ABSTRACT
Due to accuracy requirements, robots and machine-tools need to be periodically verified and calibrated through associated verification systems that sometimes use extensible guidance systems. This work presents the development of a reference artefact to evaluate the performance characteristics of different extensible precision guidance systems applicable to robot and machine tool verification. To this end, we present the design, modeling, manufacture and experimental validation of a reference artefact to evaluate the behavior of these extensible guidance systems. The system should be compatible with customized designed guides, as well as with commercial and existing telescopic guidance systems. Different design proposals are evaluated with finite element analysis, and two final prototypes are experimentally tested assuring that the design performs the expected function. An estimation of the uncertainty of the reference artefact is evaluated with a Monte Carlo simulation.
ABSTRACT
Motion capture systems are often used for checking and analyzing human motion in biomechanical applications. It is important, in this context, that the systems provide the best possible accuracy. Among existing capture systems, optical systems are those with the highest accuracy. In this paper, the development of a new calibration procedure for optical human motion capture systems is presented. The performance and effectiveness of that new calibration procedure are also checked by experimental validation. The new calibration procedure consists of two stages. In the first stage, initial estimators of intrinsic and extrinsic parameters are sought. The camera calibration method used in this stage is the one proposed by Tsai. These parameters are determined from the camera characteristics, the spatial position of the camera, and the center of the capture volume. In the second stage, a simultaneous nonlinear optimization of all parameters is performed to identify the optimal values, which minimize the objective function. The objective function, in this case, minimizes two errors. The first error is the distance error between two markers placed in a wand. The second error is the error of position and orientation of the retroreflective markers of a static calibration object. The real co-ordinates of the two objects are calibrated in a co-ordinate measuring machine (CMM). The OrthoBio system is used to validate the new calibration procedure. Results are 90% lower than those from the previous calibration software and broadly comparable with results from a similarly configured Vicon system.
