The invention belongs to the technical field of automatic control, and particularly relates to a quadruped robot foot end ground slip rate estimation method based on deep learning, which adopts a deep neural network to judge the current terrain attribute category, and further adopts a GPS global positioning result to obtain a geological data set under the corresponding longitude and latitude, therefore, the accurate terrain slip rate of the same terrain attribute in different areas is obtained; according to the method, the deep learning technology is applied to estimation of the foot end ground slip rate of the quadruped robot, the problems that a traditional manual landing judgment strategy is complex, many parameters exist, and robustness is poor are solved, and the terrain attribute judgment neural network with high reliability and high sensitivity is finally obtained by collecting data for a long time for training; and the accurate slip rate is obtained by further combining a geological positioning data set, and the method has good generalization and migration capabilities on different robot configurations and platforms.
本发明属于自动控制技术领域,具体涉及一种基于深度学习的四足机器人足端地面滑移率估计方法,其采用深度神经网络实现对当前地形属性类别的判断,进一步采用GPS全球定位结果获取对应经纬度下的地质数据集,从而得到同样地形属性在不同区域内的准确地形滑移率;本发明将深度学习技术应用于四足机器人足端地面滑移率估计,解决了传统人工着地判断策略复杂、参数多、鲁棒性差的问题,通过长期收集数据进行训练最终得到高可靠、高灵敏性的地形属性判断神经网络,进一步结合地质定位数据集获取准确的滑移率,该方法在不同机器人构型和平台上具有良好的泛化和迁移能力。
Quadruped robot foot end ground slip rate estimation method based on deep learning
一种基于深度学习的四足机器人足端地面滑移率估计方法
2022-11-25
Patent
Electronic Resource
Chinese
IPC: | G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / B62D MOTOR VEHICLES , Motorfahrzeuge / G06K Erkennen von Daten , RECOGNITION OF DATA / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen / G06V |
Foot sole structure of quadruped robot and foot sole replacing method
European Patent Office | 2023
|Switchable wheel-foot composite mechanism of mechanical foot of quadruped robot
European Patent Office | 2023
|