Safe and Efficient Operation of Mine Vehicles in Fog and Low-Visibility Conditions in Open Cast Iron Ore Mines.
Ministry of Steel · Smart Automation · Hardware
The openness and the thin field make this genuinely winnable for a team with radar hardware and mining contacts, but if you cannot demonstrate detection in real scattering media you are pitching a concept, so be certain you can before committing.
What it actually is
In NMDC's Bailadila iron ore mines the monsoon brings fog so thick that drivers of giant dump trucks can see only three to five metres ahead, so hauling either slows to a crawl or stops. That costs production and still leaves a collision risk. The ask is a system that lets those trucks keep moving safely when the driver effectively cannot see.
What to build
A driver-assistance and collision-avoidance stack for haul trucks: a fog-penetrating perception layer combining thermal imaging with radar or LiDAR returns to detect vehicles, people and haul road edges beyond the visible range, a fused positional picture from DGPS plus vehicle-to-vehicle position broadcast so trucks know where each other are without seeing each other, an in-cab display rendering the road edge, proximity ring and closing-speed warnings for the operator, and a control-room view showing live fleet positions on the haul road network with automatic alerts for proximity breaches and unsafe corridors during low-visibility events.
Smallest thing that wins the room
Fill an enclosed run with dense theatrical fog, drive a scaled vehicle through it, and show the in-cab display tracking an obstacle the mounted camera cannot see at all.
How crowded this one gets
A guess, projected from the 2025 statements — the last year where both the submission counts and the winners were published.
Quieter than 94% of the 226 · #15 of 226 by expected field
Few teams are likely to go here. The best odds on the board come from statements like this.
Why: central ministry statements sat below the average; hardware halves the field a software statement gets.
This is a guess, not a fact
Nobody has published 2026’s numbers yet. This is an analysed estimate from last year’s pattern, so please do not take it as the truth — check the live counter on the SIH portal before you decide anything. The range covers the middle half of likely outcomes, so one statement in two lands outside it. Entry closes at 500 ideas per statement, so no range goes past that — a statement that reaches the cap fills and shuts rather than drawing an unlimited crowd. The model reads only three things a team can see before choosing — software or hardware, the theme, and what kind of body posted it — and those explain about a quarter of the variation in last year’s field sizes (R² 0.25 on held-out statements). Trust the band more than the number, and the ordering more than either. It cannot see how good your idea is, which is the part that actually decides it.
The scores
The number is the shorthand. The line under it is the reason.
Acceptance potential
2/5The need is real and the openness is attractive, but the sensors are out of reach, the conditions cannot be reproduced and there is no fog haul-road data to train on, which means most submissions will be a Peltier-cooled toy and a slide deck of what the real system would do.
Feasibility
2/5The capability genuinely depends on sensors most teams cannot buy — automotive LiDAR and usable-resolution thermal cameras run into lakhs, mining radar more still — and there is no public dataset of fogged haul-road scenes, nor any way for a student team to get onto an active Bailadila haul road to collect one.
Innovation scope
5/5The description lists nine candidate technologies with 'may leverage' and specifies outcomes rather than a solution, so the entire architecture — sensing modality, autonomy level, whether it assists or intervenes — is left completely open to you.
Clarity
3/5The problem context is excellent and quantified down to 3–5 metre visibility and 37 MTPA of affected production, but the deliverable itself is never defined — a menu of technologies and a list of desired outcomes is not a specification, and two teams could build entirely unrelated things and both be responsive.
Effort
HeavySensor integration and calibration, perception under degraded visibility, sensor fusion, a V2V link, an in-cab HMI and a control room platform is a large systems-engineering effort, and most of the schedule goes to making unfamiliar hardware talk to each other.
Demo-ability
HardThe defining condition is dense fog on a mine haul road with 100-tonne dumpers, and no fog machine in a corridor with a scaled vehicle reproduces the scattering behaviour, the distances or the stakes — the demo cannot prove the claim it is making.
In its favour
- Green flag: Innovation scope is genuinely maximal — the statement asks for outcomes, not a design, so an unconventional approach cannot be marked down as non-compliant
- Green flag: NMDC has quantified the loss precisely, which means the business case writes itself and you never have to argue that the problem matters
- Green flag: Mining hardware is unfamiliar territory for most teams, so the number of serious entries will be small and a competent submission stands out
- Green flag: mmWave radar development kits are the one fog-penetrating sensor available at a few thousand rupees, which gives a resourceful team a genuine path to a working perception demo
Against it
- Red flag: You cannot reproduce dense fog at haul-road scale, so whatever you show on stage will be a scaled analogue and judges will discount the claim accordingly
- Red flag: Automotive LiDAR and decent thermal cameras cost more than most college project budgets in total, and the cheap thermal modules have resolution too low for obstacle classification at range
- Red flag: No public dataset of fogged mine haul roads exists, so any perception model is trained on synthetic fog applied to clear-weather footage, which is exactly the circular setup that collapses under questioning
- Red flag: Nine candidate technologies including digital twin and full autonomy invites teams to promise a system architecture and deliver one sensor on a breadboard
What you will be writing
- FLIR / Lepton thermal imaging
- mmWave radar (TI IWR6843) obstacle detection
- LiDAR point cloud clustering
- sensor fusion via Extended Kalman Filter
- DSRC / C-V2X vehicle-to-vehicle position broadcast
- RTK-DGPS positioning
- Mining operations safety
- Autonomous and assisted driving
- Industrial IoT
Prior art to read before you start
low-visibility obstacle detection · haul truck collision avoidance · thermal and radar sensor fusion
Analysed by Claude Opus. Every score above is a judgment call with its reasoning attached — kindly cross-check this against the official statement on the SIH portal before your team commits to it.