An Agentic AI-Driven End-to-End IoT Engineering Platform: A Proposed Workflow from Natural Language to IoT Prototype

Author

Satish Neval, Mandar Patil, Omkar Shinde, Prof. Hrushikesh Ingole


Abstract

 Developing an Internet of Things (IoT) prototype requires coordinated decisions about requirements, hardware, circuit connections, embedded firmware, and validation. These tasks can be challenging for beginners because an error in one stage may prevent the complete system from working. This paper presents the proposed design of an Agentic AI-Driven End-to-End IoT Engineering Platform that aims to translate natural-language descriptions into structured IoT prototype plans. The proposed workflow includes requirement analysis, ESP32-oriented component and pin planning, context-aware firmware generation, validation support, user approval, and deployment guidance. The current Circuitra user-interface prototype shows workspace areas for projects, components, code, simulation, AI, hardware, and console. However, the supplied screenshot shows an empty circuit canvas, no selected board, and an idle simulation. The screenshot is therefore treated as evidence of the frontend layout only; backend integration, generated circuits, firmware execution, and experimental performance have not yet been verified. This paper documents the problem, proposed architecture, workflow, and planned evaluation rather than claiming completed implementation results. 



Keywords

Agentic AI; Internet of Things; ESP32; Natural Language Processing; Firmware Generation; IoT Engineering



Full Text:

Download Paper PDF


References


  1. Peng, Y. Gao, et al., “Programming Embedded IoT Applications in Natural Language with IoTPilot,” in Proc. 23rd Annual International Conference on Mobile Systems, Applications and Services (MobiSys ’25), pp. 70–82, 2025, doi: 10.1145/3711875.3729136.
  2. Yang, M. Li, M. Han, Z. Li, and W. Xu, “AutoEmbed: LLM-driven Automated Software Development for Generic Embedded IoT Systems,” in Proc. ACM/IEEE International Conference on Embedded Artificial Intelligence and Sensing Systems (SenSys ’26), pp. 362–376, 2026, doi: 10.1145/3774906.3800466.
  3. Luo, R. Ma, X. Zhang, and L. Qiu, “IoTGen: Towards LLM-Driven IoT Hardware PCB Generation,” in Proc. 24th Annual International Conference on Mobile Systems, Applications and Services (MobiSys ’26), 2026, doi: 10.1145/3745756.3809239.
  4. Li, Y. Cheng, M. Ma, Y. Zou, N. Yang, W. Cheng, H. Li, Y. Chen, and T. Chen, “Skilled AI Agents for Embedded and IoT Systems Development,” in Proc. ACM Conference on AI and Agentic Systems (CAIS ’26), pp. 1248–1252, 2026, doi: 10.1145/3786335.3813205.
  5. Paria, A. Dasgupta, and S. Bhunia, “SPELL: An End-to-End Tool Flow for LLM-Guided Secure SoC Design for Embedded Systems,” IEEE Embedded Systems Letters, vol. 16, no. 4, pp. 365–368, 2024, doi: 10.1109/LES.2024.3447691.
  6. Wang, J. Chen, Z. Wei, L. Lin, and G. Tong, “SafeCodeRL: Security-Constrained Multi-Agent Reinforcement Learning for Trustworthy LLM-Generated IoT/CPS Software,” Sensors, vol. 26, no. 11, art. 3502, 2026, doi: 10.3390/s26113502.

Share your valuable work from Social Media Buttons