Implementation and Evaluation of Mixed Model Representation of Virtual Objects in a Haptic Based Virtual Assembly Platform Available to Purchase

Virtual assembly systems have become popular in recent years due to its ability to simulate natural interaction between parts and ease of manipulation by the user. One of the most relevant technologies used in virtual assembly systems are haptic devices that provide force feedback and allow simulating real word conditions, such as weight, inertia, texture and collisions. Physics simulation engines (PSE) are another important tool used to simulate a realistic behavior in virtual assembly systems by enabling the effect of gravity and collision response of the virtual objects, resulting in a real world behavior. However, the use of haptic systems together with physics simulation engines is costly in terms of computing resources. This cost is mainly associated to collision detection between virtual objects, and increases when the shapes represented within the PSE are more complex, resulting in a poor performance of the virtual assembly system, making very difficult to simulate the assembly of complex parts or use several parts in the assembly. The present work shows a new algorithm to simulate complex objects, by using a different representation of the same object according with its dynamic state during the assembly process. The results show that the use of mixed model representation reduce the computing time when assembling objects, thus improving the performance of the virtual assembly system and finally allowing a better comfort and performance of the user during the assembly process. The system HAMS (Haptic Assembly and Manufacturing System) was used for the experimental validation, also the simulation of four assembly tasks that simulate real assembly objects, has been conducted.