Purpose The purpose of this paper is to propose a hybrid position/force control scheme using force and vision for docking task of a six degrees of freedom (6-dof) hydraulic parallel manipulator (HPM). Design/methodology/approach The vision system consisted of a charge-coupled device (CCD) camera, and a laser distance sensor is used to provide globe relative position information. Also, a force plate is used to measure local contact forces. The proposed controller has an inner/outer loop structure. The inner loop takes charge of tracking command pose signals from outer loop as accurate as possible, while the outer loop generates the desired tracking trajectory according to force and vision feedback information to guarantee compliant docking. Several experiments have been performed to validate the performance of the proposed control scheme. Findings Experiment results show that the system has good performance of relative position tracking and compliant contact. In whole docking dynamic experiment, the amplitudes of contact forces are well controlled within 300 N, which can meet perfectly the requirement of the amplitude being not more than 1,000 N. Originality/value A hybrid position/force control scheme using force and vision is proposed to make a 6-dof HPM dock with a moving target object compliantly.

No abstract

No abstract

This article proposes a forward kinematics algorithm based on closed-loop feedback solution for real-time pose estimation of 6-degree-of-freedom parallel mechanisms. First, a feedback forward kinematics algorithm with second-order error dynamics built in joint space is presented and its convergence is analyzed. Taking a Stewart platform as an example, fast convergence and high solving accuracy of the algorithm are shown in simulation. However, while applying the algorithm to forward kinematics of a 6-degree-of-freedom redundantly actuated parallel mechanisms, solving errors will be growing slowly and non-convergent, which is caused by redundancy character of the parallel mechanisms. Then a modified algorithm, of which second-order error dynamics is built in Cartesian space, is developed. Considering velocity information is usually difficult to measure or estimate accurately in reality, the performance of feedback forward kinematics algorithm without joint velocity feedforward is also evaluated. Finally, the algorithm is implemented into real-time degree of freedom control of the redundant parallel mechanisms which is calibrated by a coordinate measuring machine. Experiment results show that the proposed algorithm can solve the pose with relatively high accuracy under static conditions and is effective and feasible for real-time pose estimation of parallel mechanisms.

No abstract