This video shows a set of jumping experiments on Solo8 v2. The controller has been designed based on a Reinforcement Learning framework we have developed in this paper [ Ссылка ]. We demonstrate that the designed controller can generate jumps that are robust to drastic changes in the environment properties, i.e. surface height, stiffness, damping, etc.
More information: [ Ссылка ]
Hardware documentation: [ Ссылка ]
Forum for questions: [ Ссылка ]
