A Multi-Agent Framework for Smart Door Automation Using Intelligent Access Control and SLMs
- Author Sudip Chakraborty
- Co-Author Deep Chakraborty
- DOI https://ww
- Country : India
- Subject : Computer Science and Information Sciences
Keywords : Smart Door, Multi-Agent Framework, SLM for Home Automation, AI in Security, Voice Interface, Home Automation, Intelligent Access Control.
Abstract
Recent advancements in home automation have highlighted the growing importance of secure and intelligent door management systems, where maintaining user convenience and data privacy is paramount. Building on prior research into smart door controls and automated security solutions, this paper introduces a novel multi-agent framework that leverages Small Language Models (SLMs) for real-time decision-making and user interaction. Unlike larger, more resource-intensive language models, SLMs offer lower latency and reduced computational overhead, making them better suited for on-device operation. To distribute tasks and optimize performance, the system comprises five specialized agents—Central Door Management, Response Management, Door Control, Identity Recognition, and Audio Response. The Identity Recognition Agent leverages advanced vision-based techniques (such as YOLO) to classify visitors (e.g., family, friends, or unknown), while the Response Management and Audio Response Agents employ SLMs to deliver contextually relevant, voice-based instructions. By offloading targeted language processing tasks to smaller, domain-focused models, the system effectively mitigates network dependencies, improving security and privacy. Preliminary evaluations indicate that this architecture not only maintains robust access control and personalized interaction but also exhibits strong scalability for diverse household scenarios. Overall, our findings suggest that a multi-agent design, powered by SLMs, represents a promising direction for next-generation smart door automation and security solutions.
Comments
No have any comment !