What Is IDTS? Inside India's Automated Driving Test System
- IDTS (often expanded as Innovative/Intelligent Driving Test System) is the sensor-and-AI stack that scores driving tests automatically at accredited centres and modern RTO tracks.
- It fuses track sensors (RFID, infrared, ultrasonic), cameras, and vehicle telemetry (GNSS, OBD-II) into a live, objective score.
- Serious faults — crossing the boundary, rolling back on the gradient, jumping a red light, identity mismatch — terminate the test automatically.
- MoRTH's ADTC framework makes real-time automated evaluation a condition of accreditation; the test output feeds Form 5B.
Why India automated the driving test
A human examiner can watch one car, for a couple of minutes, from one angle — and every candidate knows the result can depend on the examiner's mood or the right introduction. With road crashes killing more people in India than anywhere else on Earth, MoRTH's reforms pushed evaluation from opinion to measurement: instrumented tracks where the vehicle's exact path, speed and control are recorded and scored by machine, identically for everyone.
What the sensors actually measure
| Zone / skill | Instrumentation | What decides pass or fail |
|---|---|---|
| Serpentine & 8-shape | Overhead cameras + path tracking | Staying on the ribbon; no kerb strikes; smooth steering |
| Up-gradient hold | Infrared beams + telemetry | Stop on slope, restart without rolling back beyond tolerance |
| Reverse-S | Cameras + boundary sensors | Reverse path accuracy, mirror use, no boundary contact |
| Parallel parking | Ultrasonic bay sensors + camera | Vehicle fully inside the bay within permitted shunts |
| Junction & signals | Signal controller + stop-line detection | Full stop at red, correct stop-line position |
| Speed & control | GNSS + OBD-II telemetry | Zone-appropriate speed, no stalls, controlled inputs |
The instant-fail list
Some mistakes end the test immediately, regardless of points accumulated — exactly as MoRTH's guidelines prescribe. The usual set: leaving the track boundary, hitting a kerb or barrier, rolling back beyond tolerance on the gradient, jumping a simulated red light, and — in modern deployments — an identity mismatch, where the in-cabin camera detects that the driver isn't the enrolled candidate.
Where AI fits (and where it doesn't)
Computer vision tracks the vehicle's trajectory against a reference path and reads events no buried sensor can — lane discipline, stop-line precision, even whether mirrors were checked in reverse manoeuvres. In-cabin AI performs continuous face-match against the Aadhaar-verified enrolment plus seatbelt and phone-use detection. What AI doesn't do is exercise discretion: thresholds are fixed in the scoring rules, every decision is logged, and disputed calls are settled by replaying the recording — not by arguing with an algorithm.
The camera doesn't know whose nephew you are. That single fact is most of the reform.
IDTS at accredited centres vs RTO tracks
The same technology family now appears in two places: automated RTO test tracks (where the RTO's own driving test is conducted) and Accredited Driver Training Centres (where the centre's automated test replaces the RTO test via Form 5B). For candidates, the ADTC route pairs the objective test with the training to pass it honestly. For governments, both routes replace queues and discretion with throughput and evidence.
What DigiChalak adds on top
DigiChalak is a complete IDTS implementation built for the ADTC framework — and it goes past the minimum: camera + RTK-GNSS fusion that keeps scoring even when a sensor fails, hash-chained video evidence that stands up to tampering claims, offline-first edge computing for real Indian connectivity, and the entire compliance layer (syllabus gating, Forms 5B/14/15, audit packs) wired to the test engine. The test isn't a bolt-on; it's the heart of the operating system.