Abstract Computed tomography is a widely used imaging method for non-destructive testing. However, standard CT systems face fundamental limitations when scanning large objects such as car bodies, which must fit between the X-ray source and detector, both of which need freedom of movement around the specimen. Twin-robotic CT systems with high degrees of freedom address these limitations by enabling free positioning of the X-ray source and detector in space, making non-destructive CT testing of large objects feasible. However, achieving collision-free positioning of the robots is a challenging problem that is often neglected in theoretical representations of twin-robot CT configurations. This paper presents a systematic methodology for performing region-of-interest scans on large objects. The approach exploits the test object’s geometry to determine robot reachability, which serves as the foundation for trajectory planning by incorporating accessible regions. By leveraging both rotational and translational degrees of freedom, including variable source-detector distances and detector rotations, the methodology expands the range of collision-free poses, thereby increasing reachability and enabling more flexible trajectory design. The methodology is modular and adapts to arbitrary system configurations and test samples via computer-aided design (CAD)-based geometry definition, where the test object determines the collision-free workspace. It is demonstrated on a BMW 4-series body-in-white through comprehensive batch simulation across 273 Region of Interest (ROI) positions, evaluating reachability improvements achieved through the introduced degrees of freedom using a data completeness trajectory optimization criterion.
Schnitzer et al. (Thu,) studied this question.