• control
• planning
• quadrupedal robots
• quasi-velocity

Climate change is currently threatening the preservation of the Earth biodiversity. Environmental monitoring is one of the main weapons humans have to combat this problem. In this context, robotics can be a significant asset to obtain continuous and reliable monitoring of natural habits. In particular, legged robots have shown unparalleled performance when moving in such scenarios. However, careful planning and control of their motions are required to accomplish the desired mission. For instance, locomoting in unstructured scenarios may require very precise tracking to avoid collisions. Additionally, inspecting the health status of plants may require precise positioning of the camera mounted on the robot. In my thesis, I propose a novel control architecture based on the quasi-velocity method. Such an approach allows to directly control all the degrees of freedom of the base of the robots, without any alteration of the contacts with the environment. This can enable effective monitoring of the habitat to-be-inspected. Furthermore, most of the techniques proposed by the literature require controlling the position of the leg joints to obtain the desired position of the main body. This leads to a huge computational load caused by the leg inverse kinematics computation. Conversely, the approach I propose is able to reduce the computational complexity thanks to the limited use of leg inverse kinematics. To test the effectiveness of the proposed method, I implemented a crawling gait for the quadrupedal robot ANYmal C in a simulation environment.