Giancarlo Amati

A CAD-augmented reality integrated environment for assembly sequence check and interactive validation

An integrated environment based on CAD assembly software and on an Augmented Reality wearable system is used to improve the overall integration between engineering design and real prototypes manufacturing. The environment following called – Personal Active Assistant (PAA) – exploits a CAD tool connection to remarkably improve object recognition, best assembly sequence optimization, and operator instructions generation. PAA is real-time and wirelessly linked to a remote server or designer workstation where project geometric database is stored. The PAA head-mounted camera is also able to acquire the human-driven assembly sequence and check the efficiency and correctness via object recognition: an incremental sub-assembly detection algorithm has been developed in order to achieve complex dataset monitoring. On the other hand, the Augmented Reality-based assembly evaluation tool allows engineers to interact directly with the assembly operator while manipulating the real and virtual prototype components. Information from the assembly planner can be displayed, directly superimposed, on the real scene by using a see-through head-mounted display. Thus the new combined software and hardware equipment may be considered a step ahead in the support of true concurrent engineering and remote collaboration, strongly improving this latter through a better heterogeneous task integration. Several tests have been performed also to achieve personnel training and warehouse part seeking.

From spline to Class-A curves through multi-scale analysis filtering

This paper reports the work on a novel wavelet-based multi-scale filtering application used to generate very smooth subset of profiles known as Class-A curves. The multi-scale representation, based on B-spline wavelets, allows to split a spline curves in terms of a coarser least-square approximation and details coefficients. It raises that non-Class-A curves are determined by geometric imperfections strongly connected to the details coefficients. The extraction and manipulation of details with multi-scale filtering allows to select and evaluate geometric imperfections. Finally, an efficient algorithm devoted to improve a given non-Class-A profile to a Class-A curve has been implemented and tested with several examples in order to check and visualize the results.