Zhang Xuebo

A fast homography decomposition technique for visual servo of mobile robots

A fast homography decomposition algorithm for visual servo of mobile robots is presented. Motion constraints of the mobile robot are exploited and thus the complexity of the algorithm decreases significantly by avoiding the singular value decomposition (SVD) step in most cases, which is potentially competitive in such applications as visual servo task executed at the video rate. Moreover, the ambiguity problem, which is troublesome for general approaches on the unconstrained configuration such as manipulators, does not exist for the mobile robot except when the reference target plane is perpendicular to the horizontal ground. Besides, the proposed algorithm can also be exploited in other related areas such as vision-based localization, provided that the camera undergoes a planar motion. Simulation results considering the pixel noise are provided to demonstrate the performance of the proposed method.

Inverse Kinematics Analysis for a Mobile Manipulator with Redundant DOFs

A mobile manipulator is a manipulator mounted on a mobile robot. This paper proposed an inverse kinematics analysis method for a mobile manipulator with redundant degrees of freedom (DOFs). The DOFs of the end-effecter are assigned properly among the mobile robots 3 DOFs and the manipulators 5 DOFs: the rotation freedoms of the mobile robot and the manipulators 1st, 4th and 5th joint contribute to the orientation of the end-effector; while the two translation freedoms of the mobile robot and the rotation freedoms of the manipulators 2nd and 3rd joint contribute to the position of the end-effecter. Based on this DOFs assignment method, two restrictions are introduced for inverse kinematics calculation without changing the work space of the end-effector. Experiment results are included to demonstrate the performance of the results obtained by this inverse kinematics analysis. In these experiments, based on the information of the targets position and orientation obtained from an on-board CCD camera, the mobile manipulator is controlled to catch the target by utilizing the inverse kinematics analysis strategy proposed in this paper.

Dynamic hand gesture recognition using HMM-BPNN model

This paper proposes a new method, combining Hidden Markov Model and BP Neutral Network (called HMM-BPNN Model), to solve the problem of dynamic hand gesture recognition. Specifically, first extract the information of hand gesture feature from 3D depth image using the finger tracking module of Intel perceptual equipment; second, the resulting information of hand gesture feature is modeled with the method of Hidden Markov Model (HMM); third, BP Neutral Network (BPNN), as the classifier, recognizes the inputting dynamic hand gesture. Finally, the results of simulation and experiment verify the feasibility of the proposed method in this paper.

A Camera Calibration Method Based on 2D Homography

In this paper, a new method for camera calibration is presented based on 2D homography. First, intrinsic parameters are separated into two parts which are then solved out respectively using the least squares method. Compared with other techniques, the proposed method can obtain a much better initial guess which will subsequently improve the efficiency of the nonlinear-optimization algorithm and the precision of the final result. For the convenience of different applications in practice, two objective functions are proposed for nonlinear optimization, and correspondingly, two different kinds of radical distortion coefficients are obtained. While the former is well suited to infer 2D image coordinates from 3D information, the latter is more helpful to obtain 3D information from 2D image signals. Simulation results are provided to demonstrate the superior performance of the method.

Homography-based robust pose estimation method

A method for pose estimation based on homography and its robustness against image noise is investigated in this paper. This technique was implemented on the mobile robot to verify the efficiency of visual servo. Firstly, homography matrix could be calculated by comparing the corresponding target points of an object from two different camera images. Then, by decomposing the homography matrix through Faugeraspsila algorithm, the camera pose information could be obtained under the unique construction condition. Particularly, we tested the robustness of the proposed method against image noise. Finally, this method was implemented on monocular camera-based vision systems for navigating a mobile robot. Simulation results demonstrate that the proposed method show advantages such as unique decomposition result, high robust against image noise, and so on.