based on a short interval of data is not enough when sensory signals are corrupted by noise. Graph-based maps are effective in decreasing the computational cost. Two methods for constructing graph-based maps are proposed. The former is to estimate HMMs based on quantized sensory-motor signals. The latter is to estimate HMMs based on a sequence of labels obtained by modular network SOM (mnSOM). The resulting sequence of HMMs can be converted into a graph-based map in a straightforward way. Simulation results demonstrate that the proposed method is able to construct graph-based maps effectively, and to perform goal seeking efficiently.