华中科技大学学报（自然科学版）2011，Vol.39Issue(z2)：298-301,4.
一种仿人机器人行走距离预测方法
A walking distance prediction method for humanoid robot
摘要
Abstract
As environment and robot itself factors such as backlash in the joints and foot slippage affects the robot, and deviation arises between predicted distance and actual walking distance during walking. A method was proposed that can predict walking distance based on adaptive neuro-fuzzy inference system (ANFIS) for a humanoid robot. Relevant data of NAO's walking distance was acquired. Grid partitioning method was used to construct adaptive fuzzy inference system to predict the actual walking distance. Combinatorial optimization of least square algorithms and BP (back propagation) algorithm was used to train system. By analyzing and testing the trained system, the simulation results show the method is effective.关键词
仿人机器人/自适应神经模糊推理系统/网格分割/BP算法/最小二乘算法Key words
humanoid robot/ adaptive neuro-fuzzy inference system/ grid partitioning/ BP algorithm/ least square algorithm分类
信息技术与安全科学引用本文复制引用
许宪东,洪炳镕,朴松昊,刘强..一种仿人机器人行走距离预测方法[J].华中科技大学学报（自然科学版）,2011,39(z2):298-301,4.基金项目
国家自然科学基金资助项目(61075077) （61075077）
黑龙江省教育厅科学技术研究资助项目(12511446). （12511446）
