Title: Skin Based Task Hierarchy Rearrangement for Collision Avoidance and "Intentional Contact"

Author: Julio Rogelio Guadarrama Olvera, M. Sc. 

Abstract: Skin technology enabled a powerful way to sense the environment in robotic systems. It allows to simplify the formulation of safety tasks such as collision avoidance between the robot, the environment and surrounding objects. In this paper, a hierarchical policy based on tactile feedback is proposed to let a robot interact with its environment while performing a set of tasks. Any sort of collision can be prevented using simple potential functions. In this context, the concept of “Intentional Contact” is introduced to escape from undesired equilibrium points produced by local minima in the potential fields. Allowed contact with the environment enables a robot to modify its surroundings in order to fulfill the main task. Such contact is permitted as long as the generated force remains under a specific limit, otherwise a reactive action is taken to reduce it. This new concept is validated in both simulation and one real robot. 

Title: Estimation of Contact Forces and Floating Base Kinematics of a Humanoid Robot

Author: A. Mifsud 

Abstract: This paper overviewes our approach of state estimation of the floating base kinematics and contact forces of the humanoid robot HRP-2. We begin to reconstruct the flexibility dynamics, from only the inertial measurement units (IMUs), proprioceptive informations and a naive model. We then show that by adding a more complex model of the elasticity we are able to reconstruct both the contact forces and torques with the same set of measurements. Finally, we show that by adding contact wrench measurements, we significantly improve the estimation accuracy and even estimate or identificate some paramters of the elasticity model, increasing the robustness against model errors.

Title: Robustness to Inertial Parameter Errors for Legged Robots Balancing on a Level Ground

Author: N. Giftsun 

Abstract: Model-based control has become more and more popular in the legged robots community in the last ten years. The key idea is to exploit a model of the system to compute precise motor commands that result in the desired motion. This allows to improve the quality of the motion tracking, while using lower gains, leading so to higher compliance. However, the main flaw of this approach is typically its lack of robustness to modeling errors. In this paper we focus on the robustness of inverse-dynamics control to errors in the inertial parameters of the robot. We assume these parameters to be known, but only with a certain accuracy. We then propose a computationally-efficient optimization-based controller that ensures the balance of the robot despite these uncertainties. We used the proposed controller in simulation to perform different reaching tasks with the HRP-2 humanoid robot, in the presence of various modeling errors. Comparisons against a standard inverse-dynamics controller through hundreds of simulations show the superiority of the proposed controller in ensuring the robot balance. 

Title: A two-stage approach for elastic joints control based on Differential Dynamic Programming

AuthorFlorent Forget

Abstract: Elastic joint are increasingly used in humanoid robotics in order to add some compliance during contacts. The control of such kind of actuators is still an ongoing problem due to non-linearities and a reduced bandwidth as     the stiffness decreases. We propose here a two-stage approach to control elastic joints, both series elastic actuators or variable stiffness actuators.  The first stage is a whole body controller generating trajectories (position/stiffness     or torque) for each joint. Then a non-linear model predictive controller (based on differential dynamic programming) is used for each joint to generate the actuator input to track the reference trajectory computed before.


Title: Coupled method MPC - Whole body torque control for artificial locomotion

Author: Thomas Flayols

Abstract: One of the main current goals in artificial locomotion is to adapt to and recover from unexpected environments and events such as non perfectly flat terrain and pushes from outside. This is particularly needed when the robot has limited contacts with the environment, when walking for instance. To do so, we believe that a locomotion algorithm is needed, that can change online and at the control frequency, the planning of the steps and the whole body posture. We propose a coupled method that is generating trajectory for the feet and the center of mass along with an instantaneous whole body control, and we are currently working to apply it on the real robot with full force feedback. The first part of the work focuses on the motion generation. As the formulation solves a large problem with coupling constrains, the next steps placements can be adapted to real time error in COM positioning. This algorithm gives torque references for each joint. To track this torque, a low level torque control is needed. Developping it is the second part of our work. On humanoid platform like HRP-2, this is challenging because actuation was designed for position control. High gear ratio makes actuation less transparent, so motors current is a poor joint torque estimate. We are currently modeling actuation and fusing torque estimation from current measurement and from force-torque sensor only available at ankle and wrist. With this model and torque estimation, a low level control law is going to be developped that could be extended to other position controlled platforms.


Title: Stack of Tasks on ROMEO

Author: Kevin Giraud Esclasse

Abstract: For two years we have investigated in the whole body control of the Romeo robot.  We managed to implement the Stack of Tasks directly on the robot. It allowed us to provide several motions like a whole body visual tracking using Visp library. With the tasks hierarchical solver and the last features included in the SOT, the posture task is achived with a quit short and promising loop period of 3ms. This poster will give an overview of the implementation process and its issues.

Online user: 1