MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641121797 A) filed by Vel Tech Multi Tech Dr. Rangarajan Dr. Sakunthala Engineering College; Dr. T. Samraj Lawrence; R. Dharani Sri; R. Harini; L. P. Persiyal; and M. Subhashree on October 12, 2026, for Aiot And Tinyml Hybrid System For Cloudburst Prediction.
Inventors include Dr. T. Samraj Lawrence; R. Dharani Sri; R. Harini; L. P. Persiyal; and M. Subhashree.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: Cloudburst is an unpredictable and severe weather phenomenon that commonly occurs in mountainous or flood-prone regions. It has been defined as an exceptionally high amount of rainfall in a short time that results in flash floods, landslides, debris flows, loss of lives, and damage to economic activities. Traditional weather forecasting systems are often ill-suited to pick up on such localized events as they often take a large scale approach and operate on a set of general atmospheric parameters. Cloudburst conditions often entail rapid local variations in environmental parameters such as relative humidity, atmospheric pressure, temperature, rainfall, and light intensity. These sudden variations in environmental conditions can occur in a matter of hours or less, limiting the amount of time that authorities can issue warnings before the event occurs. A local monitoring system that takes local measurements on such conditions can provide an additional layer of security. An intelligent IoT-based early warning system is presented for detecting such cloudburst-associated local environmental conditions. Multiple environmental sensors are used to monitor and record continuous atmospheric pressure, relative humidity, temperature, rainfall, and light intensity readings in the region. The sensor output data is sent to a processing unit based on ESP32 to perform real-time preprocessing and analysis. These sensor readings can be used to indicate the likelihood of cloudburst occurrence based on temporal variations of one or more of these variables. The ESP32 also houses a TinyML-based pattern classification model that has been trained to pick up on local environmental conditions that indicate a cloudburst. The trained model is then used to perform inference on the sensor data to determine if the current environmental conditions warrant any alerts. The system also comprises a display unit, indicator lights, and a GSM module for sending out SMS alerts. Users can be alerted if any high-risk conditions are found. Multiple rounds of monitoring and alerting can be performed depending on the presence of such conditions. The proposed system presents a cost-effective and local means of monitoring environmental conditions and providing early warnings about impending cloudburst events in a specific region to help mitigate the damage they can cause. This system is particularly useful in remote or high risk areas that may not have sufficient data about localized weather conditions that can be provided by large-scale traditional systems.
Disclaimer: Curated by HT Syndication.
