Dimensioning and worst-case analysis of cluster-tree sensor networks
ACM Transactions on Sensor Networks (TOSN)
Time-Critical data delivery in wireless sensor networks
DCOSS'10 Proceedings of the 6th IEEE international conference on Distributed Computing in Sensor Systems
Application-level operations latency control in networked WSAN
ADHOC-NOW'12 Proceedings of the 11th international conference on Ad-hoc, Mobile, and Wireless Networks
| Hi-index | 0.00 |
This dissertation studies real-time application support in wireless ad-hoc and sensor networks. Real-time applications are performance critical applications that require bounded service latency. In multi-hop wireless ad-hoc and sensor networks, communication delays are dominant over processing delays. Therefore, to enable real-time applications in such networks, the communication latency must be bounded. The shared nature of the communication medium makes the delay characteristics of the medium access control (MAC) protocol in use very important. Furthermore, it is desirable that the MAC.protocols for such networks be distributed and be able to spatially reuse the communication channel for scalability and efficiency. In this dissertation, we derive expressions of real-time capacity that characterize the ability of a network to deliver data on time as well as develop network protocols that achieve this capacity. We introduce a hexagonal network topology based architecture for wireless ad-hoc and sensor networks for real-time applications. We present addressing and constant time routing protocols for the hexagonal network. We develop two distributed spatial-reuse time domain multiplexed MAC protocols with provable real-time properties for convergecast traffic. One protocol constructs network-wide transmission schedule and gives equal bandwidth to all the nodes. This protocol has zero scheduling message overhead and is optimal in the sense that the base-station does not idle. The other protocol supports a more general (unequal bandwidth to nodes) mixture of real and non-real time traffic. Clock synchronization is achieved implicitly. Real-time capacity expressions are obtained and analyzed for the earliest deadline first, deadline monotonic and these two scheduling algorithms. Data gathering (many-to-one) and data dissemination (one-to-many) are two of the three canonical traffic patterns in WSN. Unlike data gathering at base-stations, transmission scheduling for data dissemination is an easy problem and hence is not as much an issue as is the minimization of number of transmissions for interference suppression and energy conservation. We present and prove the properties of an optimal multicast tree in WSN, which is cast as a generalized Steiner Tree Problem. We present a distributed heuristic that constructs and maintains near-optimal multicast tree on-line.