Let us consider a launcher having a cryogenic core stage (as Ariane 5 or Delta 4 Medium), the crew being in an "Apollo-like" capsule during the ascent phase. The low composite will be made of a cryogenic core stage and strap-on boosters using hybrid propulsion; on the cryogenic stage, we will have (from the bottom to the top) an orbital service module, a capsule and an extraction tower.If we compare it to solid propulsion, hybrid propulsion (liquid oxidizer and solid propellant) will give the possibility to shut down the booster in many failure cases. If we can order the shut down of the booster while firing the crew escape tower motors, the distancing between the man-rated vehicle and the hybrid propulsion booster will be greater (in the worst case) than the distancing between the manrated vehicle and the solid propulsion boosters; the burst risk of the booster (so the projection risk of motor case or hot propellant debris towards the manrated vehicle) will decrease.If we compare it to liquid propulsion, hybrid propulsion offers more simplicity, then more reliability, inducing better crew safety. Indeed, for liquid propulsion boosters, generally, high thrust levels are required, so turbo-pumps are mandatory. With hybrid propulsion, we can avoid turbo-pumps: that simplifies the propulsion system and also its health monitoring (because failure modes are less complex); so, with hybrid propulsion, we can wish than the failure detection system will have better coverage and false alarm rates; this will increase the crew safety level.Of course, hybrid propulsion has also disadvantages:-Ground experience, but no flight experience for great boosters (but this argument forbids any innovation), -more complex than solid propulsion (e.g. specific instabilities).In that framework, we could build a failure detection system providing acceptable coverage and false alarm rates, considering flight proven technologies for the sensors (pressure measurements inside the booster tanks, gauges, control jacks measurements, Inertial Measurement Unit measurements). The design will be based upon segregation and differentiation principles. The detection system design will have to be integrated into the global system design, considering the booster propulsion aspects, structural propulsion aspects, tower escape performances aspects.Because the accidents and explosions database are relatively poor, we will have to privilege robustness if associated performance losses remain acceptable.