Publicação
Robotic locomotion combining central pattern generators and reflexes
| Resumo: | Locomotion of quadruped robots has not yet achieved the harmony, flexibility, efficiency and robustness of its biological counterparts. Therefore, the development of bioinspired controllers seems to be a good and robust way to obtain an efficient and robust robotic locomotion, mimicking their biological counterparts. Taken this matter into consideration, this thesis addresses the development of bio-inspired controllers based on the bio-inspired concepts of Central Pattern Generators (CPGs) and reflexes, implemented in the simulated Oncilla quadruped robot. Two bio-inspired controllers were developed in order to generate stable and robust locomotion on uneven terrains. Firstly, a reflex controller that has to be capable of generating locomotion in a quadruped robot, based on sensory information that results from the interplay between robot and environment. Secondly, a hybrid controller that combines CPGs and reflexes, thus exploring the advantages of the pure feedforward and feedback approaches. The hybrid controller is more resilient to external perturbations and more robust to noise, thus improving the overall performance. The results show that the reflex controller is capable of producing stable quadruped locomotion with a regular stepping pattern. Furthermore, it proved to be able to deal with slopes without changing the parameters and with small obstacles, overcoming them successfully. The hybrid controller improved the robot’s behavior by increasing its stability, harmony and displacement in the majority of the experiments. Moreover, the stepping patterns become more regular due to the inclusion of the feedforward component in the system. This combination enables to increase robustness to sensor imperfections and to anticipate the robot’s motor actions. This research’s main contribution to knowledge is based on the hybrid controller, which presents an innovative approach. CPGs have a different interpretation. They are no longer used to generate oscillatory signals to produce feedforward motor commands, but rather to assist the decoding of sensory information. Since they work as motor actions’ predictors, they improve the Oncilla’s walking behavior. |
|---|---|
| Autores principais: | Ferreira, César José Cardoso |
| Assunto: | Quadruped locomotion Central pattern generators Reflexes Predictor Feedback Feedforward Sensory information Hybrid controller Reflex controller Locomoção quadrúpede Geradores de padrões centrais Reflexos Estimador Informação sensorial Controlador híbrido Controlador de reflexos |
| Ano: | 2014 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade do Minho |
| Idioma: | inglês |
| Origem: | RepositóriUM - Universidade do Minho |
| Resumo: | Locomotion of quadruped robots has not yet achieved the harmony, flexibility, efficiency and robustness of its biological counterparts. Therefore, the development of bioinspired controllers seems to be a good and robust way to obtain an efficient and robust robotic locomotion, mimicking their biological counterparts. Taken this matter into consideration, this thesis addresses the development of bio-inspired controllers based on the bio-inspired concepts of Central Pattern Generators (CPGs) and reflexes, implemented in the simulated Oncilla quadruped robot. Two bio-inspired controllers were developed in order to generate stable and robust locomotion on uneven terrains. Firstly, a reflex controller that has to be capable of generating locomotion in a quadruped robot, based on sensory information that results from the interplay between robot and environment. Secondly, a hybrid controller that combines CPGs and reflexes, thus exploring the advantages of the pure feedforward and feedback approaches. The hybrid controller is more resilient to external perturbations and more robust to noise, thus improving the overall performance. The results show that the reflex controller is capable of producing stable quadruped locomotion with a regular stepping pattern. Furthermore, it proved to be able to deal with slopes without changing the parameters and with small obstacles, overcoming them successfully. The hybrid controller improved the robot’s behavior by increasing its stability, harmony and displacement in the majority of the experiments. Moreover, the stepping patterns become more regular due to the inclusion of the feedforward component in the system. This combination enables to increase robustness to sensor imperfections and to anticipate the robot’s motor actions. This research’s main contribution to knowledge is based on the hybrid controller, which presents an innovative approach. CPGs have a different interpretation. They are no longer used to generate oscillatory signals to produce feedforward motor commands, but rather to assist the decoding of sensory information. Since they work as motor actions’ predictors, they improve the Oncilla’s walking behavior. |
|---|