MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202641060208 A) filed by S. Hemalatha; R. Mogith; R. V. Sarvesh; Tarun Velumani; U. S. Vipul; and R. Vishnu on May 12, 2026, for Privacy Controlled Low Latency Edge-Ai Conversational Assistant.
Inventors include S. Hemalatha; R. Mogith; R. V. Sarvesh; Tarun Velumani; U. S. Vipul; and R. Vishnu.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: Voice-enabled conversational assistants are increasingly used in smart homes, workplaces, and personal environments due to their ability to provide hands-free interaction. However, conventional voice assistants typically operate in an always- on listening mode, creating serious privacy concerns and increased dependency on continuous monitoring. This project proposes a latency-optimized Edge-AI voice assistant architecture based on the ESP32-S3 microcontroller, incorporating a hardware-level Push-to-Talk trigger to ensure complete user privacy. The system integrates an I2S digital microphone for noise-resistant audio capture and utilizes a low- latency API bridge to connect with Generative Al platforms such as Gemini or OpenAI for multilingual response generation. An SSD1306 OLED display provides animated eye expressions as non-verbal feedback synchronized with assistant states such as listening, processing, and responding. By combining hardware-enforced privacy control, efficient edge processing, and cloud-based generative intelligence, the proposed design delivers a compact, user-friendly, and privacy-preserving conversational assistant suitable for home and professional applications.
Disclaimer: Curated by HT Syndication.
