Mobile ad hoc networks have surged in popularity in recent years and are now widely regarded as critically important due to their low complexity and quick expansion. However, a network's elasticity makes it vulnerable to assaults from a wide variety of vectors. The black hole attack is considered the farthest-reaching mutual category of attack within a MANET, though there are many more. Accordingly, this study proposes an intrusion detection and prevention system that utilizes machine learning approaches to identify and prevent black hole attacks in MANET. To establish suitable security regulations and automate defense operations in a large-scale MANET context, this study constructs a context-based node acceptance model based on the Dolphin Echolocation model (NA-DE). By automatically identifying malicious nodes, this approach can achieve quick and early detection of black hole attacks without negatively impacting performance. Additionally, compromised nodes are removed from the network, and traffic is routed through a healthy node if possible. Ad-hoc On-Demand Distance Vector Routing (AODV) is the protocol used for routing. Several metrics are used to gauge the efficacy of the algorithm suggested in the paper. By comparing the performance metrics of the suggested methodology to those of another state-of-the-art routing protocol, it is possible to conclude that the former achieves better results. In the experimental results, when the node count is 250, it is clear that the proposed NA-DE achieved better performance in detecting black hole nodes and allowed the transmission of secure data with less energy consumption.

近年来，移动自组织网络迅速普及，由于其低复杂性和快速扩展，现在被广泛认为至关重要。然而，网络的弹性使其容易受到来自各种媒介的攻击。黑洞攻击被认为是 MANET 中影响最远的相互攻击类别，尽管还有更多攻击类别。因此，本研究提出了一种入侵检测和预防系统，利用机器学习方法来识别和预防 MANET 中的黑洞攻击。为了在大规模 MANET 环境中建立合适的安全规则并自动化防御操作，本研究构建了基于海豚回声定位模型（NA-DE）的基于上下文的节点接受模型。通过自动识别恶意节点，该方法可以实现黑洞攻击的快速、早期检测，而不会对性能产生负面影响。此外，受感染的节点将从网络中删除，并且如果可能，流量将通过健康节点路由。自组织按需距离矢量路由 (AODV) 是用于路由的协议。使用几个指标来衡量论文中建议的算法的有效性。通过将建议方法的性能指标与另一种最先进的路由协议的性能指标进行比较，可以得出结论：前者取得了更好的结果。在实验结果中，当节点数为250时，很明显，所提出的NA-DE在检测黑洞节点方面取得了更好的性能，并且允许以更少的能耗传输安全数据。