Joint Neighbor Identification and Channel Estimation for Enabling Advanced MAC-PHY Techniques in Ad Hoc Networks
Abstract
Wireless ad hoc networks predominantly use random access schemes for data transmissions. Random accesstechniques are notoriously sub-optimal. The optimal techniques implictly assume the availability of neighborhood information {information about who are a node s neighbors{ and the wireless channel from all neighbors. Such information is not available a priori and is expensive to collect, even for static wireless networks. For dynamic networks, this cost can be prohibitively expensive, even more expensive than data, since this information changes frequently. For this reason, our ad hoc networks use random access techniques and re-main inefficient. In this research, we look at this information collection problem. We present a compressive sensing technique that allows a node to estimate its neighborhood and the channel to its neighbors by using compressive sesning techniques. Every node uses a unique preamble, generated from its own address, and sends it simultaneously with other nodes. A receiver uses our proposed technique to detect the preambles present and simultaneously estimates the channel associated with the detected preamble. Compressive sensing works only if the number of simultaneous preambles present is much smaller than all possible preambles. This is the case with ad hoc networks, where only a handful of nodes come together while the total number of all the possible nodes is extremely large (248, where 48 is the number of bits used for a node s (MAC) address). However, compressive sensing is computationally expensive even though it is very efficient in detecting many preambles in just one slot. We develop many techniques to make compressive sensing practical and study the tradeoffs thereof. Thus, our work allieviates the overhead associated with collecting neighborhood and channel information. It has a lot of potential in making theoretically-optimal solutions practical. We believe that it s appriorate for dynamic and large-scale wireless networks, and is a step towards making such networks efficient. This research effort will allow our Ph.D. students gain domain-specific knowledge in wireless ad hoc networks. It will also allow the PIs to enrich their courses.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- May 05, 2017
- Source ID
- N000141712412
Entities
People
- Kannan Srinivasan
Organizations
- Office of Naval Research
- Ohio State University
- United States Navy