Human-rating means redesigning and testing a rocket so it is safe enough to carry people, including a way for the crew to escape if anything goes wrong.
A rocket that carries satellites can accept some risk: a failure costs money and time. A rocket that carries people cannot. Human-rating is the work of making a rocket — and everything around it — safe enough for a crew.
For the Gaganyaan programme, ISRO is using a human-rated version of LVM3, called HLVM3. ISRO says the rocket has been reconfigured to meet human-rating requirements. Like LVM3, it has a solid stage, a liquid stage and a cryogenic stage. It is designed to carry the orbital module — the crew module plus service module — to a 400 km low Earth orbit, for a planned three-day mission with a crew of three. The crew module gives the crew an Earth-like environment in space and is built to keep them safe during re-entry and descent until touchdown.
The most visible part of human-rating is the Crew Escape System (CES). It sits on top of the crew module and is powered by quick-acting, fast-burning solid motors. If something goes wrong on the pad or during the climb, it pulls the crew module away from the rocket to a safe distance.
ISRO tested this in October 2023 with TV-D1, a single-stage test rocket powered by a modified Vikas engine. At about 11.7 km altitude, travelling at Mach 1.2, the escape system pulled the crew module away. The crew module was designed to come down into the sea under ten parachutes, touching the water at about 8.5 m/s.
ISRO lists the tests that must come before people fly:
ISRO's rule is simple: the safety and reliability of every system must be shown on uncrewed flights first.
Check which test flights and uncrewed missions have flown and which must still fly before astronauts launch. Follow Gaganyaan
Human-rating is about probability. A crew survives if the rocket works, or if it fails and the escape system works:
P(crew safe) = P(rocket ok) + P(rocket fails) × P(escape works)
Take hypothetical numbers: the rocket succeeds 98% of the time and the escape works 90% of the time when needed.
P(crew safe) = 0.98 + 0.02 × 0.90 = 0.98 + 0.018 = 0.998
The chance of losing the crew drops from 2% (1 in 50) to 0.2% (1 in 500). An escape system multiplies safety without making the rocket itself any more reliable. This is why so much testing goes into it.
The TV-D1 test conditions. For this first in-flight test, ISRO set the abort at about 11.7 km and Mach 1.2, faster than sound.
The descent after an abort. On TV-D1, ISRO's sequence shows the deceleration system starting at about 17 km, the drogue parachute at about 16.7 km and the main parachutes below about 2.5 km. The nominal touchdown speed was 8.5 m/s. In free-fall terms, v²/2g = 8.5² / (2 × 9.81) ≈ 3.7 m: hitting the sea at 8.5 m/s is like a drop of about 3.7 m.
Why test one piece at a time. Each test isolates part of the chain — parachutes (IADT), pad escape (PAT), in-flight escape (TV flights) — so a fault shows up where it is cheap to fix. Uncrewed orbital flights then test the whole system together before a crew rides it.
To pull the crew module quickly away from a failing rocket, on the pad or during the climb, so the crew can land safely by parachute.
0.98 + 0.02 × 0.90 = 0.998, or 99.8%.
An in-flight abort: the Crew Escape System pulled the crew module away from a test rocket at about 11.7 km and Mach 1.2, and the module landed in the sea under parachutes.