| 基于最小二乘支持向量机的移动机器人导航 |
| |
| 引用本文： | 侯艳丽.基于最小二乘支持向量机的移动机器人导航[J].电子设计工程,2011,19(23):11-12,15. |
| |
| 作者姓名： | 侯艳丽 |
| |
| 作者单位： | 商丘师范学院计算机与信息技术学院,河南商丘,476000 |
| |
| 基金项目： | 河南省科技厅基础与前沿技术研究项目，河南省科技厅科技攻关研究项目(112102210210):河南省教育厅自然科学基金项目 |
| |
| 摘 要： | 针对未知环境下的移动机器人导航,提出将最小二乘支持向量机与强化学习相结合的导航方法。首先以移动机器人CASIA-I和它的工作环境为实验平台,确定出强化学习的回报函数;然后利用基于滚动窗的最小支持向量机解决强化学习中的泛化问题。最后对所提方法进行了实验,实验结果表明所提方法能够避免导航陷入局部极小,并对未知环境具有较强的适应性。
|
| 关 键 词： | 移动机器人 强化学习 支持向量机 导航 泛化 |
Mobile robot navigation based on least squares support vector machine |
| |
| HOU Yan-li.Mobile robot navigation based on least squares support vector machine[J].Electronic Design Engineering,2011,19(23):11-12,15. |
| |
| Authors: | HOU Yan-li |
| |
| Affiliation: | HOU Yan-li(Department of Computer and Informatiion Technology,Shangqiu Teachers College,Shangqiu 476000,China) |
| |
| Abstract: | According to the problem of mobile robot navigation in the unknown environment, a navigation method using reinforcement learning based on least squares support vector machine （LS-SVM） is proposed. Firstly, according to CASIA-I and its working environment, an approach is proposed to determine the reward/penalty function of reinforcement learning. Secondly, a LS-SVM based on a time windows is used to solve the generalization problem. At last, experimental results are included to show that the proposed method can avoid getting into local minimum and it is adaptive to unknown environments. |
| |
| Keywords: | mobile robot reinforcement learning SVM navigation generalization |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
