Peg-in-Hole Assembly Based on Six-Legged Robots with Visual Detecting and Force Sensing

Abstract

Manipulators with multi degree-of-freedom (DOF) are widely used for the peg-in-hole task. Compared with manipulators, six-legged robots have better mobility performance apart from completing operational tasks. However, there are nearly no previous studies of six-legged robots performing the peg-in-hole task. In this article, a peg-in-hole approach for six-legged robots is studied and experimented with a six-parallel-legged robot. Firstly, we propose a method whereby a vision sensor and a force/torque (F/T) sensor can be used to explore the relative location between the hole and peg. According to the visual information, the robot can approach the hole. Next, based on the force feedback, the robot plans the trajectory in real time to mate the peg and hole. Then, during the insertion, admittance control is implemented to guarantee the smooth insertion. In addition, during the whole assembly process, the peg is held by the gripper and attached to the robot body. Connected to the body, the peg has sufficient workspace and six DOF to perform the assembly task. Finally, experiments were conducted to prove the suitability of the approach.

Hand Guiding a Virtual Robot Using a Force Sensor

“…A very interesting way of using an FT sensor is presented by Zhao et al (2020). In their work, the peg-in-hole approach for a six-parallel-legged robot is proposed.…”

Stability-Guaranteed and High Terrain Adaptability Static Gait for Quadruped Robots

“…In addition, an exploratory gait planning on uneven terrain with touch sensing is proposed as shown in Figure 5 . Exploratory gait incorporates force sensing [ 48 ] on the foot, and it is a planning method in which a robot can adjust gait planning actively when its foot meets obstacles during swing phase. Four three-dimensional force sensors are installed at the foot end-effectors of quadruped robot to achieve the proposed touch sensing.…”

A Coarse-to-Fine Method for Estimating the Axis Pose Based on 3D Point Clouds in Robotic Cylindrical Shaft-in-Hole Assembly

“…Particularly, in shaft-in-hole assembly, challenges are introduced from the location of the target hole and the smooth insertion of shaft-in-hole. Concerning the location of targets, previous methods relying on pure tactile/force sense [6,7] or hybrid vision-tactile sense [8] can work successfully. Typically, as illustrated in Figure 1a, with the rough pre-positioning of the hole, robots employ force/torque (F/T) sensors to further locate holes exactly to insert the shaft.…”

“…With admittance control, robots can adaptively adjust the pose of the end effector toward the hole bottom during the insertion process of shaft-in-hole. Zhao et al [8] showed the successful implementation of admittance control in peg-in-hole assembly.…”

“…To the best of our knowledge, although some researchers have solved the peg-in-hole assembly [6][7][8], which is similar to the shaft-in-hole one, their methods are more dependent on force sensing than vision, limiting the efficiency of assembly. As shown in Figure 1, intuitively, it is time-consuming to adjust the shaft pose relative to the hole according to the contact force feedback.…”