Automatic Realignment of Defective Assemblies Using an Inverse Kinematics Analogy

Abstract

Alignment and plumbness of construction assemblies is a challenging and fundamental problem because it relies on manual solutions to the underlying geometric feedback control problems inherent in practices such as pipefitting and steel structures erection. Where defective components and segments are not well controlled, the errors propagate in larger components and therefore cause more severe challenges and large costs consequently. This paper presents a framework for automatic and systematic development of realignment actions required to achieve a desired state by employing and combining two theories: (1) three-dimensional (3D) imaging that enables the identification of the as-built status and then quantification of incurred discrepancies as a feedback signal by comparing the captured as-built status with the designed state existing in the building information model (BIM), and (2) an inverse kinematics analogy that results in the calculation of required changes in the degrees of freedom defined where manipulations and changes can be applied. Experimental results show that the framework can generate the required actions for achieving a desired state systematically and with a high level of accuracy.