Lee Center for Advanced Networking

Effros’s colleague, Robert McEliece, is developing error-correcting codes that make the transfer of information as accurate and fast as possible. Such error correction must overcome any network noise as well asmulti-user interference.

McEliece applies the recently developed coding and decoding technique known as “turbo” codes to high-speed data networks. Already a proposed international standard for wireless networks, turbo codes may become the standard way to communicate reliably and efficiently on many networks. However, McEliece’s group recently invented a new technique, called “repeat-accumulate” codes (patent pending), that appears to offer significant advantages. McEliece and his group are actively pursuing this possibility.

In cooperation with the Jet Propulsion Laboratory (JPL), McEliece is also working on a theory for organizing networks of remote sensors on distant planets. For example, to explore the surface of Mars, JPL is considering scattering robotic sensors on the surface of the red planet. Each sensor would send its data to an orbiting base station, which would transmit the data back to Earth.

This is a more difficult problem than networking on our own planet. For one thing, each sensor would have a limited power supply, presumably batteries. One challenge, then, is to get as much data from each sensor as possible before the battery is exhausted. So rapid and efficient data transmission will be critical.

Another problem is distance. Each sensor needs to send its data packets to the base station. But if the orbiting base station is too far away, a sensor would instead transmit its data to a neighboring sensor, which could relay the data through successive neighboring sensors to the base station. Each successive transmission is called a “hop,” and the problem is to get as many packets of data to the base station as possible, using multiple hops, without over-using the battery power of any one sensor.

Once all the data is gathered together in an orbiting base station, it will need to be compressed before it is sent to Earth, to eliminate redundancy and to save battery power.