Guardian

Inspiration

Resources like water, electricity are used everywhere, A simple system can aim to reduce usage instead of warning the user rather it turns it off by itself. Thus, eliminating human laziness

What it does

It uses valves and sensors combined to read data and shut off the specific valve if any unwanted usage is detected, It DOES NOT have hard-coded value for when to shut-off, rather it learns for the initial few weeks [using ML forest] which then combined with a PIR detector sensor to account for presence of human. It AUTOMATICALLY shuts the valves off when it detects anomally [the chances of errors reduce due to ML forests unsupervised learning] aswell as the device aims to be private for each user

How we built it

Prototype using replit Raspberry pi 5 [8gb ram] Sensors for Water detection, Electricity main circuit with external smart switches and (if possible) internal smart switches, lpg gas detector and PIR sensor Valves for Each tap, main water circuit, lpg gas valves [Apex certified]

Challenges we ran into

Simulation challenges[WE DID NOT TEST IT IRL YET] Error rates are low but not zero and due to the web dashboard displaying data - user raise question whether the control is truly private and what if anyone breaches it.

Accomplishments that we're proud of

Using ML forest with unsupervised training the model works predictably, mostly the web ui is the part which we like the most- usage of 3d model in web ui

What we learned

Questions bout privacy, general law against fixtures of valves in an infrastructure, web development, critical thinking and quick and smart answering of difficult questions.