MUMBAI, India, Oct. 5 -- Intellectual Property India has published a patent application (202621098962 A) filed by Dr. D. Y. Patil Institute Of Technology, Pimpri, Pune -; and Dnyaan Prasad Global University, School Of Technology And Research, Pimpri, Pune - on August 16, 2026, for Agentic Ai-Driven Multilingual Agricultural Advisory System With Multi-Tier Offline Functionality.
Inventors include Dr. Shubhangi Vairagar; Aditya Khalkar; Omkar Kokate; Sarang Kokane; and Ayush Kandekar.
The application for the patent was published on October 02, 2026, under issue no. 40/2026.
Abstract: The present invention relates to an intelligent agricultural advisory system designed for rural Indian farmers, addressing critical barriers in agricultural technology adoption including language accessibility, internet dependency, and lack of agentic intelligence. The invention uses a ReAct-pattern tool orchestration framework for agentic capability with any LLM API via XML instruction decoding and parallel tool execution. It has an inference fallback hierarchy of agentic cloud AI with tool calling, then cloud LLM, then device agnostic LLM inference using llama.cpp, followed by device agnostic RAG retrieval from 3,073 pre-computed embeddings of Indian agriculture crops. This invention uses a voice pipeline which eliminates conversation lag by streamlining TTS delivery of speech by executing TTS generation in parallel with text generation. This invention includes vision-based detection of plant disease, real-time weather forecast with farming advice generated by AI, GPS- based soil composition, and language capabilities in Hindi, Marathi, and English. This invention uses a device agnostic RAG component based on BERT WordPiece tokenization, ONNX INT8-quantized all-MiniLM-L6-v2 embeddings, and dot-product similarity search for sub-5ms retrieval irrespective of internet availability. This invention works on off-the-shelf ARM64 Android devices without internet and gracefully escalates capabilities with cloud computing upon network availability.
Disclaimer: Curated by HT Syndication.
