PAGO: Pelvic analysis and gait output

Seung-Joo Lee, University of Pennsylvania


Human locomotion is one of the most fundamental and routine human actions we perform and, at the same time, we see every day. It is natural therefore to expect good synthesized virtual human walking in virtual reality, computer games, animation, and training. However, human walking is not such a simple motion even if its basic structure is cyclic. It is not easy to synthesize robust and realistic human walking only using patterns or rules without using empirical data stored in a motion capture library. It is even more difficult to select proper subsequences of motion from large motion (capture) library. We introduce a new locomotion system, PAGO (Pelvis Analysis and Gait Output), which reuses motion capture data by accepting the pelvic motion stream as a control input. PAGO explores a solution to synthesize natural locomotion based on a minimal control motion which has at least pelvic movements. PAGO analyze and obtain a pelvic cycle with high accuracy and search pelvic cycles from stored motion. The final motion re-creates precisely synthesized leg movements.

Subject Area

Computer science

Recommended Citation

Lee, Seung-Joo, "PAGO: Pelvic analysis and gait output" (2006). Dissertations available from ProQuest. AAI3246187.