We are developing a unified approach to computer-aided mechanical assembly design in which all design tasks are performed within a single computational paradigm supported by integrated design software. Mechanical assembly design is the task of devising an assembly of parts that performs a function reliably and economically. It is a ubiquitous activity that spans mechanical, electrical, and biomedical engineering. Designers need to devise, analyze, and compare competing design prototypes to produce optimal designs. Computer-aided design reduces design time and improves quality by allowing designers to substitute electronic prototypes for physical prototypes in diverse tasks.
We have developed a prototype design environment called HIPAIR for general planar assemblies. The environment supports the key design tasks of simulation, parametric design, and functional tolerancing for a broad range of mechanical assemblies, such as camera shutter mechanisms, part feeders, robotic arms, and knee prostheses. HIPAIR organizes the design tasks around the fundamental task of contact analysis, which we automate by configuration space computation. Configuration space is a complete, concise, and explicit representation of rigid body interactions and contains the requisite information for design tasks involving contacts. HIPAIR allows designers to perform computations that lie outside the scope of previous software and that defy manual analysis. It allows them to visualize assembly function under a range of operating conditions, to find and correct design flaws, and to evaluate the functional effects of part tolerances. It has been tested on hundreds of pairs and on a dozen assemblies. HIPAIR performs at interactive speed on assemblies of ten parts with tens of thousands of contacts.
We are expanding this approach to selected spatial assemblies and to other design tasks, such as toleranced assembly planning, validation, and parametric design. We plan to integrate HIPAIR with existing CAD systems and to validate it with practicing engineers on real design examples. This research should lead to a better understanding of mechanical design and pave the way for the next generation of rapid prototyping tools that will help industry reduce design time and increase product quality.
Computer-Aided mechanical assembly design using configuration spaces (1,335 Kb, Postscript)
L. Joskowicz and E. Sacks
Purdue U. CS Tech Report 97-001.
![[Back]](arrow2.gif)