No abstract

No abstract

A buffer overflow at a network node occurs if the sum of the amount of data it generates and the amount of data it receives from upstream nodes exceeds its transmission throughput. Densely connected wireless sensor networks (WSNs) with intermittently available sinks (such as mobile sinks) strongly inherit such a buffer overflow problem as sensor nodes in WSNs have limited buffering capacity, and the intermittent nature of these sinks can further influence the network throughput. We propose distribution and caching of sensor-generated data at sensor nodes that belong to any of the minimum cost paths leading to a mobile sink path. We further propose to store the data at each source node for a certain period of time instead of immediately forwarding them to the downstream nodes so as to reduce unnecessary network congestion. Both mathematical and simulation evaluation show that these schemes significantly reduce packet loss probability at network of any size. Consequently, the amount of information lost in the system due to packet overflow/retransmission will be significantly reduced.

No abstract

<p>A new approach is described for investigating the accuracy of positioning active long-term evolution (LTE) users. The explored approach is a network-based method and depends on signal level measurements as well as the coverage of the serving cell. In a two-dimensional coordinate system, the algorithm simultaneously applies LTE measured data with a combination of a basic prediction model to locate the mobile device’s user. Furthermore, we introduce a unique method that combines timing advance (TA) and the measured signal level to narrow the search region and improve accuracy. The developed method is assessed by comparing the predicted results from the proposed algorithm with satellite measurements from the global positioning system (GPS) in various scenarios calculated via the number of cells that user equipment concurrently reports. This work separates seven different cases starting from a single reported cell to five reported cells from up to 3 sites. For analysis, the root mean square error (RMSE) is computed to obtain the validation for the proposed approach. The study case demonstrates location accuracy based on the numbers of registered cells with the mean RMSE improved using TA to approximately 70-191 m for the range of scenarios.</p>