Visible to the public Biblio

Filters: Keyword is 1521617  [Clear All Filters]
Schmittle, Matt, Lukina, Anna, Vacek, Lukas, Das, Jnaneshwar, Buskirk, Christopher P., Rees, Stephen, Sztipanovits, Janos, Grosu, Radu, Kumar, Vijay.  2018.  OpenUAV: A UAV Testbed for the CPS and Robotics Community. Proceedings of the 9th ACM/IEEE International Conference on Cyber-Physical Systems. :130–139.

Multirotor Unmanned Aerial Vehicles (UAV) have grown in popularity for research and education, overcoming challenges associated with fixed wing and ground robots. Unfortunately, extensive physical testing can be expensive and time consuming because of short flight times due to battery constraints and safety precautions. Simulation tools offer a low barrier to entry and enable testing and validation before field trials. However, most of the well-known simulators today have a high barrier to entry due to the need for powerful computers and the time required for initial set up. In this paper, we present OpenUAV, an open source test bed for UAV education and research that overcomes these barriers. We leverage the Containers as a Service (CaaS) technology to enable students and researchers carry out simulations on the cloud. We have based our framework on open-source tools including ROS, Gazebo, Docker, PX4, and Ansible, we designed the simulation framework so that it has no special hardware requirements. Two use-cases are presented. First, we show how a UAV can navigate around obstacles, and second, we test a multi-UAV swarm formation algorithm. To our knowledge, this is the first open-source, cloud-enabled testbed for UAVs. The code is available on GitHub:

D. Orol, J. Das, L. Vacek, I. Orr, M. Paret, C. J. Taylor, V. Kumar.  2017.  An aerial phytobiopsy system: Design, evaluation, and lessons learned. 2017 International Conference on Unmanned Aircraft Systems (ICUAS). :188-195.
Anastasia Mavridou, Joseph Sifakis, Janos Sztipanovits.  2017.  WebGME-BIP: A Design Studio for Modeling, Analyzing and Generating Systems with BIP.

When building large concurrent systems, one of the key difficulties lies in coordinating component behavior and, in particular, managing the access to shared resources of the execution platform. Components may interact through buses, message buffers, etc. leading to resource contention and potential deadlocks compromising safety-critical operations. The concurrent nature of such interactions is the root cause of the complexity of the resulting software. Thus, the complexity of software systems is exponential in the number of their components, making a-posteriori verification of their correctness practically infeasible. An alternative approach, taken by the BIP framework, consists in ensuring correctness-by-construction by applying automatic transformations to obtain executable code from formally defined models. Following this latter approach, we have designed and implemented a BIP design studio. We have studied extensions of the BIP language for specifying parameterized models and integrated them in the design studio to enhance scalability, reusability, and reduce model size. Additionally, we have studied and implemented a set of necessary and sufficient conditions for validating the consistency and encodability of BIP models at design time. We have developed code generation plugins from graphical BIP models to equivalent Java and BIP code. The generated BIP code can be verified for deadlock-freedom or safety properties using compositional verifications tools offered by the BIP framework.

Anastasia Mavridou, Joseph Sifakis, Janos Sztipanovits.  2017.  A Design Studio for Architecture-based Design with BIP.

The Behavior-Interaction-Priority (BIP) framework, rooted in rigorous semantics, allows modeling heterogeneous component-based systems. BIP is supported by a textual modeling language, as well as a tool-set including run-time platforms and verification tools. We present a web-based design studio that allows specifying BIP behavior and interaction models in a purely graphical way and generating the equivalent textual specifications. To facilitate scaling and reusability of BIP models, we have extended architecture diagrams, a graphical language for modeling architecture styles, to define parameterized BIP models. We present the various services provided by the design studio, including model repositories, design guidance mechanisms, code generators, and integration with the BIP tool-set.

L. Vacek, E. Atter, P. Rizo, B. Nam, R. Kortvelesy, D. Kaufman, J. Das, V. Kumar.  2017.  sUAS for Deployment and Recovery of an Environmental Sensor Probe. IEEE International Conference on Unmanned Aircraft Systems (ICUAS) 2017.

Small Unmanned Aircraft Systems (sUAS) are already revolutionizing agricultural and environmental monitoring through the acquisition of high-resolution multi-spectral imagery on-demand. However, in order to accurately understand various complex environmental and agricultural processes, it is often necessary to collect physical samples of pests, pathogens, and insects from the field for ex-situ analysis. In this paper, we describe a sUAS for autonomous deployment and recovery of a novel environmental sensor probe. We present the UAS software and hardware stack, and a probe design that can be adapted to collect a variety of environmental samples and can be transported autonomously for off-site analysis. Our team participated in an NSF-sponsored student unmanned aerial vehicle (UAV) challenge, where we used our sUAS to deploy and recover a scale-model mosquito trap outdoors. Results from indoor and field trials are presented, and the challenges experienced in detecting and docking with the probe in outdoor conditions are discussed.

Jonathan Sprinkle, Chris vanBuskirk, Stephen Rees, Jnaneshwar Das, Vijay Kumar, Joris Kenanian, Paulo Tabuada.  2017.  Compiling CPS Model Repositories through Student Competitions. 2nd Workshop on Monitoring and Testing of Cyber-Physical Systems.

This talk describes how the Cyber-Physical Systems Virtual Organization (CPS-VO) is hosting competitions for the purpose of improving CPS verication tools. We describe the 2016 Challenge, which focused on quadrotor control and codesign of payload, and the 2017 Challenge which focuses on populating a ground vehicle simulator with realistic obstacles. In addition, the interfaces by which participants compete are described, in order to articulate the means by which models can be decoupled from the system for the purposes of evaluation by external tools.