High Impact Factor : 4.396 icon | Submit Manuscript Online icon |

AURA Software Architecture and System One Decision Integration for an Edge Artificial Intelligence Departmental Assistant Robot

Author(s):

Siddhant Ramesh Dube , Trinity College of Engineering and Research Pune; Abhishek Devanand Jadav, Trinity College of Engineering and Research Pune; Rusha Gajanan Panzade, Trinity College of Engineering and Research Pune; Aditya Mahadev Jaybhay, Trinity College of Engineering and Research Pune; Prof. Dhiraj Tikar, Trinity College of Engineering and Research Pune

Keywords:

AURA, Edge-AI Robotics, ROS 2 Jazzy, Nav2, Retrieval-Augmented Generation, System One Decision Model, Jev by TypeSafe, Digital Twin

Abstract

University academic offices struggle with distributed, disconnected informational channels for scheduling, course curricula, instructor availability, circulars, and course attendance. Mobile robotics presents an integrated interactive medium, yet industrial mobile bases and anthropomorphic platforms remain excessively intricate and costly for localized departmental use. This paper conducts a software architectural review of AURA (Academic Utility Robotic Assistant), an edge-AI assistant configured for voice interaction and currently undergoing simulation-stage evaluation. The work reports no synthetic or unverified experimental metrics; instead, it outlines the system software infrastructure—incorporating ROS 2 Jazzy, Nav2, Gazebo Harmonic simulation, FastAPI service management, ChromaDB-indexed retrieval-augmented generation (RAG), dual-factor BLE/TOTP verification, and role-based security policies—alongside the introduction of TypeSafe's Jev System One decision model. Delegating input validation, task classification, and threshold-gated escalation away from generative large language models (LLMs) toward typed Choice, Score, and Noul questions reduces parsing errors, bounds query latency, and keeps every AI component outside the real-time ESP32 wheel control loop. A measurement plan and a set of unmeasured V1 targets are fixed in advance so that later prototype results can be judged against clear criteria.

Other Details

Paper ID: IJSRDV14I90077
Published in: Volume : 14, Issue : 9
Publication Date: 01/12/2026
Page(s): 170-175

Article Preview

Download Article