Straight legged walking of a biped robot

A Bio-Robotic Toe & Foot & Heel Models of a Biped Robot for More Natural Walking: Foot Mechanism & Gait Pattern

“…This was due to the fact that humans also generate an almost knee stretch motion during toe-off. The knee stretch walking not only gives natural and humanlike walking pattern but also better performance in the knee torque requirement (Kurazume et al, 2005;Ogura et al, 2006). This gives an opportunity to employ smaller actuators for the knee.…”

“…Our algorithm can generate stable knee stretch walking patterns without singularity. With knee-stretched motion, biped robot walking pattern will not only become more similar to human, but also will require less torque and thus making the pattern more energy efficient as human walking itself is optimized to energy efficiency (Kurazume et al, 2005;Ogura et al, 2006). Moreover, with our proposed mechanism, utilizing extra addition of heel joint is comparably less sophisticated than those biped robots, which have waist joints.…”

“…Knee stretch motion has often been related to the naturalness of biped walking. Recently some researchers tried to achieve stable walking patterns involving knee stretch motion (Kurazume et al, 2005;Ogura et al, 2006). There is even a commercial product for humanoid robot that is able to walk with straight knee (Garage, 2008).…”

“…There is even a commercial product for humanoid robot that is able to walk with straight knee (Garage, 2008). The research utilized a conventional biped robot with flat foot to generate straight walking pattern (Kurazume et al, 2005). One of the main reasons why knee stretch motion for biped walking is hard to achieve is that during the knee stretch, the inverse kinematics solution for the leg becomes singular.…”

Singularity-Tolerant Inverse Kinematics for Bipedal Robots: An Efficient Use of Computational Power to Reduce Energy Consumption

“…Apart from model-free walking approaches [9], which deal with singularity problem differently, a large number of model-based algorithms are based on preplanned knee trajectories, where other joint angles are adjusted to realize desired swing or Center of Mass (CoM) motions [10], [11]. Likewise, the walking planner proposed in [12] modifies a parametrized 2D CoM trajectory to limit large knee velocities close to singular positions. For arbitrary balancing tasks, however, such periodic trajectories can not be found.…”