Holly H. Jin Postdoctoral Researcher
	Systems Optimization Laboratory Department of Management Science & Engineering

Contact information

Terman Engineering Center, Rm 325
380 Panama Way
Stanford, CA 94305
E-mail: hollyjin@stanford.edu

Research interests

PhD thesis research

Localization using Peer-to-Peer Communication: A Suite of Algorithms for 2D, 3D, Real-time and Distributed Network Localization

The market for real time location service market is quickly expanding and is estimated to reach $1.26 billion in 2011. Many localization applications require accuracy and coverage that are greater than simple triangulation or trilateration approaches can deliver. Even GPS-based localization is insufficiently accurate in some applications; and of course, it is ineffective in obstructed environments or ultra-low-power deployments. An alternative approach is to utilize distances not just to nearby base stations or anchors but also to nearby peers to improve localization accuracy. Yet even for small peer to peer networks, determining nodes' locations accurately and efficiently has been a difficult computational problem. A suite of high-performance algorithms has been developed at Stanford University to estimate node positions for wireless sensor networks. These algorithms are also applicable to WiFi networks, cell phone networks, RFID networks, and public safety networks. The algorithms are capable of estimating node positions in static, dynamic, and distributed networks of arbitrary size. These algorithms help achieve the speed necessary for large real-time network applications without sacrificing accuracy in the estimated locations.

Advantages:

• Speed: capable of localizing ten thousand nodes within a minute (with Matlab implementation)
• Scalability: near-linear complexity
• Real-time Capability: fast and accurate node localization and tracking in real-time
• Accuracy: can localize low-anchor-density networks traditional methods cannot, and special optimization routine effectively combat noisy measurement data

Software

• Software available for localization in 2D, 3D, real-time tracking and in distributed networks.
• First place winner in software category of The Stanford-Berkeley Innovators' Challenge, a joint technology contest organized by BASES, the Business Association of Stanford Engineering Students, and VERTEX, the Engineering Entrepreneurship Club at UC Berkeley.
• For a demo or further information, please contact: hollyjin@stanford.edu.

Publications

• Localization paper: SpaseLoc: An adaptive subproblem algorithm for scalable wireless sensor network localization, SIAM Journal on Optimization 17(4), 1102-1128 (2006).
• Location Intelligence 2007 conference presentation: Slides
• SpaseLoc Demo Figures: Fig1_LowSensorFullSDP. Fig2_LowSensorSpaseLoc. Fig3_LowAnchorFullSDP36Nodes. Fig4_1_LowAnchorSpaseLoc36Nodes. Fig4_2_LowAnchorSpaseLoc36Nodes. Fig4_3_LowAnchorSpaseLoc36Nodes. Fig4_4_LowAnchorSpaseLoc36Nodes. Fig4_5_LowAnchorSpaseLoc36Nodes. Fig4_6_LowAnchorSpaseLoc36Nodes. Fig4_7_LowAnchorSpaseLoc36Nodes. Fig4_8_LowAnchorSpaseLoc36Nodes. Fig4_9_LowAnchorSpaseLoc36Nodes. Fig5_LowAnchorSpaseLoc1000Nodes. Fig6_NoiseDataFullSDP. Fig7_NoiseDataFullSDPWithRef. Fig8_NoiseDataSpaseLoc. Fig9_NoiseDataSpaseLocWithRef. Fig10_1_SpaseLocDistributed. Fig10_2_SpaseLocDistributed. Fig10_3_SpaseLocDistributed. FigFig10_3_SpaseLocDistributed. Fig14_SpaseLocDynamic1Moving. Fig15_SpaseLocDynamic10Moving. Fig16_SpaseLoc3D.

Other affiliations

• Cardinal Optimization Inc., scientist.
• Stanford Racing Team, team member in Planning and Optimization Group.
• Systems Optimization Laboratory, Stanford University, conducted my PhD thesis, then studied as a postdoctoral researcher.
• Nortel, worked on various telecommunication products.

Updated: April 24, 2007