Using the time-sequential set of LRF measurement results
until time t expressed by z1:t and measurement results from
sensors for DR by the odometer and gyroscope expressed by
0:t1 u , the probability 1: 0: 1 , t t t p x z u of state vector xt of the
mobile robot at the given time t is formulated by the
recurrence formula Eq. (6), where t t p z x is the probability
distribution of the measurement model and 1 1 , t t t p x u x is
the probability distribution of the motion model.
presents our conclusions.