Papers by Keyword: Motion Capture

Abstract: The analysis and implementation of this paper is applied to a real game boxing machine. The contents of this game boxing machine are consists of three parts that are the UNITY, programming and hardware development. Therefore, in order to have RPG role-play boxing game inside of real game boxing machine, this paper firstly analyzes user requirement and then implement the combined-use in the traditional large boxing machines that will let the players to face the need considering how to fight the contents of RPG role-play boxing game. Moreover, in order to save more time to make 3D animations, motion capture equipment is used called MOCAP to generate characteristics, hero, monsters, and beautiful girls. All body movements of characteristics inside of game boxing machine captured by MOCAP are tied to the body of the head, limbs and torso. Furthermore, those body movements will PhaseSpace to our servers via the Internet passing out after the signal processing MotionBuilder software for recording and output operation. The output file can be transformed into the skeleton and action graphic model and using MAYA software can also do the same way. All procedures in this RPG role-play boxing game design have been successfully interacted both machine side and play side with fun and game new life, and to integrate the actual situation of the game for the future.

Abstract: For Motion Capture in Gait Analysis using Known Spherical Markers one simple direct approach is to compute the projection of the Marker Center using its projection in the Pixel Plane and based on it to find the location of the Marker on the line that connects the Marker Center Projection and the camera Focal Point. For various positions of the Marker in the workspace the exact image of the marker is computed using a genuine approach and compute back the approximation of the position based on the generated image. Various algorithms are taken in consideration and finally the results are assessed from the point of view of Gait Analysis and two directions for calculus improvement are identified.

Abstract: This paper investigates the problem of composed LED light sources placement for a particular case of motion capture with reflexive spherical markers in Gait Analysis in a controlled lighted environment. After the addressed problem is introduced the way the rays are propagated from the individual LEDs to the pixel array of video camera is computed. Based on this calculus the number of ray hits on distinct pixels is computed for a marker placed in the workspace. Next by varying some scene parameters we obtain the total number of hits on the overall set of makers placed in the scene and based on this number we obtain finally the best placement for the composed LED light sources using an iterative approach.

Abstract: In recent years, robots have found extensive applications in automating repetitive, defined, position dependent tasks such as painting and material handling. However, continuous contact type tasks (such as finishing, deburring and grinding) that require both position and force control are still carried out manually by skilled labor. Majorly, because it is difficult to program experienced user skills in a robotic setup without having clear knowledge of underlying model used by the operators. In this paper we present a preparation for capturing human operator’s dynamics using an instrumented hand-held tool and a motion capture setup. We first present the design of an instrumented tool and later present a method for reliably capturing kinematics using redundant markers removing effects of marker occlusions, and effect of gravity caused by the tool's mass. Kinematic information is used for deriving the forces/torques on the tool end effector.

Abstract: This paper investigates the limits of precision caused by color depth discretization in motion capture with known sphere marker in a uniform lighted scene for an ideal pinhole camera using a black box model. For each marker position its ellipse projection in the pixel plane is computed and the final pixel values are established based on how match the pixel area is covered by the marker image. The color of some boundary pixels loses information due to the color depth discretization and therefore the same image represents a range of marker location in space and not a single location. The maximum variations on all the major axis are computed with a precision of 100 nm for each marker location and a final analysis is performed using this findings.

Abstract: Tube bending is one of the most relevant manufacturing processes for the production of structural elements, but it suffers from the problem of springback that requires the tuning of the process parameters at every launch of new production batches. Off-line optimization approaches can be found in literature, but they often require complex characterization of the material properties or the application of approaches based on numerical simulation analyses. So, the development of new and more flexible on-line approaches to measure and correct the springback is crucial especially for highly automated machines as for example the tube benders. The paper presents a new measurement approach, based on the application of motion-capture techniques, to provide real-time measurements of the bent tube orientation, in order to decrease the time for the set-up of the main process parameters. A new methodology as well as a new experimental apparatus for the in-line monitoring of the tube springback is presented, as well as the evaluation of its accuracy when applied to the industrial process. An Inertial Measurement Unit (IMU) is linked to the tube during bending and the measurements from three gyroscopes and three accelerometers are used to perform the computation of the tube orientation in the 3D space. The proposed approach appeared promising for the evaluation of the springback through the measurement of the final angular configuration reached after bending.

Abstract: This paper designed the posture initialization calibration method by the inertial sensor in human limb movement with any attitude toward. By initializing the target specific actions can be implemented to identify timing corresponding sensors and joint, and calculate the coordinate transformation relation of human skeletal coordinates corresponding to each inertial sensor's coordinate system and the 3D human skeleton model. Then through the coordinate conversion of inertial sensor attitude coordinates and depth first traversal calculation on human skeletal tree, real-time update of human motion body attitude data, driven simulation of human skeletal model by human motion, realize the real-time tracking of motion capture.

Abstract: Kaedah A is a fend off technique engaged in Seni Silat Cekak Malaysia, upon confronting a punch force exerted within the vicinity of the thorax area. Hitherto, there is still lack of biomechanical analysis on the execution of Kaedah A. Therefore, this study aims at analysing the effectiveness of Kaedah A based on the total execution time as well as to describe the kinematic characteristics of the hand movement upon its execution. The experiment was carried out by means of motion capture. Microsoft Kinect was utilised to detect the hand movement whilst the post processing of the captured motion was performed via Virtual Sensei Lite. Kaedah A was executed five times by an experienced Seni Silat Cekak Malaysia practitioner to investigate the accuracy and repeatability of the system. The data obtained serves as an input for the trajectory mapping for both initial and end point identification. The time difference, Δt between the points demonstrates that the total time execution for Kaedah A is less than 0.1 s. Further analysis involves filtering the coordinate data obtained in order to generate the polynomial function of the hand movement during the execution of Kaedah A. It could be concluded that the Kaedah A execution has the features of a ballistic movement. The findings provides useful data for reliability prediction as well as further enhancement of the Kaedah A itself.

Abstract: Resulting in an optimal grasping pose of satisfying task requirements is a critical problem for hands manipulation. Based on the problem, taxonomy of grasping poses was created. This paper presents key-based search method that can obtain all hand manipulation poses matching the grasped object. However, these grasping pose must be ranked by combining with task requirements and object geometry. In this way, optimal grasping pose can be achieved. The effectiveness of the method was demonstrated by using grasping pose taxonomy comparisons from motion capture example database.

Abstract: With the development of the society, the study of folk craft and protection still appears imminent. We not only should attach importance to and protection of folk art, and to look for more advanced and effective means of protection. On the protection and development of national folk craft, directly involved in the sustainable development of national culture. Using the digital technology can realize the collection and protection of the folk craft. In this paper, under the background of digital folk craft as the research object, the digital protection of folk craft key technology and related application problem discussed in this paper, aimed at to protect, inherit and carry forward the near extinction of folk cultural heritage to provide the reference.