Floor selection for structure location tracking
There is described a floor selection system for location tracking within a structure. A communication component receives sensor data, associated with a beacon received from a tag, from sensors. A processor identifies features based on the sensor data, identifies decision tree stumps arranged in a particular order, assigns weights to the decision tree stumps based on the features and the feature thresholds. The processor further determines whether an aggregate of the weights is greater than a resultant threshold and selects a floor location of the tag based on whether the aggregate of weights is greater than the resultant threshold. The floor location is selected from floor locations of the structure proximal to the sensors.
1. A floor selection system for location tracking within a structure comprising:
a communication component configured to receive sensor data from a plurality of sensors, the sensor data being associated with a beacon received from a tag by the plurality of sensors; and
a processor configured to identify a plurality of features based on the sensor data, identify a plurality of decision tree stumps arranged in a particular order, assign weights to the plurality of decision tree stumps based on the plurality of features and a plurality of feature thresholds of the plurality of decision trees, determine whether an aggregate of the weights is greater than a resultant threshold, and select a floor location of the tag based on whether the aggregate of weights is greater than the resultant threshold,
wherein the plurality of features include a difference between distances to a transport of adjacent floors,
wherein the plurality of features are derived from an original feature set and have at least 50% association with ground truth, and
wherein the floor location is selected from a plurality of floor locations of the structure proximal to the plurality of sensors.
2. The floor selection system as described in claim 1 , wherein the processor gene rates the aggregate of the weights by repeatedly executing the assigning the weights to the plurality of decision tree stumps.
3. The floor selection system as described in claim 1 , wherein the plurality of features includes a difference of a quantity of sensors between adjacent floors.
4. The floor selection system as described in claim 1 , wherein the plurality of features include an area of probability where a sensor may be located for each floor of adjacent floors.
5. The floor selection system as described in claim 1 , wherein the plurality of features include a floor identification associated with a specific sensor of the plurality of sensors with a strongest radio frequency signal.
6. The floor selection system as described in claim 1 , wherein the plurality of features include a difference in sums of a plurality of radio frequency signals between adjacent floors.
7. The floor selection system as described in claim 1 , wherein the plurality of features include a difference of weighted motion values of a specific sensor, the weighted motion values being weighted by a distance from the specific sensor.
8. The floor selection system as described in claim 1 , wherein the weights and the corresponding feature thresholds are determined by using decision tree stumps as weak classifiers.
9. The floor selection system as described in claim 1 , wherein the processor assigns the weights to the plurality of decision tree stumps by determining whether the individual features of the plurality of features are greater than the corresponding feature thresholds for each decision tree stump.
10. A floor selection method for location tracking within a structure comprising:
receiving sensor data from a plurality of sensors, the sensor data being associated with a beacon received from a tag by the plurality of sensors;
identifying a plurality of features in response to receiving the sensor data from the plurality of sensors, wherein the plurality of features include a difference between distancesto a transport of adjacent floors, wherein the plurality of features are derived from an original feature set and have at least 50% association with ground truth;
identifying a plurality of decision tree stumps arranged in a particular order, the plurality of decision tree stumps including a plurality of feature thresholds;
assigning weights to the plurality of decision tree stumps based on the plurality of features and the plurality of feature thresholds;
determining whether an aggregate of the weights is greater than a resultant threshold; and
selecting a floor location of the tag in response to determining whether the aggregate of weights is greater than the resultant threshold, wherein the floor location is selected from a plurality of floor locations of the structure proximal to the plurality of sensors.
11. The floor selection method as described in claim 10 , further comprising generating the aggregate of the weights by repeatedly executing the assigning of the weights to the plurality of decision tree stumps.
12. The floor selection method as described in claim 10 , wherein identifying the plurality of features includes identifying a particular feature of the plurality of features as a difference of a quantity of sensors between adjacent floors.
13. The floor selection method as described in claim 10 , wherein identifying the plurality of features includes identifying a particular feature of the plurality of features as an area of probability where a sensor may be located for each floor of adjacent floors.
14. The floor selection method as described in claim 10 , wherein identifying the plurality of features includes identifying a particular feature of the plurality of features as a floor identification associated with a specific sensor of the plurality of sensors with a strongest radio frequency signal.
15. The floor selection method as described in claim 10 , wherein identifying the plurality of features includes identifying a particular feature of the plurality of features as a difference in sums of a plurality of radio frequency signals between adjacent floors.
16. The floor selection method as described in claim 10 , wherein identifying the plurality of features includes identifying a particular feature of the plurality of features as a difference of weighted motion values of a specific sensor, the weighted motion values being weighted by a distance from the specific sensor.
17. The floor selection method as described in claim 10 , further comprising determining the weights and the corresponding feature thresholds by using decision tree stumps as weak classifiers.
18. The floor selection method as described in claim 10 , wherein assigning the weights to the plurality of decision tree stumps includes determining whether the individual features of the plurality of features are greater than the corresponding feature thresholds for each decision tree stump.