Today, more and more bipedal humanoid robots are being used, for a variety of purposes. These can range from helping out in everyday situations, to looking through debris in devastated areas. However, bipedal robots are still far from perfect: they lack versatility and robustness to be used more readily in more wide-ranging situations. Two main approaches exist to control the gait of bipedal robots: top-down approaches concentrate on mathematically proven methods to design stable gait controllers; these approaches often result in high computational costs, low energy-efficiency and low versatility. The second kind of approaches are bottom-up approaches. These make use of human motion analysis, biomechanics and neuro-scientific research and try to transfer those to the control of bipedal robots. Such an approach is used in this work, trying to make an existing bio-inspired gait controller more versatile by enabling precise control of the robot's gait. The existing bio-inspired controller is first introduced, as well as all the tools necessary for the development and simulation. Then, the optimization procedure used to obtain satisfying gaits throughout this work is introduced, together with an example of 'natural gait' obtained with the controller. After that, the work is presented as emph{4 successive experiments} to adapt the gait controller and enable the control of gait features. In the first experiment, a sensitivity analysis is performed to help guide the development of the next experiments. In the second experiment, a first controller is developed enabling to control two gait features (step height and step length) by continuous modulation. In the third experiment, a second way to develop a controller for step height and step length is explored. In the last experiment, new modulations are tested, allowing to modify the gait for a few steps only, enabling obstacle avoidance. Finally, the main results and take-aways from all the experiments are summarized in the conclusion, and future perspectives are discussed ; Master [120] : ingénieur civil électromécanicien, Université catholique de Louvain, 2017
Gait modulation of a humanoid robot, using bio-inspired mechanisms
2017-01-01
Hochschulschrift
Elektronische Ressource
Englisch
DDC: | 629 |
IEEE | 2014
|Emotional gait generation for a humanoid robot
British Library Online Contents | 2010
|Humanoid Robot Gait on Sloping Floors Using Reinforcement Learning
Springer Verlag | 2016
|Gait control and optimization of humanoid soccer robot
IEEE | 2019
|Gait Generation of Humanoid Robot Based on Analysis of the Human's Gait
Trans Tech Publications | 2014
|