Flow-based Cyber Physical Systems

Abstract:

Monitoring flow-based systems (FBS) (e.g., water distribution systems) is of paramount importance, due to their economic and health impacts. FBS monitoring has been typically achieved by strategically placed, costly and complex, static sensors. To reduce the cost of monitoring, we propose a mobile wireless sensor network (WSN) system comprised of mobile sensors (their movement aided by the inherent flow) and static beacons which aid locating sensors. This project presents the first complete architectural design, algorithms, and protocols for optimal monitoring of FBS. Our proposed solution includes sensing and communication models, MAC and group management protocols for sensor and beacon communication, and algorithms for sensor and beacon placement. We compare our proposed solution with state of art through extensive simulations and a proof-of-concept system implementation. We demonstrate performance improvements, such as a dramatic reduction (a factor of 91) in the number of sensors when sensing range is marginally (2.5 times) increased.

I. OVERVIEW
Flow-based systems (FBS) are physical systems, such as water distribution systems (WDS), oil & gas pipelines, and human cardiovascular system, that can be modeled as a flow network in which a fluid, e.g., water, oil or blood, flows through the edges of the network. Such systems are vulnerable to attacks that can potentially have severe health and economic impacts (e.g., blockage of arteries in a human circulatory system is a critical problem that needs to be identified in early stages to avoid the risk of heart attack). Monitoring such systems to identify these attacks is of paramount importance. Recent advancements in nanotechnology and wireless sen- sor networks enable the design of monitoring systems using mobile devices [3]. WDS monitoring has been traditionally solved using static sensors placed strategicically. Static sensors are often expensive and fail to effectively cover the entire WDS, since the sensors are sparsely distributed and the WDS are, typically, large scale and complex. Therefore, monitoring such systems with minimal monitoring infrastructure (while guaranteeing low false positives and negatives for detecting the attack and accurate attack localization) is of paramount im- portance. As a consequence, mobile sensors traversing pipeline networks (e.g., in water pipelines [7] [5]) have been developed. Most previous solutions, however, do not aim at reducing the number of devices used. Only recently, there has been research on optimal FBS monitoring using mobile sensors. In [11], the authors propose a preliminary solution for event detection and localization using mobile sensors and static beacons in acyclic flow networks, such as WDS. Mobile sensors are injected into an acyclic flow network at select points (e.g., where manholes are present). Beacons are static devices, placed strategically, used for mobile sensor localization. While sensors are propelled by the flow of fluid in the flow-based system, they detect events. The authors proved that the problem of optimal monitoring of acyclic flow networks is a NP-Hard problem. In [11], the authors consider a simplistic binary sensing model that imposes a large cost on the number of sensors needed to ensure a maximum false negative rate (as set by a user) in event detection. [11] also requires the physical capture of sensors due to the absence of communication among sensors and beacons.
We propose to enable sensor to sensor and sensor to beacon communication, along with a more realistic sensing model, to address the limitations of current solutions [11]. These additions raise several interesting research questions. Sensor- sensor communication can potentially have a large number of collisions in environments with high propagation delays, such as a WDS, where acoustic communication is used. In addition, the randomness in node movement can lead to high inefficiency in sensor-sensor communication. The first research question that we aim to address is to design a MAC protocol and group management protocols to maximize the benefit of communication among sensors. Additionally, [11] is only concerned with where and how many sensors are deployed. However, we can leverage sensor-sensor communication to aid event localization. Therefore, we also consider when to insert the sensors in order to ensure that there is a signifi- cant amount of sensor-sensor communication among sensors traversing different paths. We explore how the new sensing and communication models affect the algorithms for FBS monitoring (i.e., sensor and beacon placement). Finally, we encapsulate the different pieces of the system in an integrated architecture for FBS monitoring. The integrated solution for FBS monitoring consists of models for sensing, communication, a group management protocol for sensor nodes and algorithms for the deployment of the infrastructure, such that user requirements (e.g., for maximum false negative rate of event detection and maximum event localization error) are met. The contributions of our project are:

* We present the first integrated architecture for FBS monitoring using mobile WSN.
* We propose the first sensing and communication mod- els, and the first group management protocol for FBS monitoring.
* We present the first algorithms to solve the problem of monitoring a FBS using sensors with communication and more sophisticated sensing capabilities.
* We demonstrate the feasibility of our solution through a proof-of-concept system implementation and testbed evaluation. We demonstrate its efficiency in large networks through extensive simulations.

REFERENCES
[1] N. Aitsaadi, N. Achir, K. Boussetta, and G. Pujolle. Differentiated underwater sensor network deployment. In OCEANS, 2007.

[2] K. Brumbelow, J. Torres, S. Guikema, E. Bristow, and L. Kanta. Virtual cities for water distribution and infrastructure system research. In World Environmental and Water Resources Congress, 2007.

[3] J. C. Claussen, A. D. Franklin, A. ul Haque, D. M. Porterfield, andT. S. Fisher. Electrochemical biosensor of nanocube-augmented carbon nanotube networks. ACS Nano, 3(1), 2009.

[4] K. Hunt, J. Niemeier, and A. Kruger. RF communications in underwater wireless sensor networks. In EIT, 2010.

[5] J. Kim, G. Sharma, N. Boudriga, and S. Iyengar. Spamms: a sensor- based pipeline autonomous monitoring and maintenance system. In COMSNETS, 2010.

[6] G. Kokosalakis, A. Gorlov, E. Kausel, and A. Whittle. Communications and power harvesting system for in-pipe wireless sensor networks. US Patent App. 20,070/209,865.

[7] T. Lai, W. Chen, K. Li, P. Huang, and H. Chu. Triopusnet: automat- ing wireless sensor network deployment and replacement in pipeline monitoring. In IPSN, 2012.

[8] X. Li, W. Yu, X. Lin, and S. Iyengar. On optimizing autonomous pipeline inspection. Robotics, IEEE Transactions on, 28(1), feb. 2012.

[9] J. Ma and W. Lou. Interference-aware spatio-temporal link scheduling for long delay underwater sensor networks. In SECON, june 2011.

[10] Z. Sun, P. Wang, M. C. Vuran, M. A. Al-Rodhaan, A. M. Al-Dhelaan, and I. F. Akyildiz. MISE-PIPE: Magnetic induction-based wireless sensor networks for underground pipeline monitoring. Ad Hoc Netw., May 2011.

[11] M. Suresh, R. Stoleru, E. Zechman, and B. Shihada. On event detection and localization in acyclic flow networks. Systems, Man, and Cybernetics: Systems, IEEE Transactions on, 43(3), 2013.

[12] D. Trinchero and R. Stefanelli. Microwave mobile sensor networks within underground conduits filled of fluids. In Electromagnetic Theory, International Symposium on, 2010.

[13] R. Urick and R. Urick. Principles of underwater sound, volume 3.
McGraw-Hill New York, 1983.

[14] A. J. Whittle, L. Girod, A. Preis, M. Allen, H. B. Lim, M. Iqbal, S. Sri- rangarajan, C. Fu, K. J. Wong, and D. Goldsmith. WATERWISE@SG: A testbed for continuous monitoring of the water distribution system in singapore. In WSDA, 2010.

[15] S. Xiong, L. Yu, H. Shen, C. Wang, and W. Lu. Efficient algorithms for sensor deployment and routing in sensor networks for network- structured environment monitoring. In INFOCOM, 2012.

[16] A. Yazdani and P. Jeffrey. Complex network analysis of water distribution systems. Arxiv preprint arXiv:1104.0121, 2011.