论文标题
基于信任的Winnow线性乘法分类，用于MANET中的安全多径路由
Trust-Based Winnow Linear Multiplicative Classification For Secure Multipath Routing In Manet
论文作者
论文摘要
移动临时网络（MANET）中的多路径路由通过避免网络中的攻击节点来对安全数据传输产生重大关注。为了克服此类局限性，提出了基于Winnow Trust的多路径路线发现（WT-MRD）机制。它以更高的安全性和更少的时间构建了从源到目标的多个路径。最初，根据节点合作计数，数据包转发率和数据包下降速率（通过基于邻居节点的信任计算（NN-TC）模型）计算每个节点的信任值。计算信任值后，通过使用Winnow线性乘法分类（WLMC）算法将节点分为正常或恶意。在普通节点的帮助下，WT-MRD机制通过发送两个控制消息RREQ和RREP来从源到目的地找到多路径。源节点将路由请求RREQ传输到相邻节点，以构建多个路由路径。接收RREQ消息后，相邻节点将维护源信息和下一个HOP信息的路由表。然后路由回复（RREP）消息将从相邻节点发送到源节点。通过这种方式，从源到目标的多个路径路径以更高的安全级别构建。关键字：合作计数，数据包删除率，数据包转率，多路径路由，安全性，信任值，Winnow线性乘法分类
Multipath routing in Mobile Ad Hoc Network (MANET) plays a significant concern for secured data transmission by avoiding the attack nodes in the network. In order to overcome such limitations, a Winnow Trust based Multipath Route Discovery (WT-MRD) Mechanism is proposed. It constructs multiple paths from source to destination with higher security and lesser time. Initially, the trust value of each node is calculated based on the node cooperative count, data packet forwarding rate and packet drop rate by Neighbor Node-based Trust Calculation (NN-TC) Model. After calculating the trust value, the nodes are classified as normal or malicious by using Winnow Linear Multiplicative Classification (WLMC) Algorithm. With the help of normal nodes, the WT-MRD Mechanism finds multipath from source to destination by sending a two control message RREQ and RREP. Source node transmits a route request RREQ to the neighboring node for constructing the multiple route paths. After receiving the RREQ message, the neighboring node maintains the route table where the source information and next hop information are present. Then Route Reply (RREP) messages are sent from neighboring node to source node. By this way, multiple route paths from source to destination are constructed with a higher security level. Keywords: Cooperative Count, Data Packet Dropped Rate, Data Packet Forwarded Rate, Multipath Routing, Security, Trust Value, Winnow Linear Multiplicative Classification
