Independent road audits of how bus and truck ADAS behaves in Indian traffic
TL;DR Assistaudit fits a small logger and camera to buses and trucks that have driver-assistance systems, records every warning and automatic brake for a few weeks and reports which ones were right, wrong or missed. Fleets and insurers pay per audit, and OEMs buy the pooled benchmark.
India is about to put driver-assistance systems on a big share of its heavy vehicles. A MoRTH draft from April 2025 proposed drowsiness warning (AIS-184) and lane departure warning (AIS-188) on buses and trucks from April 2026 for new models, plus emergency braking (AIS-162)[1]. Check the final Gazette dates though, as reports have differed.
Here's the catch - nobody knows how these systems behave on Indian roads. Most effect sizes come from US data, where truck AEB cut rear-end crash rates by 41%[2]. Even in the US, front crash prevention worked less well when the vehicle hit was a motorcycle, 41% against 53% for a car[3]. On Indian roads, full of two-wheelers, animals and faded lane paint, false alarms could make drivers ignore warnings or switch the systems off.
What it does, in regulatory terms: measurement and reporting only. Assistaudit does not certify vehicles or grant type approval, which stays with government-notified test agencies such as ICAT[4]. Its reports are independent evidence for buyers, not a compliance stamp.
A fleet picks 10–30 vehicles. Assistaudit fits a forward camera and a logger that reads the vehicle's ADAS signals, and records four to eight weeks of normal duty. Software flags every warning, every brake intervention and every near-miss that came without a warning. Trained reviewers then label each one as a correct alert, a false alert or a missed event.
The fleet gets a plain report covering false alarms per 1,000 km, what triggers them and where drivers override or disable the system. Pooled across fleets with consent, the labels become a benchmark that OEMs and suppliers can buy to tune their systems for India.
I've reframed the seed, which pitched independent safety checks for self-driving cars. India has no driverless cars on public roads and the transport minister has said he won't allow them[5]. Mandated ADAS on heavy vehicles is the real, near-term version of the same question. Does the automation work here?
Big private bus and truck fleets pay per audit, because they're the ones who eat the crashes and driver complaints. Motor insurers covering those fleets are a second buyer if the data helps them price risk. OEMs and Tier-1 suppliers pay for pooled benchmark data.
There's no public price for this service, so a pilot will have to set it.
The draft ADAS rule for buses and trucks dates from April 2025, with start dates in 2026[1]. Once the systems are fitted, fleets will be living with them every day. Camera and edge-AI hardware is cheap enough now too, and one fleet video company says they've analysed 30B+ driving miles[6].
Nobody publishes a figure for ADAS verification spending in India. The narrowest signal is the mandate itself - four ADAS functions proposed for M2, M3, N2 and N3 vehicles[1], which covers every bus and medium or heavy truck.
Fleet video safety is an established neighbouring market, as that company's mileage claim shows[6]. That's context, not your market.
ICAT and similar agencies test ADAS on proving grounds for type approval[4], but that happens before sale and on test tracks. Fleet video firms sell cameras that coach drivers and score behaviour[6], which isn't the same as checking the vehicle's own ADAS. OEMs collect field data on their own systems but rarely publish it.
I found no independent group in India auditing how fitted ADAS performs once vehicles are in service.
| Difficulty | Hard (52/100) |
| Starting capital | ₹90 L–1.8 Cr |
| Time to launch | 6–9 months |
Logging is the easy part, since cameras and data loggers are cheap and proven. The hard parts are reading ADAS signals that each OEM exposes differently, labelling events consistently enough to defend them and getting OEMs to see an outsider's report as fair and not hostile.
No licence is needed to measure and report, but you must never present the report as certification.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't let the model have the final word on whether an event was a missed hazard. Those labels end up in disputes with OEMs, so a trained reviewer has to own each one.
Want to work out the AI layer for Assistaudit? Talk to Yash →
Here's what can go south -
I like this for a founder with vehicle-testing or automotive-electronics experience and the patience to sell to fleets and OEMs. If you're a pure software team then it's the wrong fit.
Prove two things first - that you can read and label ADAS events on two different OEMs' vehicles, and that one fleet pays for the report.
Plan for roughly ₹90 L–1.8 Cr, mostly on loggers, cameras, the review team and the software to read different OEMs' signals.
No licence to measure and report, but you can't certify vehicles. That stays with notified agencies like ICAT. You'll need DPDP compliance for video and the fleet's consent to tap vehicle data.
Yes, anywhere ADAS is becoming mandatory on trucks and buses. Type approval and privacy rules differ, so check who certifies vehicles locally.
The free Assistaudit validation kit has the line-by-line capital breakdown, the effort scoring, the ADAS-rule and data-protection checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for fleets, insurers and OEMs, and 5 assumptions to test. Download the kit →
Thinking about building Assistaudit or something close to it? The make-or-break work is getting access to ADAS signals and making your labels hold up when an OEM pushes back, and I'm glad to help you plan for both. If you need a hand, you know where to find me. Talk to Yash →
For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.