Markus Ryll

Research Group:

Supervisor:

I am a Ph.D. student with a background in Medical Engineering and Mechatronics. In the group group of (Max Planck Institute for Biological Cybernetics) I am working on the further development of quadcopter control.

Currently I am interested in:

• Micro aerial vehicles (MAV)
• Advances MAV designs
• Quadcopter flight behavior
• Quadcopter control
• Aerial grasping
• Mechanical Engineering
• Teleoperation
• Software Engineering

Actual I am working on the design of a new Micro aerial vehicle. For more information check my projects.

Introduction

Micro aerial vehicles (MAVs) are an increasingly important research field. On the one hand, the idea of using MAVs in cluttered indoor environments and for rescue and surveillance tasks is becoming increasingly popular. On the other hand, the development of new MAV concepts and the related control issues represent scientific challenges. For instance: what is necessary to control a micro aerial vehicle in an unknown and cluttered environment? What are the requirements for such a MAV? What kind of sensor inputs is necessary and what kind of autonomy is needed? Shall the MAV be completely autonomous or under the supervision of a human operator as in teleoperation scenarios?

Goals

Custom quadrotor helicopters have the drawback of being underactuated: they only possess four independent control inputs w.r.t. their 6-dof pose in space. A rotation in roll or pitch causes a Cartesian acceleration into the same direction. As a consequence, it can be proven that quadrotors are only able to independently act on their Cartesian position and yaw angle, this means 4 dofs. This imposes some limitations. For example, a sensor or gripper attached to the robot cannot be arbitrarily oriented during flight nor the quadrotor can hover in place with any body orientation. In the Human-Robot-Interaction group, we are developing a new kind of MAV. Based on the concept of a quadrotor helicopter with four propellers, we extend this design to allow the four propellers to tilt about their axes w.r.t. the quadrotor body. This gives us full control over the 6 dofs of the quadrotor body, and better chances to interact with and manipulate the environment.

Methods

Starting from first principles, we aim at deriving a realistic dynamical model of our envisage quadrotor and at devising advanced nonlinear controllers which can exploit the additional actuated degrees of freedom. Inertial and aerodynamics effects will also be considered in order to improve the realism of the model. Besides providing a matlab implementation, we will also port the quadrotor model into our own real-time multi-robot simulation environment with a physically realistic engine.

In a future step, we plan to build a prototype of this new quadrotor concept in order to test and validate in real conditions both the models and controllers obtained during the simulation stage. To this end, we will make use of our two Flying-Arenas equipped Vicon tracking systems for obtaining an external ground truth.

Initial results

We developed a dynamic model of the quadrotor with actuated propellers and devised a nonlinear controller based on dynamic feedback linearization. Simulation results confirm the validity of our methods.

Initial conclusion

The Human-Robot-Interaction group develops a new kind of quadrotor helicopter with tilt rotors to control all 6 Cartesian space dofs. This MAV will be able to interact with and manipulate the environment.