The Internet of Things (IoT) has become an increasingly popular study area, with billions of devices deployed globally in recent years. These devices can speak with one another without the need for human involvement because they are all connected to the Internet. But this also gives rise to new security problems that are becoming more and more prevalent and important areas for study. A novel INtrusion Detection in Iot using Advanced deep learning NETwork (INDIA-NET) has been presented in this paper to address these shortcomings. It effectively detects intrusions using this innovative Deep Learning (DL) technique. The dataset's duplicate and redundant data are first eliminated by preprocessing using the Minkowski distance-based closest neighbour algorithm. After preprocessing an enhanced transient search optimization algorithm has been used for feature selection. After selecting the features, the features are sent to the novel convolutional neural network combined Generative Adversarial Network (CNN-GAN) which classifies the output into DOS attack, UR3 attack and normal. The accuracy, precision, recall, and detection rate, among other particular metrics, have all been used to calculate the suggested method's performance. Using the NSL-KDD dataset, the suggested INDIA-NET system has been assessed. Its results show improvements in accuracy, precision, recall, and F1-score, with respective values of 99.02%, 99.38%, 98.29%, and 98.83%.