MUMBAI, India, March 13 -- Intellectual Property India has published a patent application (202641025051 A) filed by Vellore Institute Of Technology, Vellore, Tamil Nadu, on March 3, for 'a smart road safety system for autonomous or driver-assist vehicles configured for obstacle detection and collision risk prediction.'

Inventor(s) include Parvathi R; and Tharun M.

The application for the patent was published on March 13, under issue no. 11/2026.

According to the abstract released by the Intellectual Property India: "The present invention relates to a smart road safety system for autonomous or driver-assist vehicles that enables real-time obstacle detection and collision risk prediction using a single forward-facing monocular camera. The system captures continuous road scene images and generates multi-modal pseudo-sensor data including deep-learning-based object detection with confidence scoring, 10 monocular depth estimation producing dense distance maps, and optical-flow- based motion vectors representing relative object velocity. An orientation-aware collision risk assessment engine determines obstacle motion direction with respect to the ego vehicle, classifies obstacles as converging, crossing, or diverging, and selectively computes time-to-collision for threatening objects while suppressing 15 non-critical risks. An adaptive sensor fusion module dynamically assigns weights to perception cues based on temporal stability and reliability, producing a fused collision risk value. An uncertainty-aware forecasting module evaluates persistence of risk across frames while accounting for noise, occlusions, lighting variations, and depth or motion fluctuations to reduce false alerts. An edge- 20 adaptive execution controller adjusts inference frequency and computational load depending on predicted risk levels. Safety alerts are persistently generated only when collision risk exceeds a threshold for a sustained duration, thereby enabling low-latency, camera-only collision prediction with improved decision reliability under dynamic environmental conditions."

Disclaimer: Curated by HT Syndication.