基于激光数据特征提取的一般环境下实时定位方法

Real-time Positioning Methods Based on the Characteristics of the Laser Data to Extract the General Environment

实时定位技术是机器人在日常工作中完成各项任务的基础,为此本文给出了一种一般环境条件下基于激光数据特征提取的实时定位方法。该方法采用鲁棒的激光数据特征提取,通过对比实时样本和预定义模板的匹配程度以确认环境特征点。借助粒子滤波方法,利用里程计数据和当前观测到的环境特征点信息实时估计和验证机器人自身的位置和姿态。实验通过建立精确的机器人运动模型和激光数据观测模型,实现了仅用100个粒子就能进行机器人的实时定位。实验结果表明该方法能够准确提取环境中的疑似特征并依次实现了精确、快速的自定位。

Real-time positioning technology is the basis of the robot to complete various tasks in their daily work, this paper gives a general environmental conditions based on the characteristics of the laser data to extract real-time positioning. The method uses laser data of robust feature extraction, and to confirm the environment through the comparison of real-time sample and predefined templates match feature points. With the particle filter method, odometer data and the current environmental characteristics of the observed point information in real time to estimate and verify the position and attitude of the robot itself. Experiment through the establishment of accurate model of the robot movement and laser data observation model, only 100 particles will be able to conduct real-time positioning of the robot. The experimental results show that the suspected characteristics of this method can accurately extract the environment and in order to achieve accurate, fast, self-positioning.