基于机器学习的四足移动机器人视觉导航方法

传统四足移动机器人视觉导航方法以导航监测为主,视觉定位的地图实时性较低,影响了视觉导航效果。为此,设计基于机器学习的四足移动机器人视觉导航方法。该方法基于机器学习提取四足移动机器人的视觉特征,从机器人实时获取的视觉图像中提取导航信息,并通过机器学习分析提供视觉环境的三维结构信息。实验结果表明,该方法能够快速地找到最优路径,完成视觉导航任务,减少机器人能量消耗,有效提高整体效率。

Motion design of a hybrid wheeled/legged robot for lunar exploration

The robot consists of a quadruped mechanism and two active dual-wheel casters possesses the advantages of wheeled and legged mechanism, and can quickly move on the relatively plane ground with the wheeled mechanism, and can walk on the extremely uneven terrain with the legged mechanism. The effectiveness of the motion design of the hybrid robot is iHustrated by simulation results.