1 of 1

Multi-Robot Testbed for Search

and Rescue Mission

11th Annual COE Graduate Poster Presentation Competition

Student: Justin Blackman (MS)

Advisors: Dr. Redwan and Dr. Karimoddini

Cross-Disciplinary Research Area: Aerospace and Transportation Systems

MOTIVATION

OBJECTIVE

PROBLEM DESCRIPTION

REFERENCES

  • To fully integrate robotic platforms in the Robotic Operating System (ROS) using an Unmanned Ground Vehicle (UGV), a Robotic Arm, and a Drone that work cooperatively to complete a Rescue mission.
  • Develop a decentralized and formal method for task allocation in a search and rescue mission by combining market-based task allocation and behavior tree decision making.
  • A multi-robot platform can offer more functionality when robots with different sets of capabilities can complement each other to carry out a cooperative mission.
  • A distress signal is received from an injured person in an environment. The mission is to search for and deliver a care package to that person.
  • Multiple robots will be deployed to complete the mission.
  • Tasks are assigned to the robots subject to the most efficient allocation of resources.
  • The end goal is to demonstrate the ability of autonomous robotic platforms to work cooperatively to complete shared missions in an effective manner.

ROBOT AGENTS

  • Turtlebot
    • Ground vehicle that is able to search environment.
  • Parrot Bebop Drone
    • Aerial vehicle that is able to search environment.
  • Jaguar Robot w/ Robotic Arm
    • Ground vehicle that is able to deliver a package using attached robotic arm.

Figure 9:Turtlebot

Figure 11: Jaguar

Figure 10: Parrot Bebop Drone

Figure 1: Robot Collaboration

  • Wang, T., Shi, D., & Yi, W. (2020, November). Extending Behavior Trees with Market-Based Task Allocation in Dynamic Environments. In Proceedings of the 2020 4th International Symposium on Computer Science and Intelligent Control (pp. 1-8).
  • Burgard, W., Moors, M., Stachniss, C., & Schneider, F. E. (2005). Coordinated multi-robot exploration. IEEE Transactions on robotics21(3), 376-386.
  • Tadewos, Tadewos G., Laya Shamgah, and Ali Karimoddini. "Automatic Decentralized Behavior Tree Synthesis and Execution for Coordination of Autonomous Vehicles."
  • Tadewos, Tadewos G., Laya Shamgah, and Ali Karimoddini. "Behaviour Tree Based Decentralized Coordination of Autonomous Vehicles."

MARKET-BASED TASK ALLOCATION

BEHAVIOR TREE DECISION MAKING

RESCUE MISSION TASKS

  1. Search for and locate an injured person in an indoor environment.
  2. Deliver a care package to the location of the injured person.
  3. Return each robot to it’s respective base.
  4. The entire Rescue mission including tasking, planning, coordination, and control will be conducted autonomously.

Mission : T1, T2 ... TN

Market Based Auctioning level 1

Market Based Auctioning level 2

Task Assigment

Local BTi Synthesis

Task Execution

Resources are available

Tj

fij

Ak

fdk

Proceed to the next task

Figure 2: Auctioning Level-1

Figure 3: Auctioning Level-2

Figure 4: Market-based Auctioning

Search – deploys agent to search for P.

Transmit – transmits P.

Grab – holds CP in robot arm.

WaitTransmit – waits for P TxRx.

MoveTo – moves agent to P.

Place – places CP down at P.

Figure 6: Rescue Mission Action Bank

Figure 7: Synthesized BT for Search Action

Figure 8: Synthesized BT for Deliver Action

Figure 5: Legend