Coarse location estimation for mobile devices
Coarse location estimation for mobile devices is disclosed for detecting mobile device presence at general locations of interest and switching operating modes and services for one or more location context aware applications. In some implementations, sensor data is received from a plurality of data sources at a location. For each data source, a first probability is estimated that the mobile device is at the location based on sensor data from the data sources. A second probability is estimated that the mobile device is not at the location based on sensor data from the data sources. The first and second estimated probabilities are statistically combined to generate a third estimated probability that the mobile device is at the location.
1. A method comprising:
receiving, by a processor of a computing device operating at a location, sensor data from a plurality of data sources;
determining, by the processor, a first estimated probability that the computing device is at the location based on first sensor data from a wireless network data source and a first data source model;
determining, by the processor, a second estimated probability that the computing device is at the location based on second sensor data from a Global Navigation Satellite System (GNSS) data source and a second data source model;
determining, by the processor, a third estimated probability that the computing device is at the location based on particle filter data;
combining, by the processor, the first, second and third estimated probabilities; and
generating, by the processor, a coarse estimate of the location of the computing device based on the combined first, second and third estimated probabilities; and
displaying the coarse estimate on the computing device.
2. The method of claim 1 , where combining the first, second and third estimated probabilities further comprises:
combining the first, second and third estimated probabilities in a Bayesian classifier.
3. The method of claim 2 , where the first estimated probability is given by
P
(
L
i
|
d
wifi
(
t
)
)
=
P
(
d
wifi
(
t
)
|
L
i
)
η
and the second estimated probability is given by
P
(
L
i
|
d
gps
(
t
)
)
=
P
(
d
gps
(
t
)
|
L
i
)
η
,
where η is a normalization constant, d gps (t) is GNSS data, L i is a state of being at location i and d wifi (t) is wireless network data.
4. The method of claim 3 , where the third estimated probability is given by
P
(
L
i
|
d
p
(
t
)
)
=
P
(
dp
(
t
)
|
L
i
)
η
,
and d p (t) is the particle filter data.
5. The method of claim 1 , where generating a coarse estimate of the location of the computing device based on the combined first, second and third estimated probabilities further comprises:
determining a maximum likelihood that the computing device is at the location based on the first, second and third estimated probabilities.
6. A computing device, comprising:
one or more processors;
memory coupled to the one or more processors and storing instructions, which, when executed by the one or more processors, causes the one or more processors to perform operations comprising:
receiving sensor data from a plurality of data sources;
determining a first estimated probability that the computing device is at a location based on first sensor data from a wireless network data source and a first data source model;
determining a second estimated probability that the computing device is at the location based on second sensor data from a Global Navigation Satellite System (GNSS) data source and a second data source model;
determining a third estimated probability that the computing device is at the location based on a particle filter model;
combining the first, second and third estimated probabilities;
generating a coarse estimate of the location of the computing device based on the combined first, second and third estimated probabilities; and
displaying the coarse estimate on the computing device.
7. The computing device of claim 6 , where combining the first and second estimated probabilities further comprises:
combining the first, second and third estimated probabilities in a Bayesian classifier.
8. The computing device of claim 7 , where the first estimated probability is given by
P
(
L
i
|
d
wifi
(
t
)
)
=
P
(
d
wifi
(
t
)
|
L
i
)
η
and the second estimated probability is given by
P
(
L
i
|
d
gps
(
t
)
)
=
P
(
d
gps
(
t
)
|
L
i
)
η
,
L i is a state of being at location i, where η is a normalization constant, d gps (t) is GNSS data and d wifi (t) is wireless network data.
9. The computing device of claim 8 , where the third estimated probability is given by
P
(
L
i
|
d
p
(
t
)
)
=
P
(
dp
(
t
)
|
L
i
)
η
,
and d p (t) is the particle filter data.
10. The computing device of claim 6 , where generating a coarse estimate of the location of the computing device based on the combined first, second and third estimated probabilities further comprises:
determining a maximum likelihood that the computing device is at the location based on the first, second and third estimated probabilities.
11. A non-transitory, computer-readable storage medium storing instructions, which, when executed by one or more processors, causes the one or more processors to perform operations comprising:
receiving sensor data from a plurality of data sources;
determining a first estimated probability that a computing device is at a location based on first sensor data from a wireless network data source and a first data source model;
determining a second estimated probability that the computing device is at the location based on second sensor data from a Global Navigation Satellite System (GNSS) data source and a second data source model;
determining a third estimated probability that the computing device is at the location based on a particle filter model;
combining the first, second and third estimated probabilities;
generating a coarse estimate of the location of the computing device based on the combined first, second and third estimated probabilities; and
displaying the coarse estimate on the computing device.
12. The computer-readable storage medium of claim 11 , where combining the first and second estimated probabilities further comprises:
combining the first, second and third estimated probabilities in a Bayesian classifier.
13. The computer-readable storage medium of claim 11 , where the first estimated probability is given by
P
(
L
i
|
d
wifi
(
t
)
)
=
P
(
d
wifi
(
t
)
|
L
i
)
η
and the second estimated probability is given by
P
(
L
i
|
d
gps
(
t
)
)
=
P
(
d
gps
(
t
)
|
L
i
)
η
,
where η is a normalization constant, d gps (t) is GNSS data and d wifi (t) is wireless network data.
14. The computer-readable storage medium of claim 13 , where the third estimated probability is given by
P
(
L
i
|
d
p
(
t
)
)
=
P
(
dp
(
t
)
|
L
i
)
η
,
and d p (t) is the particle filter data.
15. The computer-readable storage medium of claim 11 , where generating a coarse estimate of the location of the computing device based on the combined first, second and third estimated probabilities further comprises:
determining a maximum likelihood that the computing device is at the location based on the first, second and third estimated probabilities.