Information Technology Journal1812-56381812-5646Asian Network for Scientific Information10.3923/itj.2013.1160.1167WangZai-jun FangBao-fu ShiGuo-qiang 62013126Omni-directional walking is a humanoid, flexible and efficient
walking way for biped humanoid robot. A method based on motion decomposition,
inverse kinematics, key-frame and cubic spline interpolation is proposed to
realize Omni-directional gait and walking planning. The Omni-directional gait
can be seen as an integration of the three independent movements, so each movement
can be planned separately. It significantly reduces the complexity of implementation,
as well increases the accuracy of gait. Base on the implementation of omni-directional
gait, reinforcement learning is used to optimize the walking path while tracking
dynamic target in order to be more stable and efficient. Experiment shows the
method is valid and efficient.]]>Alur, R., K.E. Arzen, J. Baillieul, T.A. Henzinger and D. Hristu-Varsakelis et al.,2005Pleite, P., J. Maria, C. Llover and Enric,2000Fujita, M.,2003Xu, Y. and Y. Tan,2007Behnke, S.,2006Tang, Z. and M.J. Er,2007Chang, G., W. Hao and F. BaoFu,2009Huang, Q., K. Yokoi, S. Kajita, K. Kaneko, H. Arai, N. Koyachi and K. Tanie,2001Mellmann, H. and Y. Xu,2010Bica, A.M.,2012Mitchell, T.M.,1997