Session: 08-11-01: Mobile Robots and Unmanned Ground Vehicles I

Paper Number: 165358

Terrestrial Multi-Robot Testbed for Cluster-Space Navigation Experiments 

Multi-robot connected systems are gaining popularity due to their advantages, including redundancy, increased coverage and throughput, flexible reconfigurability, and spatially diverse functionality. Our lab has demonstrated cluster control in simulation and on an indoor hardware test bed. We require a newer test bed for conducting experiments at scale. The new test bed is based on the “modern” set of hardware and software tools. The experiments include formation control, reshaping of the cluster, and cluster motion control. Central to our strategy are the concepts of considering the n-robot system as a single entity, a cluster, and specifying motions with respect to cluster attributes, such as position, orientation, and geometry. A supervisory operator or real-time pilot specifies and monitors system motion, and centralized control computations are executed with respect to the cluster space variables. Kinematic transforms allow compensation commands to be derived for each individual robot, and they also allow data from a variety of sensor packages to be converted to cluster space state estimates enabling the coordinate motion of the n-robot system.

 

Contributions:

- An upgraded “avionics architecture” for a system of mobile robots based on the modern hardware and software architecture,  Raspberry pi and Robot Operating System (ROS), respectively.

- A ROS based framework for a n-robot control is presented with ability to switch between simulation and real environment.

- Experimental and simulation results for various sized formation and discussion of the performance, limitations, and tradeoffs for the choice of cluster frame placement.

 

Method used - A simple cluster of 3 robots in a triangle configuration is defined by the length of sides of the triangle and the angle between them. The control scheme is designed for a n-robot cluster. RViz (a ROS visualiszation tool) is used simulate and validate the multirobot system motion in a defined cluster in the virtual world. A set of test are then performed on the real hardware. The hardware consists 3 outdoor mobile robots, each of which is equipped with GPS and IMU for its position and orientation sensing,. and an onboard Raspberry Pi compute unit running ROS. The test environment consists of an open field (about 100x100 m2) where the methodology used are as follows:

- Formation control - transient response of the robots moving from home/initial position to the final formation defined by the cluster space.

- Reshaping of the cluster on the fly - a sequence is devised to change the shape of the cluster space paired along with the motion of the cluster.

- Comparing the simulation and the real world performance of 5 robot cluster.

 

Preliminary results validate our cluster space model in the simulation environment. We are currently performing field tests with 3 robot clusters. An additional 2 robots are under construction to support a 5 robot cluster experiments.

Practical applications -  The cluster space control of robots support our ongoing work on Adaptive Navigation (AN) involves modifying a vehicle’s direction or motion path based on measurements taken during movement. In the context of exploring scalar fields, such as temperature or pollutant concentration across a region of interest, adaptive navigation facilitates the identification of locations of interest, like the maximum temperature or the pollutant source, without the need for exhaustive mapping of the entire region. 

Presenting Author: Manoj Sharma Santa Clara University

Presenting Author Biography: Manoj Sharma received his M.S. and Ph.D. degrees from University of Dayton in 2016 and 2020. He is currently a Research Associate in the dept. of Mechanical Engineering, Santa Clara University, Santa Clara, CA, USA. His area of interests include multirobot systems, collaborative robots, and field robotics.

Authors:

Ricky Schober Santa Clara University
Shumeng Li Santa Clara University
Christian Pedrigal Santa Clara University
Weston Tierney Santa Clara University
Yuichiro Neo Megachips LSI USA
Manoj Sharma Santa Clara University
Michael Neumann Santa Clara University
Christopher Kitts Santa Clara University