Wireless Sensor Networks (WSNs) are a relatively new technology with many potential applications, including military and homeland security surveillance operations. Accurate classification of WSN contacts has been attempted using various sensor combinations over the past few years, yet video and photographic imagery remain the only choices for attaining context specific contact classification. While cameras have been successfully installed within some WSNs, there are serious limitations to this solution. Most stemming from the scarce power resources, immobility, and small form factor common among conventional WSN nodes. An efficient, low cost answer to this problem involves the use of unmanned aerial vehicles (UAVs) to acquire imagery of WSN contacts. For this system to scale to the wide expanses that WSNs deploy over, UAV contact surveillance operations must be controlled autonomously. The objective of this thesis is to research and implement an autonomous UAV WSN system, where an optimized two-dimensional flight plan is produced in response to WSN contact detection. Flight plans autonomously guide the UAV on a course to either an estimated interception point with the WSN contact or to the instigated WSN cluster, depending upon user input. The event driven application produced in this study functions in the periphery of the Kestrel Autopilot System, communicating flight plans to the UAV through properly crafted Kestrel packets.
| Identifer | oai:union.ndltd.org:nps.edu/oai:calhoun.nps.edu:10945/2663 |
| Date | 09 1900 |
| Creators | Schall, Stephen A. |
| Contributors | Singh, Gurminder, Vaidyanathan, Ravi, Naval Postgraduate School (U.S.)., Department of Computer Science |
| Publisher | Monterey, California. Naval Postgraduate School |
| Source Sets | Naval Postgraduate School |
| Detected Language | English |
| Type | Thesis |
| Format | xx, 169 p. : col. ill. ;, application/pdf |
| Rights | Approved for public release, distribution unlimited |
Page generated in 0.0013 seconds