Search

In this paper, we propose a fast feature points-based object tracking method for robot grasp. In the detection phase, we detect the object with SIFT feature points extraction and matching. Then we compute the object's image position with homography constraints and set up an interest window to accommodate the object. In the tracking phase, we only focus on the interest window, detecting feature points from the window and updating the window's position and size. Our method is of special practical meaning in the case of service robot grasp. Because when the robot grasps the object, the object's image size is usually small relative to the whole image, it is unnecessary to detect the whole image. On the other hand, the object is partially occluded by the robot gripper. SIFT is good at dealing with occlusion, but it is time consuming. Hence, by combining SIFT and an interest window, our method gains the ability to deal with occlusion and can satisfy the real-time requirements at the same time. Experiments show that our method exceeds several leading feature points-based object tracking methods in real-time performance.

In the field of machine vision, the multi-solution problem of the perspective-three-point (P3P) camera pose estimation problem has been widely researched. Many researchers have tried to find unique solutions that will enable the use of the P3P method in practical applications such as the calibration of external parameters of a camera. Gao found a special case that is very simple and intuitive, but can only be used for a camera with a very large field of view. Therefore, it did not attract much attention. The purpose of this study was to analyse Gao's special case and use it for calibrating the external parameters of an omnidirectional camera, which is used widely because of its very large field of view. We first verify Gao's special case and extend its application. We then propose an iterative method for finding the unique solution. After this, we perform a simulation experiment to determine the effect of the configuration of the three control points on the accuracy of the unique solution. Finally, we use an application to examine the efficiency of the method.

A more suitable motion planning method for an omni-directional mobile robot (OMR), an improved APF method (iAPF), is proposed in this paper by introducing the revolving factor into the artificial potential field (APF). Accordingly, the motion direction derived from traditional artificial potential field (tAPF) is regulated. The maximum velocity, maximum acceleration and energy consumption of the OMR moving in different directions are analyzed, based on the kinematic and dynamic constraints of an OMR, and the anisotropy of OMR is presented in this paper. Then the novel concept of an Anisotropic-Function is proposed to indicate the quality of motion in different directions, which can make a very favorable trade-off between time-optimality, stability and efficacy-optimality. In order to obtain the optimal motion, the path that the robot can take in order to avoid the obstacle safely and reach the goal in a shorter path is deduced. Finally, simulations and experiments are carried out to demonstrate that the motion resulting from the iAPF is high-speed, highly stable and highly efficient when compared to the tAPF.

This paper presents a rapid traversability assessment approach based on an extended 2.5D grid-based representaion of the rough terrain. Stereo vision system is used to perceive the environment surrounding robot. Conventional 2D, 3D and other 2.5D grid maps determine the traversability indices of the grids directly from the sensor feedback, while our approach attempts to address the indices of terrain from multiple grids instead. By analyzing the properties of multiple grids that the robot is to traverse, passable grids are distinguished, which also takes the robot's size into account. Fuzzy logic framework is applied to extract traversabiltiy indices from the terrain characteristics. A soccer robot equipped with a stereo vision system is adopted for experiments. The results show that our map is capable of speeding the process of traversability assessment and providing an autonomous mobile robot with a appropriate representation of 3D uneven terrain profile.

This paper presents a rapid traversability assessment approach based on an extended 2.5D grid-based representaion of the rough terrain. Stereo vision system is used to perceive the environment surrounding robot. Conventional 2D, 3D and other 2.5D grid maps determine the traversability indices of the grids directly from the sensor feedback, while our approach attempts to address the indices of terrain from multiple grids instead. By analyzing the properties of multiple grids that the robot is to traverse, passable grids are distinguished, which also takes the robot's size into account. Fuzzy logic framework is applied to extract traversabiltiy indices from the terrain characteristics. A soccer robot equipped with a stereo vision system is adopted for experiments. The results show that our map is capable of speeding the process of traversability assessment and providing an autonomous mobile robot with a appropriate representation of 3D uneven terrain profile.