ROB4FAM

Robots For the Future of Aircraft Manufacturing

Project Details

  • Type : Industrial Project
  • What We Did : Motion Planning/Whole body control/Walking/Coordination
  • Tools Used : Stack-of-Tasks, jrl-walkgen, HPP
  • To be completed on : 2022
  • Skills : Model Predictive Control / Walking
  • Partners : Airbus / Florent Lamiraux

ROB4FAM - Robotics For the Future of Aircraft Manufacturing

The goal of this project is to develop and port on industrial robots reactive technologies developed in the frame of our research projects.

The humanoid robot is our scientific flagship which impose to think out-the-box and develop new technologies.

WP-1 Real time/Interactive motion planning

Description

The goal of this workpackage is to provide a path planning scientific approach in order to create a hierarchical finite state machine based on the task to perform, the robot model and the object to be manipulated. The core scientific contribution is described in the IEEE Transactions on Robotics paper from Florent Lamiraux (WP1-TRO1). It consists in the capabilities to associate task functions following a graph of constraints, to start from a configuration of the robot and the objects to reach a final state. It was applied to the humanoid robot TALOS (WP1-SII) and the mobile robot Tiago (WP1-CASE)

Publications

WP-2 Instantaneous whole body control (including the task to perform)

Description

This workpackage plan was to implement a task space inverse dynamics for a complex or humanoid robot. In this framework, using Dynamic Differentiable Programming for one actuator over a trajectory was tested (see WP2-ECC). As there is a large litterature showing that it is easier to perform walking with position controler vs torque control we perform some comparisons between both in simulation (see WP2-ICAR), and on the real robot (see WP2-Frontiers). This led us to develop a passivity based scheme for Task Space Inverse Dynamics.

Publications

WP-3 Estimation-SLAM

Description

This workpackage aims at providing a localization system for robots such that they can localize accurately and start performing motion planning. In the frame of a the H2020 Memmo project, we have deployed the system developed by the Dynamic Robot Systems Group, Oxford Robotics Institute. It prodives a system precise up to 2 cm.

Publications

WP-4 Balance

Description

This workpackage aims at generating balanced trajectories for walking and perform loco manipulations.

Publications