Objectives: To implement Wireless Sensor Node Authentication and Data Security Framework Using Machine Learning. Methods: The study is centered on a Generative Adversarial Networks (GAN) learning technique which explains about authenticity of secured wireless sensor network. This method is developed using two main components like Generator (G) network which produces false data for confusing the attacker and Discriminator (D) network which consists of multiple layers that can efficiently differentiate between real and fake data for interpretation at the end node (Rx) for providing secure communication. Unauthenticated sensor nodes can be detected with higher accuracy, throughput and low end-to end delay using machine learning based framework. In this work, Probability of False Detection, Reliability and Uniqueness, Delay and throughput are evaluated to determine the performance of presented approach. Findings: Simulation results reveal that the proposed IFAs antennas can provide isolation with a mutual coupling reduction of 18 dB with respect to the transmission coefficients and can also obtain sufficient bandwidth by the proposed antenna array for the 5G mobile terminals applications. Novelty: GANs algorithm is used for the first time to resolve WSN middleware having security issues.