Human-Like Robotic Handwriting with Inverse Kinematic Using a Mobile Telescopic Arm

Abstract

In this paper, we present a method to enable the Stretch robot, developed by Hello Robots Inc., to perform letter writing tasks in a predefined configuration, using its telescopic arm gripper. Our method leverages inverse-forward kinematics techniques and is implemented in both Python and MATLAB for simulation purposes while, for real-world applications, we solely utilize Python, to interface directly with the Stretch robot through its SDK. To facilitate this capability, we modify the description of the Stretch, incorporating 14 distinct parts along with a virtual base component to enable controlled base movement. This virtual base acts as a prismatic joint, allowing for the adjustment of the base position. However, during the implementation of the inverse-forward kinematics in both simulation and real-world scenarios, we encountered an unexpected base movement. To address this issue, we propose the use of a threshold, mitigating such behavior. Moreover, we provide detailed insights into the results obtained with and without the proposed threshold. Specifically, we observe a significant reduction in precision errors when employing the threshold, with an average decreasing from 7.4 cm to 0.2 cm in real-world scenarios. This improvement underscores the efficacy of our proposed approach in enhancing the accuracy and stability of the Stretch robot during letter writing tasks.