Abstract: Mobile robot localization has been considered to be an important task in the ?eld of robotics research. This paper proposes Recurrent Convolutional Neural Networks-Based Mobile Robot Localization（RCNN-MRL） algorithm. RCNN（Recurrent Convolutional Neural Networks） is a neural networks model that combines Convolutional Neural Networks（CNN） and Recurrent Neural Networks（RNN）, and RCNN-MRL estimates the self-position from the first person view captured by a camera on a robot by using RCNN. Specifically, it uses a regression model for localization by using RCNN capable of processing consecutive images. This paper uses simulated environments where a two-wheel robot moves randomly, and analyzes the performance of localization. The experiments show that RCNN model can estimate the self-position of the robot.

Key words: mobile robot localization, first person view, time series image, convolutional neural networks, two-wheel robot

摘要： 移动机器人定位已成为机器人研究的重要任务。提出基于递归卷积神经网络的移动机器人定位（Recurrent Convolutional Neural Networks-Based Mobile Robot Localization，RCNN-MRL）算法。递归卷积神经网络（Recurrent Convolutional Neural Networks，RCNN）结合卷积神经网络（Convolutional Neural Networks，CNN）和递归神经网络（Recurrent Neural Networks，RNN）的特性，并依据机器人上嵌入的照相机拍摄的第一人称视角图像，RCNN-MRL算法利用RCNN实现自主定位。具体而言，先通过RCNN有效地处理多个连续图像，再利用RCNN作为回归模型，进而估计机器人位置。同时，设计双轮机器人移动，获取多个时间序列图像信息。最后，依据双轮机器人随机移动建立仿真环境，分析机器人定位性能。实验数据表明，提出的RCNN模型能够实现自主定位。

关键词: 移动机器人定位, 第一人称视角, 时间序列图像, 递归卷积神经网络, 双轮机器人