Towards Modeling The Behavior Of Physical Intruders In A Region Monitored By A Wireless Sensor Network
| dc.contributor | Krishnamoorthy, Pranav | en_US |
| dc.date.accessioned | 2010-07-19T19:55:02Z | |
| dc.date.accessioned | 2011-08-24T21:43:47Z | |
| dc.date.available | 2010-07-19T19:55:02Z | |
| dc.date.available | 2011-08-24T21:43:47Z | |
| dc.date.issued | 2010-07-19 | |
| dc.date.submitted | January 2010 | en_US |
| dc.description.abstract | A priority task for homeland security is the coverage of large spans ofopen border that cannot be continuously physically monitored forintrusion. Low-cost monitoring solutions based on wireless sensornetworks have been identified as an effective means to perform perimetermonitoring. An ad-hoc wireless sensor network scattered near aborder could be used to perform surveillance over a large area withrelatively little human intervention. Determining the effectiveness of such an autonomous network in detecting and thwarting an intelligent intruder is a difficult task. We propose a model for an intelligent attacker that attempts to find a detection-free path in a region with sparse sensing coverage. In particular, we apply reinforcement learning (RL) - a machine learning approach, for our model. RL algorithms arewell suited for scenarios in which specifying and finding an optimalsolution is difficult. By using RL, our attacker can easily adapt to newscenarios by translating constraints into rewards. We compare ourRL-based technique to a reasonable heuristic in simulation. Our resultssuggest that our RL-based attacker model is significantly moreeffective, and therefore more realistic, than the heuristic approach. | en_US |
| dc.identifier.uri | http://hdl.handle.net/10106/4947 | |
| dc.language.iso | EN | en_US |
| dc.publisher | Computer Science & Engineering | en_US |
| dc.title | Towards Modeling The Behavior Of Physical Intruders In A Region Monitored By A Wireless Sensor Network | en_US |
| dc.type | M.S. | en_US |