Multiple task execution algorithm for whole-body control of humanoid robots

Motivation

1. Discontiunous transition between tasks leads to unstable and inefficient operation of robot.

2. Multiple Tasks Exacution is considered as significant control issue in compelx robotic system when robot interacts between enviornments.

3. In humanoid robot, multiple tasks must be contorolled in uncertain environments continuously.

robot_task

 

 

 

 

 

 

 Goals

When multiple tasks are exacuted in sequence, the cases which tasks are added or removed is significant important to deal with robots efficiently.

In these situations, tasks can be transited by the proposed algorithm continuously.

Stablity and efficiency of performing tasks are warrentied by the proposed algorithm.

robot_motion

 

 

 

 

 

 

 

– Specific goals –

1. Validation of continuous task transition algorithm by using 2 DOF or 3 DOF manipulator

2. Validation complex case of continuous task transition algorithm by using 6 DOF manipulator

3. Continous mutiple tasks control by using Two arms manipulator

new_3dof manipulator

 

 

 

Achievements

1. Position control of 2 DOF planar by using velocity motors kinematically

2. Partial Orientation control of 2 DOF planar {including one position and one orientation} by using velocity motors kinematically

3. Verifying the discontinuity of tasks exacution when task is removed or added { [x,y] -> [y,th] : x is removed and th is added }

disandcontinuity

 < Position Variation of end-effector by operational space control >
4. Implementing stable operation near joint limits or avoiding joint limits using activation parameters for operational space control framework.

1stjointactivation
  < Position and velocity of the 1st joint and activation parameter during the experiment of joint limit avoidance >

5. Using the proposed continuous task transition algorithm by de-activating the task along the singular direction.

singularSVD

  < Activation parameter of the task is designed as a function of the singular value by singular value decomposition (SVD) >