求高手迅速英语翻译成中文when the robot senses, and when it moves, respectively. Suppose the robot just sensed s. Markov localization then P (l j s) = ff P(s j l) P(l)where ff is a normalizer that ensures that the resulting prob- abilities sum up to one. When the robot moves, Markov localization updates P(l) ability: P ( 0l ) = using the Theorem of total prob- Z P ( 0l j a;l) P(l) dl Here a denotes an action command. These two u

求高手迅速英语翻译成中文
when the robot senses, and when it moves, respectively.
Suppose the robot just sensed s. Markov localization then
P (l j s) = ff P(s j l) P(l)
where ff is a normalizer that ensures that the resulting prob-
abilities sum up to one. When the robot moves, Markov
localization updates P(l)
ability:
P ( 0l ) =
using the Theorem of total prob-
Z
P ( 0l j a;l) P(l) dl
Here a denotes an action command. These two update
equations form the basis of Markov localization. Strictly
speaking, they are only applicable if the environment meets
a conditional independence assumption (Markov assump-
tion), which specifies that the robot's pose is the only state
therein. Put differently, Markov localization applies only to
static environments.
Unfortunately, the standard Markov localization ap-
proach is prone to fail in densely populated environments,
since those violate the underlying Markov assumption. In
the museum, people frequently blocked the robot's sensors,
as illustrated in Figure 1. Figuratively speaking, if people
line up as a "wall" in front of the robot—which they often
did—, the basic Markov localization paradigm makes the
robot eventually believe that it is indeed in front of a wall.
To remedy this problem, RHINO employs an "entropy
filter" (Fox et al. 1998b). This filter, which is applied to all
proximity measurements individually, sorts measurements
into two buckets: one that is assumed to contain all cor-
rupted sensor readings, and one that is assumed to contain
only authentic (non-corrupted) ones. To determine which
sensor reading is corrupted and which one is not, the en-
tropy filter measures the relative entropy of the belief state
before and after incorporating a proximity measurement:
P(l) logP(l) dl + P(l j s)logP(l j s) dl
l Sensor readings that increase the robot's certainty
(_H(l;s) > 0 ) are assumed to be authentic. All other sen-
sor readings are assumed to be corrupted and are therefore
not incorporated into the robot's belief. In the museum,
certainty filters reliably identified sensor readings that were
corrupted by the presence of people, as long as the robot
knew its approximate pose. Unfortunately, the entropy fil-
ter can prevent recovery once the robot looses its position
entirely. To prevent this problem, our approach also incor-
porates a small number of randomly chosen sensor readings
in addition to those selected by the entropy filter. See (Fox
et al. 1998b) for an alternative solution to this problem.