A Framework To Create Humanoid Robot Animations Through Motion Capture

Producing realistic character animation remains one of the great challenges incomputer graphics. At present, there are three methods by which this animation canbe produced. The first one is key framing, in which the animator gives important keyposes for the character at specific frames. A second one uses physical simulation todrive the character’s motion its results are good, due to lack of control its difficult touse and it’s costly and with characters it’s not much successful. The last one is motioncapture, has been widely used to animate characters. It uses sensors placed on personand collects the data that describes their motion as they are performing the desiredmotion. The main challenge that an animator is confronted with is to generatecharacter animation with a realistic appearance. Some seemingly simple behaviorssuch as human walking are difficult to model because of their inherent instability.Motion capture is useful in a wide array of field such as biomechanics, gait analysis,computer animations, gesture recognition, sign language, music and also fine artperformance. In sport science, motion capture data is used for analyzing andperfecting the sequencing mechanics of premiere athletes, as well as monitoring therecovery progress of physical therapies. Computer animation of human locomotionhas become trendy in recent years because of the need to use human beings asartificial actors in three-dimensional simulation environments. It has also helped ingreat deal to acquire actions for learning real robotic patterns. In the field ofanimation and gaming industry, it is common that motion information are capturedare in different formats and also limited for specific tasks. This research describes theframework in which the different human movements have been taken from motioncapture and that information is animated which sets the direction to study the virtualcharacter models and its locomotion in the virtual environment. It also gives feasibleapproach in understanding of walking gait patterns in that environment. The analysisphase of this research focuses on motions concerned with retargeting of motion capture data and animation by studying humanoid robot. This framework also leads tothe study issues related to safety engineering. The framework is presented which isdivided into three modules i.e. joint structure system which gives attributes about all3D models used in animation. Trajectory pattern defines the walking gaits ofhumanoid robot. The experiment shows its advantages over different systems. Withthe combination of Bezier curves it simulates the smooth curve for walking gait inhumanoid robot and last of all the third module is visual computing is an extendedpart of framework which calculates the falling factors of the body mechanics. Theresults show from motion capture to animation gives extra flexibly in the field ofergonomics for e.g. certain actions performed with in the motion capture equipmentare limited but with animation into it we can achieve the desired results. Other resultsinvolve that we can get information from different kind of virtual character modelsbased on motion capture from the framework. It means that if subject is in differentsize and shape or animal within any different virtual environment all its informationcan be synchronized with this enhanced framework.