Studies On Haptic Rendering Algorithms

The sense of force is a very important sensory modal of human being, just like vision and audition. In real world, many properties of physical objects are explored by force cues, such as the shape information, the temperature, and the stiffness and so on. While, in virtual world, the geometry and other properties of computer-generated virtual objects are haptically perceived through all kinds of electro-mechanical devices which are called haptic devices. The process of exploring virtual objects by haptic devices is called haptic rendering.During the past several decades, many effective haptic rendering algorithms have been developed. However, in order to make users interact with virtual objects more realistically, more effective haptic rendering algorithms should be designed to achieve more stable and realistic feedback force. Developing such algorithms is full of challenges. The challenges lie in three aspects. Firstly, the efficiency of haptic rendering algorithms is required to be rather high, normally higher than 1 KHz. Secondly, haptic rendering algorithms should be designed regarding the specific physical features of a haptic device, for haptic rendering is a highly device-dependent process. Lastly, the optimal running time of haptic rendering algorithms intrinsically depends on the complexity of input models.There is a kind of interaction named area-contact interaction and many haptic devices provide area-contact interaction interface, which is called area-contact haptic devices such as CyberGrasp. We proposed a new area-based haptic rendering algorithm for area-contact haptic devices. Unlike previous work, the algorithm employs clusters of line segments with common original points as haptic proxies, not just a point or 3D mesh. Considering the feature of area-contact interaction, our haptic proxies are constructed by sampling the contact areas. The target objects are uniformly partitioned to accelerate the haptic rendering process. Our algorithm can haptically render stable and realistic feedback force, for it not only respects thefeature of area-contact haptic devices but has reasonable efficiency.Haptically rendering high complexity objects in real-time is difficult and has been received many interests. Haptic rendering algorithms, whose efficiency depends on the complexity of input models, can not render high complexity objects fairly good. To overcome this problem, we proposed a new haptic-box-based real-time haptic rendering algorithm. Haptic-box is a kind of representation of target object for haptic rendering, which encodes the contact information of sampled rays with target object. With haptic-box represented target object, collision detection algorithm only takes 6 operations to calculate contact information between a line segment and target object. One operation is to determine the contact information between a line segment and a box, which is fast. The other five operations are bilinear interpolation, which are fast too. Thus, many unnecessary contact determination operations are avoided and the process of collision detection is very rapid. The contact information from collision detection process can be directly used to compute feedback force. In the process of haptic rendering, our algorithm doesn't directly deal with the input high complexity models, hence it is independent of the complexity of input models. Experiments show that our algorithm is both rapid and effective.