Cascaded constrained optimization for cheetah-inspired galloping robot leg mechanism

The paper describes one of the key steps in creating a bio-inspired galloping robot, namely parametric optimization of a multi-link leg mechanism. We introduce the cascaded constrained optimization procedure for finding optimal values for geometric parameters of a femur, knee, and foot segments for given design constraints and various gait specifications. The presented approach is computationally efficient and guarantees convergence to the unique solution. Steps on free parameters, constraints, and cost function selection are discussed. The efficiency of the proposed procedure is illustrated via a series of simulations. The proposed approach can also be implemented for the design of wearable robots like upper- and lower-limb exoskeletons and rehabilitation systems that should follow sometimes sophisticated human-like trajectories.