Citation
Building a Multi-Lateral Cyber-Physical Drone Network using XBee

Material Information

Title:
Building a Multi-Lateral Cyber-Physical Drone Network using XBee
Creator:
Yehua Zhang
Bastian Tenbergen
Publication Date:

Notes

Abstract:
In recent years, the software engineering community have used the term “CyberPhysical System” (CPS) to denote autonomous systems, which collaborate on tasks, no individual CPS can achieve alone ([1-3]). With increasing quality and ubiquity of high-performance communication equipment and computing hardware, previously remote controlled unmanned autonomous vehicles for the military, search and rescue robots, or autonomous cars are becoming more intelligent and can roam the world with minimal human intervention. In consequence, these autonomous CPS are often entrusted with safety-critical behavior. However, this implies new and unforeseen challenges to their development, as the development process must ensure that no human is harmed or injured during operation. In other words, the CPS must be verifiably safe. This research program focuses on the challenges during early stages of development of collaborative CPS, when requirements document assertions the future behavior of the CPS. These assertions can then be used to guide safety assessment techniques such as FHA or FMEA. To this end, we used several small instructional robot kits as a simple surrogate platform to investigate the challenges and possible solutions to CPS development.
Acquisition:
Collected for SUNY Oswego Institutional Repository by the online self-submittal tool. Submitted by Zach Vickery.

Record Information

Source Institution:
SUNY Oswego Institutional Repository
Holding Location:
SUNY Oswego Institution
Rights Management:
All applicable rights reserved by the source institution and holding location.
Embargo Date:
9/13/2020

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

RESEARCH POSTER PRESENTATION DESIGN © 2015 www.PosterPresentations.com Introduction In recent years, the software engineering community have used the term Physical (CPS) to denote autonomous systems, which collaborate on tasks, no individual CPS can achieve alone ([ 1 3 ]) . With increasing quality and ubiquity of high performance communication equipment and computing hardware, previously remote controlled unmanned autonomous vehicles for the military, search and rescue robots, or autonomous cars are becoming more intelligent and can roam the world with minimal human intervention . In consequence, these autonomous CPS are often entrusted with safety critical behavior . However, this implies new and unforeseen challenges to their development, as the development process must ensure that no human is harmed or injured during operation . In other words, the CPS must be verifiably safe . This research program focuses on the challenges during early stages of development of collaborative CPS, when requirements document assertions the future behavior of the CPS . These assertions can then be used to guide safety assessment techniques such as FHA or FMEA . To this end, we used several small instructional robot kits as a simple surrogate platform to investigate the challenges and possible solutions to CPS development . Objectives To further explore the conceptual challenges of Cyber Physical Systems development, specifically human in the loop ( HiL ) control of multiple drones, this project expanded on the previous Student/Faculty Challenge Grant project by incorporating a multi modal communication system between several robots and a single human operator . To this end, we : 1. Determine the global position of each CPS robot in the environment through ultrasonic trilateralization . 2. with regard to other robots. 3. E xchange relative movement information between each robot through Xbee radios. 4. Develop and implement an algorithm to dispatch operator commands within the meshed network. We made use of the ultrasonic indoor position trilaterlization system Mav 09 by Marvelmind Robotics . And as well as XBee 3 radio module which enables the ability that allows us to transmit data through dedicated mesh network . Contact Web https://github.com/tenbergen/CPS Rover/tree/master/dante/Formation Email yzhang9@oswego.edu | bastian.tenbergen@oswego.edu References 1. Wolf, W. (2009) Cyber physical Systems. IEEE Comp. 42(3). 2. Broy , M.; Cengarle , M.; Geisberger , E. (2012) Cyber Physical Systems: Imminent Challenges. Monterey Workshop. 3. Daun , M.; Salmon, A.; Tenbergen, B.; Weyer Solutions for the Engineering of Collaborative Embedded Systems. 2 nd Intl. EITEC WS. Implementation It was crucial to maintain a connection between the server and the robot mesh network . To do so, the first thing we did was to reverse the relation between robots and the human operator . In the previous project, we chose the robot as our server to receive commands and data from the human operator . However, this method isn't ideal for our project because in the event of losing connection, the human operator would not know which robot to connect to if there's a change in the remote network . Yehua Zhang , Dr. Bastian Tenbergen Building a Multi Lateral Cyber Physical Drone Network using XBee A proper network protocol was created as the combination of both UDP and TCP, thus in the event of losing connection to the robot mesh network, our server would be able to start to broadcast UDP signal until an automatically selected master robot is connected from the mesh network, after that the program would switch to TCP for better connection and functionality . These classes serve as the top level of our connection procedure and they always run in a loop so we could recover from the loss of signal safely . A formation class "formation . py" was created to allows our robots squadron to move in a certain formation and maintain their respective safe distance to each other . We used our XBee mesh network to share and respond each position data . This is shown in the image below, in which the robots automatically change formation according to their relative position ( click any image for a demo video ) . Accomplishments Using this research, we successfully added more features on previous platform to study the development of autonomous behavior in cyber physical drones during early stages of development . Moreover, it aids in the identification of challenges, pitfalls, and interaction modes in human in the loop autonomous vehicle control, i . e . when a human must be informed about the possible safety critical decisions made by a robot during autonomous control mode . To this end, we implemented the following features : Customized network protocol allowing to recover from loss of signal ; Universal GUI template to improve future updates ; A mesh network layer to abstract from single robot communication and enable multi robot coordination ; Formation capability to enable robots to move as seemingly one unit, depending on commands received from a single user .