A ZMP based interval criterion for rollover-risk diagnosis

This paper presents an interval criterion used to anticipate rollover accidents of all-terrain wheeled vehicles. The Zero Moment Point concept, initially designed for humanoid robots, is presented in the case of a 4 wheels vehicle, first in an ideal case of a perfectly flat ground and finally in the off-road context. This adaptation is done by assuming that the tyre/ground contact is no more punctual but distributed in a set enclosed in a computable box (3D interval). It is then possible to derive an interval version of the ideal ZMP-based criterion. The method is experimented by using data collected during a real experiment done with a vineyard harvester that is a typical case of wheeled reconfigurable vehicle. The interval criterion gives a flexibility for its interpretation that is more adapted to the complex and varying all-terrain context.