The widely-used protocol in Mobile Ad hoc Networks (MANET) achieves a dynamic, self-organizing and on-demand multi-hop routing by means of the AODV routing protocol. In AODV, wireless links disconnect occasionally because the nodes on the routing path are unreachable, which makes AODV inefficient and unreliable. By constructing a mesh structure, AODV-BR improves AODV routing protocols and provides multiple alternate routes. The algorithm establishes the mesh and multi-path with the RREP of AODV, which does not transmit many control messages. In this paper, a new improved protocol AODV-BRL is proposed to increase the adaptation of routing protocols to topology changes by modifying AODV-BR. In AODV-BRL the alternate routes are created by the Extended Hello Message as well as RREP packets. Then the Extended Routing Table and the least hop count first (LHF) are proposed to determine the optimal alternative route that significantly reduces the distance between the repair node and the destination. Finally, the performance improvement is evaluated by simulations. According to simulation results it is evident that compared with AODV-BR and AODV, AODV-BRL has a higher packet delivery ratio and lower routing overhead.
Recent advances in camera-equipped drone applications increased the demand for visual object detection algorithms with deep learning for aerial images. There are several limitations in accuracy for a single deep learning model. Inspired by ensemble learning can significantly improve the generalization ability of the model in the machine learning field, we introduce a novel integration strategy to combine the inference results of two different methods without non-maximum suppression. In this paper, a global and local ensemble network (GLE-Net) was proposed to increase the quality of predictions by considering the global weights for different models and adjusting the local weights for bounding boxes. Specifically, the global module assigns different weights to models. In the local module, we group the bounding boxes that corresponding to the same object as a cluster. Each cluster generates a final predict box and assigns the highest score in the cluster as the score of the final predict box. Experiments on benchmarks VisDrone2019 show promising performance of GLE-Net compared with the baseline network.
Multipath routing is attractive for load-balancing, fault-tolerance, and security enhancement. However, constructing and maintaining a set of node-disjoint paths between the data source and sink is non-trivial in a dynamic environment. In this paper, we study the problem of route recovery in vertex-disjoint multipath routing for sensor networks with many-to-one traffic patterns. We identify the sufficient conditions for multipaths to be recovered when the existing node-disjoint paths are broken, and provide a simple framework for multipath maintenance. This framework is very efficient in time when multipath source routing is employed. Our findings can help to conserve network resource by not launching any route discovery when the data source realizes that a new route may not exist, to guide mobile data sources to relocate themselves in order to reconstruct the new multipaths, and to help newly-deployed data sources quickly determine whether the required number of multipaths exist for sure or not and then compute them. The technique proposed in this paper is a good complement to the classic max-flow algorithm when node-disjoint multipaths are needed.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.