Software Framework for Simulation and Control of Quadruped Robot Solo12

Simulation tools are commonly used in robotics, because they enable people to conduct research without the hardware. Unfortunately, simulation never fully replicates the real world. For that reason there is an abundance of simulators where each one has it’s own strengths and weaknesses. We have built a software framework that enables us to run our control algorithms on different simulators or on the real robot just by modifying a single flag. That helps utilize the strengths of each simulator while removing the effort needed to run same the code on the real robot or different simulator. Currently supported simulators are PyBullet and MuJoCo. The focus of this paper is on quadruped Solo12 robot and mimicking the the real robot in the simulation. That type of robots is gaining popularity because they are proven to be very agile and stable. We will also utilize the tools provided by ROS2 to visualize the robot. The approach was verified on a real robot and two simulators. All described code is available online and is free to use.