Developed mainly in control theory literature, *linear sensing
models* are some of the most common and important. For all of the
sensors in this family, assume that
(nonsingular
linear transformations allow the sensor space to effectively have
lower dimension, if desired). The simplest case in this family is the
*identity sensor*, in which . In this case, the state
is immediately known. If this sensor is available at every stage,
then the I-space collapses to by the I-map
.

Now nature sensing actions can be used to corrupt this perfect state observation to obtain . Suppose that is an estimate of , the current state, with error bounded by a constant . This can be modeled by assigning for every , as a closed ball of radius , centered at the origin:

Figure 11.11 illustrates the resulting nondeterministic sensing model. If the observation is received, then it is known that the true state lies within a ball in of radius , centered at . This ball is the preimage, , as defined in (11.11). To make the model probabilistic, a probability density function can be defined over . For example, it could be assumed that is a uniform density (although this model is not very realistic in many applications because there is a boundary at which the probability mass discontinuously jumps to zero).

A more typical probabilistic sensing model can be made by letting and defining a probability density function over all of . (Note that the nondeterministic version of this sensor is completely useless.) One of the easiest choices to work with is the multivariate Gaussian probability density function,

in which is the covariance matrix (11.64), is its determinant, and is a quadratic form, which multiplies out to yield

If is a Gaussian and is received, then must also be Gaussian under this model. This will become very important in Section 11.6.1.

The sensing models presented so far can be generalized by applying linear transformations. For example, let denote a nonsingular matrix with real-valued entries. If the sensor mapping is , then the state can still be determined immediately because the mapping is bijective; each contains a unique point of . A linear transformation can also be formed on the nature sensing action. Let denote an matrix. The sensor mapping is

(11.70) |

In general, and may even be singular, and a linear sensing model is still obtained. Suppose that . If is singular, however, it is impossible to infer the state directly from a single sensor observation. This generally corresponds to a projection from an -dimensional state space to a subset of whose dimension is the rank of . For example, if

(11.71) |

then yields and . Only of each can be observed because has rank . Thus, for some special cases, singular matrices can measure some state variables while leaving others invisible. For a general singular matrix , the interpretation is that is projected into some -dimensional subspace by the sensor, in which is the rank of . If is singular, this means that the effect of nature is limited. The degrees of freedom with which nature can distort the sensor observations is the rank of . These concepts motivate the next set of sensor models.

Steven M LaValle 2012-04-20