It is crucial to create robust strategies for protecting Internet of Things gadgets. There is a risk that cyberattacks could compromise IoT devices. This justifies the need for their safety. They will be safe from harm if they have a solid identity. For instance, the gadget should only talk to the systems it manages, and its information should be kept in a safe place. This information needs to be encrypted the minute it is transmitted and stored. To prevent malicious actors from gaining access to private data, the device must be encrypted by an external entity. None of these things have been standardised yet. In addition, the use of more lightweight cryptographic systems is required because big data sources, such as sensors placed in IoT applications, produce enormous data. Diffie Hellman became an effective method for secure and lightweight data exchange for linked devices in IoT settings after it was combined with Elliptic Curve (EC-DH). Even with these enhancements, there are still security holes in the Internet of Things. We propose a Hybrid Technique that improves security in IoT applications as a solution to this issue. We analysed a real-world example of anomaly detection in the medical field. Protocols like MQTT and XMPP are set up with the Adafruit IO cloud and the Watson IoT platform for the experiments. Several experiments validated the scheme's efficacy, demonstrating its ability to increase security while simultaneously improving key characteristics associated with the Internet of Things (uploading, downloading, encrypting, and decrypting). The proposed approach is useful for real-time IoT applications.

Internet of things (IoT), IoT Security, Hybrid Approach. AI, Machine Learning