Modelling, Simulation, and Control for a Bipedal Walking Robot

The goal of this thesis is to design a controller that enables the Advanced Biomechatronics and Locomotion Laboratory- Biped I (ABL-BI), a biped with 13 active degrees of freedom, to walk continuously in a stable manner in the presence of small external disturbances. With this goal in mind, two walking controllers are designed. The first controller (the ZMP-based controller) is a modification of an existing real-time controller that is based on the Linear Inverted Pendulum Model (LIPM) and the Zero Moment Point (ZMP). Two major changes are made to this controller: 1) the addition of a Center of Gravity (COG) position control loop, and 2) the application of a task prioritization method in the inverse differential kinematics algorithm of the biped. These changes are made to simplify reference motion planning and postural control. The second controller (the ZRAM-based controller) is a new bio-inspired control method that regulates the centroidal moment of the biped by minimizing the distance between two ground reference points: the reference Centroidal Moment Pivot (CMP) and the Center of Pressure (COP). The main advantage of this method is its ability to modify the reference feedforward motion in real-time to attenuate external disturbances. It is also a relatively simple control strategy that focuses on the fundamental relationships and state variables that are key to the dynamic stability and high level motion of the biped. In addition, a new approach in kinematic modelling and solution of the equations has been proposed and implemented. The full kinematics of the biped is treated as separate open kinematic chains in order to delegate task space objectives to specific segments of the biped and simplify postural control.;;A 3-D simulated biped that is based on the kinematic and mass specification of ABL-BI is developed in a simulation environment called Webots. Walking tests are performed on the simulated biped to validate the two proposed walking controllers and the results show that both the ZMP-based controller and the ZRAM-based controller can generate stable walking gaits. In addition, a series of perturbation tests are performed on the ZRAM-based controller to evaluate its robustness to external disturbances. It is shown that the level of disturbance that is attenuated by the ZRAM-based controller in the perturbation tests is comparable to a similar walking controller in the literature.