Design of a modular and reconfigurable robotic manipulator for evaluating cobotic approaches

Context

Recent progresses in robotics, which tend toward assistive systems or versatile collaborative robots, could make a reorganization of industrial sites possible in order to preserve the health of human workers, by improving their working conditions, while enhancing the value of human expertise and skills. Such a transformation of work requires to rethink robotic control approaches, both in terms of safety and physical capabilities, to enable the robots to help humans in their activity and evolve in a shared environment. The conventional architectures of industrial robots, which are heavy and rigid, must also be adapted to the needs of interaction with humans (safety, perception), in the present context of climate change that cannot be ignored (energy consumption).

To meet such functional requirements, the development of the physical cobot architecture and its controller cannot be separated. Codesign methodologies [Fad22] address this dependency by cyclically optimizing the mechatronic design and the control of the robotic system. Codesign approach exploits dynamic modeling tools (Pinocchio [Carp19]), physical simulation engines (Gazebo [Koe04], Simulation and Active Interfaces [Sai2]), and trajectory planners (MoveIt! [Gor19], Crocoddyl [Mas20]) to predict the behavior of a robotic architecture associated with a control law. But it remains difficult to perfectly model the physics of the system and the interaction, to predict the influence of kinematic or inertial parameters of the robot architecture on the evolution of its control law, or to simulate the impact of control parameters on the robot real-world behavior. Experimental evaluation is essential to validate developments based on simulations and to quantify induced errors. However, it is often limited to controller evaluations on commercial robots, or to the validation of a full and fixed prototype at the end of the design phase.

Research activities

This internship project aims at designing a modular and reconfigurable robotic manipulator to improve experimental evaluation of cobotic approaches (design and control). The robot will be composed of a set of standardized joints and segments, which can be assembled together to form different architectural configurations (number and orientation of joints). A number of kinematic (length, joint position, etc.) and inertial (position of the center of mass, inertia, etc.) parameters of the links must be adjustable, to assess their impact on the behavior of the robotic arm. The joint modules will include a generic guidance solution and integrate several types and sizes of actuation (motors/transmission systems).

A prototype will be manufactured and assembled during the internship. Open-source solutions can be integrated for the low-level motor control (control boards [ODRI], SimpleFOC library [Sku22], etc.). The robotic manipulator will be described and modeled using the URDF standard and interfaced with the laboratory controllers (Qontrol library) via ROS.

The project goal is to set up a modular, reconfigurable, open and shared experimental platform for a comprehensive and parameterized evaluation of control approaches and mechanism optimization strategies.

Research environment

This research internship will take place within the auctus team at the Inria center of the University of Bordeaux (Talence). The auctus team aims at meeting challenges of collaborative robotics for humans at work. The team’s research is divided into three scientific axes: analysis and modeling of human behavior (biomechanical and cognitive); human-robot interaction and coupling; design and control of cobotic systems. This internship falls within the scope of the third axis, by participating in the implementation and evaluation of joint robotic design/control approaches. The project will benefit from the hardware and software resources of the lab experimental platform. The student will be supervised throughout the internship by the three supervisors.

Skills

The candidate should have solid skills in mechanical design, CAD, robot modeling, prototyping, motor/sensor integration. Some additional experiences in robotic control, programming (C++, Python), and ROS would be appreciated.

Contacts

To apply at this offer, please email a detailed CV, a motivation letter, and your Master courses/grades to the three supervisors:

References

[Carp19] CARPENTIER, Justin, SAUREL, Guilhem, BUONDONNO, Gabriele, et al. The Pinocchio C++ library: A fast and flexible implementation of rigid body dynamics algorithms and their analytical derivatives. In : 2019 IEEE/SICE International Symposium on System Integration (SII). IEEE, 2019. p. 614-619.

[Fad22] FADINI, Gabriele, FLAYOLS, Thomas, DEL PRETE, Andrea, et al. Simulation aided co-design for robust robot optimization. IEEE Robotics and Automation Letters, 2022, vol. 7, no 4, p. 11306-11313.

[Koe04] KOENIG, Nathan et HOWARD, Andrew. Design and use paradigms for gazebo, an open-source multi-robot simulator. In : 2004 IEEE/RSJ international conference on intelligent robots and systems (IROS)(IEEE Cat. No. 04CH37566). IEEE, 2004. p. 2149-2154.

[Gor19] GÖRNER, Michael, HASCHKE, Robert, RITTER, Helge, et al. Moveit! task constructor for task-level motion planning. In : 2019 International Conference on Robotics and Automation (ICRA). IEEE, 2019. p. 190-196.

[Mas20] MASTALLI, Carlos, BUDHIRAJA, Rohan, MERKT, Wolfgang, et al. Crocoddyl: An efficient and versatile framework for multi-contact optimal control. In : 2020 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2020. p. 2536-2542.

[Sku22] SKURIC, Antun, BANK, Hasan Sinan, UNGER, Richard, et al. SimpleFOC: A Field Oriented Control (FOC) Library for Controlling Brushless Direct Current (BLDC) and Stepper Motors. Journal of Open Source Software, 2022, vol. 7, no 74, p. 4232.

[Sai2] Stanford Simulation and Active Interfaces framework, https://github.com/manips-sai-org/

[ODRI] Open Dynamic Robot Intiative Master board et Microdriver https://github.com/open-dynamic-robot-initiative