The Internet of Things (IoT) is a vast network of diverse devices that has expanded into various aspects of our daily lives. However, its growth comes with increased security challenges, particularly in developing efficient identity management systems. General-purpose identity management solutions are impractical due to IoT limitations related to centralized architecture, resource constraints, and impersonation susceptibility. These limitations raise the need for an IoT-specific identity management scheme. Research has shown excellent efforts employing Distributed Ledger Technologies (DLT), Fog Computing, and Machine Learning techniques. Blockchain, as a type of DLT, has taken a vast interest in current IoT security research. However, while it mitigates the centralization issue in IoT, it has multiple limitations. As an alternative, IOTA is proposed as an IoT-specific DLT to overcome these challenges. In this work, we propose a conceptual framework for an identity management scheme that utilizes IOTA technology for the secure, distributed storage of IoT device identities. Alongside IOTA, we integrate Fog Computing to mitigate the resource limitations in IoT devices and employ Machine Learning to provide continuous authentication on connected IoT devices. Through this scheme, we aim to achieve decentralization, scalability, efficiency, and security of the IoT.