Build a lightweight keyword spotting model that runs locally on a low-power microcontroller with under 256KB RAM and streams audio to a remote server upon detection.
Indian Space Research Organisation(ISRO) · Department of Space / Indian Space Research Organisation · Hardware
Signing in saves it for your whole team — everyone on your invite link sees the same two entries. Anything you shortlisted while signed out comes with you.
Decomposed from what the description asks for. Nothing added.
Local KWS Model
An ultra-lightweight keyword spotting model built using open-source TinyML frameworks to run on low-power devices.
Audio Streaming Pipeline
A module that instantly streams subsequent audio to a remote ASR server once the custom keyword is detected.
Resource Monitor
Ensures the edge application operates within the limits of under 256KB RAM and less than 10% CPU usage while idling.
Custom Keyword Trainer
Training pipeline utilizing custom keywords without relying on pre-trained generic smart assistant models.
Both columns are read off the brief's own wording. Nothing here is inferred from the ministry's name.
Evaluators will check model size, RAM and Flash footprint, CPU usage during idle listening, true-positive rate with near-zero false activations, and latency between the keyword ending and the cloud receiving the audio stream.
A jury can still ask about these. Decide them deliberately rather than by accident.
Generated from the brief's own wording and the competition's published rules — never from a guess about what this ministry prefers.
The specific custom keyword to be used?
The brief never answers this, so a panel will. Whatever you decide, say it the same way twice.
The exact target hardware microcontroller model for final testing?
The brief never answers this, so a panel will. Whatever you decide, say it the same way twice.
Has any part of this been shown at a previous event, hackathon or college project?
The guidelines are explicit: your solution must not have appeared in any previous event or programme, of any sort. A recycled project is what a team under time pressure reaches for.
The brief asks for local kws model. How would you build that?
Decoded from SIH26172 itself — An ultra-lightweight keyword spotting model built using open-source TinyML frameworks to run on low-power devices. The brief asks for it by name.
Each one is quoted from a gap in the brief, not a guess about your team.
Needs physical hardware
Listed as a hardware problem statement, so a working demo needs physical components you have to source yourself.
No measurable target
While it demands a high true-positive rate and near-zero false activations, it never states exact numerical thresholds for these accuracy metrics.
What the organisers attached, and what the brief assumes you can get.
Same organisation, same year. Reading two of theirs tells you more about what they care about than reading one.
Pick what you are about to do and copy the prompt. It carries the organisers' own wording, the constraints they never spell out, and an instruction not to invent requirements they never set.
Who has this problem, what already exists, and what you would have to find out.
The brief asks for audio streaming pipeline. How would you build that?
Decoded from SIH26172 itself — A module that instantly streams subsequent audio to a remote ASR server once the custom keyword is detected. The brief asks for it by name.
The brief asks for resource monitor. How would you build that?
Decoded from SIH26172 itself — Ensures the edge application operates within the limits of under 256KB RAM and less than 10% CPU usage while idling. The brief asks for it by name.
The brief asks for custom keyword trainer. How would you build that?
Decoded from SIH26172 itself — Training pipeline utilizing custom keywords without relying on pre-trained generic smart assistant models. The brief asks for it by name.
Where does your data come from — a published source, one you collect, or one you generate?
No dataset is attached to this problem statement, so sourcing it is part of the work and nobody told you that.
Why not use what already exists? Name the closest thing to this that is already running.
A team that has not named the alternative themselves is answering this for the first time in the room.
Which single thing will you demonstrate end to end, start to finish, with nothing skipped?
Ours, not a rule: a narrow thing that fully works survives questioning better than a broad thing that half works. If nobody on the team can name it, that is the finding.
Show me this working: evaluators will check model size, RAM and Flash footprint, CPU usage during idle listening, true-positive rate with near-zero false activations, and latency between the keyword ending and the cloud receiving the audio stream.
This is the evaluator read for your problem statement, decoded from the brief's own wording.