MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202621074367 A) filed by Mayur Kakasaheb Jadhav on June 16, 2026, for A Hardware-Implemented Explainable Ai System For Automated Control Of Studio Equipment.
Inventors include Dr. Shilpa S. Laddha; Mayur Kakasaheb Jadhav; Ms. Priyanka S. Sonwane; Mr. Tushar K. Jadhav; and Ms. Rohini S. Gore.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: This work is related to a specialized hardware optimized system (101) specifically designed for newsroom studio equipment controlling, video recording and telecast automation. The system comprises a set of tightly integrated subsystem modules, including a secure feed ingestion unit (201) that performs elliptic curve Diffie– Hellman key exchange (202) algorithm and HMAC-SHA256 verification (203) directly on chip, a graph processing accelerator (301) with an explainability engine (302) that assigns Explanation Identifiers (307) and manages provenance scoring, and a real-time interrupt controller (303) with DMA channels (311) that provides predictable response to sensor inputs (310). The specific actuator driver interface encodes the EIDs into the instruction/control packets for devices such as the Pan-Tilt- Zoom enabled video camera devices (401), the studio lighting rigs (402) and the text-based teleprompters (403). This is predominantly carried out by NDI logic (404), DMX512 frame generation (405) and UART framing (406). In parallel, the trace export subsystem (501) builds correct audit records with specific integrity tests (505). Export is controlled by the system via the controllers (506) also to flash storage or USB storage (507). This system’s operations are power and thermal stable. The Power and Thermal Management subsystem (601) includes DVFS controllers (602), Sleep/wake gates (603), UPS-triggered checkpoints (604), graph state preservation (605), and explainability buffers (308). The system tags every command with an EID so actions can be traced. It works fast with latency under 100 ms, tested at 40–60 ms for heavy loads. The design is well scalable. By unifying siloed modules into one compact platform, the invention reduces external dependencies and guarantees uninterrupted automated operation in live newsroom recording environments with explainable artificial intelligence.
Disclaimer: Curated by HT Syndication.
