1 of 5

Research Objective: To develop offline-coordination planning techniques for multi-robot systems with partially shared workspaces that enables them to allocate and perform manufacturing tasks in a time-effective and computationally efficient manner.

Task Allocation and Dead-Lock-Free Trajectory Planning for Collaborative Multi-Robot System

Collaborative robots performing a manufacturing task

Mohammad Zainullah Khan

Advisors: Andrew Murray, Ph.D & David Myszka, Ph.D

Department of Mechanical & Aerospace Engineering

Robot 1

Robot 3

Robot 2

Divide task into raster segments

Allocate rasters to each robot using Tabu Search Heuristic

Perform Combinatory Collision checking

Introduce delays

Resolve Deadlocks

Path Modification

Task-time calculation

Tabu Search Heuristic

Tabu Search Heuristic is employed for task allocation and to find the local optimal path with minimum timespan.

Combinatory Collision Checking

Generally, for m robots, each with n links:

Future Work

The coordination planning techniques now need to be implemented on a 5/6 DOF multi-robot systems for manufacturing environment.

Applications

  • Spray Painting
  • 3D Printing (Additive Manufacturing)
  • Pressure Washing
  • Loading Docks (Object manipulation)

Modeling of three 3P robots in MATLAB

Collision�Checks

=

Reference:�https://www.therobotreport.com/top-5-countries-using-industrial-robots-2018/

2 of 5

3 of 5

Research Objective: To develop offline-coordination planning techniques for multi-robot systems with partially shared workspaces that enables them to allocate and perform manufacturing tasks in a time-effective and computationally efficient manner.

Task Allocation and Dead-Lock-Free Trajectory Planning for Collaborative Multi-Robot System

Modeling of three 3P robots in MATLAB

Collaborative robots performing a manufacturing task

Mohammad Zainullah Khan

Advisors: Andrew Murray, Ph.D & David Myszka, Ph.D

Department of Mechanical & Aerospace Engineering

Robot 1

Robot 3

Robot 2

Divide task into raster segments

Allocate rasters to each robot using Tabu Search Heuristic

Go through the robots’ motion

Perform Combinatory Collision checking

Collisions Detected?

Introduce delays iteratively in every robot’s path to resolve collisions

Select the minimum delay

Check for deadlocks

Iteratively modify each robot’s path to resolve deadlock

Select path modification leading to minimum increase in task time

Restart collision checking from start

Generate a matrix of each robot’s end-effector and time

Calculates the task-time

Animates the robots

Flowchart for Offline Path Planning with task-allocation, collision avoidance & deadlock resolution

Tabu Search Heuristic

Tabu Search Heuristic is employed for task allocation and to find the local optimal path with minimum timespan.

Combinatory Collision Checking

Generally, for m robots, each with n links:

4 of 5

Research Objective: To develop offline-coordination planning techniques for multi-robot systems with partially shared workspaces that enables them to allocate and perform manufacturing tasks in a time-effective and computationally efficient manner.

Applications

  • Spray Painting
  • 3D Printing (Additive Manufacturing)
  • Pressure Washing
  • Loading Docks (Object manipulation)

Simplified Problem

Two DOF cranes collaborating to arrange boxes in a specific order

References

“Performance of Collaborative Robot Systems.”NIST,8 May 2018, https://www.nist.gov/programs-projects /performance-collaborative-robot-systems.

Hieu Bui, Harry A. Pierson, Sarah Nurre Pinkley & Kelly M. Sullivan (2021) Toolpath planning for multi-gantry additive manufacturing, IISE Transactions, 53:5, 552-567, DOI: 10.1080/24725854.2020.1775915

Task Allocation and Dead-Lock-Free Trajectory Planning for Collaborative Multi-Robot System

Task Allocation & Path Planning

Tabu Search Heuristic is employed for task allocation and to find the optimal path with minimum timespan.

Optimal Path with collisions

Collision Avoidance

In order to avoid collisions, delays are introduced in either crane 1 or crane 2’s path (which ever results in lower timespan).

Pick and arrange four boxes using two cranes in the optimum manner without collisions.

Three collaborative robots performing a manufacturing task

Mohammad Zainullah Khan

Advisors: Andrew Murray, Ph.D & David Myszka, Ph.D

Department of Mechanical & Aerospace Engineering

5 of 5

3rd robot introduced

3rd Robot’s lengths & attachment point defined

3rd robot drawn

Incorporate Robot motion

Include in Tabu Search Heuristic

Perform Combinatory Collision Checking

Introduce delays to resolve collisions

Check for any deadlocks

Resolve Deadlocks

Make Time-vs-Coordinates matrix

Show collaborative motion

Robot 1

Robot 3

Robot 2