This project follows the work started for the implementation of human-like movement planning and control on a single arm-hand anthropomorphic robot system, in a first version, and later coordinated asynchronous bimanual movement planning. The next step and the main goal of this work is to endow ARoS with coordinated synchronous bimanual movements. These bimanual movements have to be collision free and human-like.
In a summary form, the objectives for this project are the following:
· Study of the human motor control for an better understanding of coordination and cooperation between the two arms;
· Definition of bimanual tasks and related grips;
· Extension of the posture-based model to the bimanual manipulation problem;
· Adapting the new model to changes in the scenario;
· Modelling new temporal and spatial constraints which have to be introduced in the optimization problem in order to satisfy requirements of synchronization, essentials to accomplish cooperative bimanual tasks;
· Extending the optimization problem in order to take into account the end-state comfort effect: introduction of a way to formulate the comfort of a posture;
· Modelling the dominant arm-hand behavior often present in dual-arm manipulation in humans.
Task 1 – State of the Art
· Bibliographic review about currently existing work on bimanual manipulation in robotics;
· Bibliographic review about human motor control in order to understanding the bimanual manipulation.
Task 2 – Study of the theoretical concepts implemented in ARoS
· Direct and inverse kinematics of the two robotic arms and hands of ARoS.
· Review of the ARoS’s cognitive architecture.
Task 3 – Study the method of movement planning and optimization applied to the ARoS
· Study the code of the movement planning applied for the bimanual asynchronous movements and the non-linear optimization approach.
Task 4 – Improve the robustness of the optimization process
· Switch the mathematical software AMPL by OPTI (that is a MATLAB toolbox for constructing and solving linear, nonlinear, continuous and discrete optimization problems).
Task 5 – Extend the optimization process applied to ARoS
· Extending the optimization problem in order to take into account the end-state comfort effect: introduction of a way to formulate the comfort of a posture;
Task 6 – Implement the mechanism that adapts the movements to changes in the scenario
· Endow the system to update the movement planning during the movement of the arm, in order to consider possible changes in the environment.
Task 7 – Implement synchronous bimanual coordination in ARoS
· Modelling new temporal and spatial constraints which have to be introduced in the optimization problem in order to satisfy requirements of synchronization, essentials to accomplish cooperative bimanual tasks;
· Modelling the dominant-arm behavior often present in dual-arm manipulation in humans.
Task 8 – Final tests
· Realization of physical tests with ARoS robot.
Task 9 – Write/ Review Dissertation
· Writing and constant review of the Dissertation.