1: \begin{abstract}

2: We will report our efforts in designing feedback for the thruster-assisted walking of a bipedal robot. We will assume for well-tuned supervisory controllers and will focus on fine-tuning the desired joint trajectories to satisfy the performance being sought. In doing this, we will devise an intermediary filter based on the emerging idea of reference governors. Since these modifications and impact events lead to deviations from the desired periodic orbits, we will guarantee hybrid invariance in a robust fashion by applying predictive schemes within a short time envelope during the double support phase of a gait cycle. To achieve the hybrid invariance, we will leverage the unique features in our robot, i.e., the thruster.   

3: 

4: % This paper introduces a 5 link biped with point feet and thrusters fixed at its torso. The thrusters are inactive during swing phase to ensure the existence of zero dynamics. An impact map where both feet are fixed to the ground is used, which leads to a prolonged double support phase where the thrusters are utilized to regulate ground contact force. The kinematic constrained brought on during this phase reduces the degree of freedom on the lower half of the body, leading to an over-actuated system. During this phase a nonlinear Model Predictive Controller (MPC) is used such that friction cone condition is guaranteed to satisfy physical constraint, and the states after impact are steered back to the zero dynamic manifold to ensure periodicity.

5: 

6: %  Despite significant impulsive effects in the dynamics, stable thruster-assisted walking gaits were achieved with the extended double support phase and the results suggest an enhanced performance and superior post-impact convergence to the zero dynamics manifold in comparison to traditional event-based schemes.

7: 

8: \end{abstract}

9: