Home

Syllabus

Projects

Books and Readings

Assignments

Quizzes

Academic Integrity

About this site

GPS Sensor Array Management

Client

Brady O'Hanlon, graduate student, Electrical and Computer Engineering
bwo1@cornell.edu

Background

A GPS receiver is being developed for use in large arrays for ionospheric monitoring. The hardware consists of a TI Digital Signal Processor, a single-board computer running Linux, and other custom hardware. It is envisioned that these receivers (nodes) will operate semi-autonomously in the field, in arrays of 10, 100, or more. Data are produced continuously by each node (on the order of 10 KB per second per node).

Requirements

1. The data must be reliably collected from all nodes in a particular subnet and either stored onto a monitoring PC, then sent to an analysis center (i.e., central server) or uploaded directly to the server. Data transmission to the analysis center will be via the internet.
2. Monitoring of data quality and receiver health would be useful.
   
3. Dynamic node reconfiguration (both enabling/disabling nodes, and sending short command messages).
4. A graphical interface for both node reconfiguration and network monitoring.

Restrictions

1. The single-board computer is both memory and CPU constrained (this piece of hardware cannot be changed unless a similarly-capable, less-expensive alternative is found).
2. Both data integrity and latency are important.
3. Possible network interfaces include Ethernet, WiFi, ZigBee.

Questions

1. What are the data rate limits for various network configurations?
2. What are the tradeoffs (in power, cost, number of sustainable nodes, geographic node distribution) for various network interfaces?
3. Can some lossless compression algorithm (e.g., LZW) be implemented (without exceeding CPU limitations) to increase throughput?