Multicontact Robot Arm Path Planning for Opening Handleless Doors
针对无把手门，提出一种机器人通过多个接触点推开门板的方法，利用门板形状预计算可行接触姿态，减少在线搜索和碰撞检测，经仿真和实验验证有效。
Robotic manipulation in human-centric environments often requires interaction with everyday objects, including furniture doors. While previous research has primarily focused on opening doors equipped with handles, handleless furniture doors pose unique challenges due to the absence of a convenient grasping point. Existing methods for opening doors are usually based on single-point contact strategies in which a robot pulls on a door handle. When using these strategies, the space of possible solutions is limited by a specific contact point. We investigate an approach in which a robot opens a door by pushing the door panel using multiple contact points. This expands the potential solution space compared to single-contact methods and enables more versatile and robust manipulation strategies. In addition, we propose a strategy that takes advantage of the simple shape of the door panel and enables a precomputation of the feasible contact poses, significantly reducing the need to search for suitable contact points and tool orientations as well as collision checking in the online planning process. We conduct comprehensive evaluations of our method through both simulation and real-world experiments, demonstrating its effectiveness across a range of door configurations and properties.