Automated WLAN radio map construction method and system

1. An automated Wireless Local Area Network (WLAN) radio map construction method, comprising:

generating a learning model based on Finite State Automata (FSA) in which divided areas of an indoor map are expressed by location states; and

learning to arrange collected WLAN fingerprints in corresponding location states through machine learning based on the learning model,

wherein the learning comprises:

setting representative WLAN fingerprints for the location states; and

arranging the collected WLAN fingerprints in the corresponding location states with reference to the representative WLAN fingerprints, and

wherein the setting comprises:

finding Access Points (APs) from the collected WLAN fingerprints,

randomly arranging the found APs in the location states, and

assigning the representative WLAN fingerprints to the location states with reference to an arraignment of the APs and a WLAN signal propagation model,

wherein a number of APs included in an initial value of the learning model is variable,

wherein the number of APs included in the initial value is dynamically determined by increasing the number of APs until a condition in which a location range determined by trilateration is not distributed and is determined to be a single range less than or equal to a predetermined size is satisfied, and

wherein an order of the APs included in the initial value is determined with reference to Information Gain (IG).

2. The automated WLAN radio map construction method of claim 1 , further comprising collecting WLAN fingerprints obtained by a mobile device in an indoor space,

wherein the learning comprises arranging the collected WLAN fingerprints in corresponding location states.

3. The automated WLAN radio map construction method of claim 2 , wherein the WLAN fingerprints obtained by the mobile device do not comprise location information and movement information.

4. The automated WLAN radio map construction method of claim 1 , further comprising storing a result of the arranging.

5. The automated WLAN radio map construction method of claim 1 , wherein the setting and the arranging are repeated a predetermined number of times.

6. The automated WLAN radio map construction method of claim 1 , wherein the setting comprises changing the representative WLAN fingerprints for the location states based on the result of the arranging.

7. The automated WLAN radio map construction method of claim 6 , wherein the setting comprises changing the representative WLAN fingerprints for the location states with reference to all of the WLAN fingerprints arranged.

8. The automated WLAN radio map construction method of claim 1 , wherein the setting is performed through learning using at least one of an A* algorithm, a hill climbing algorithm, a genetic algorithm, and a memetic algorithm.

9. The automated WLAN radio map construction method of claim 1 , wherein the arranging is performed through learning using at least one of an Expectation Maximization (EM) algorithm and a segmental K means algorithm.

10. The automated WLAN radio map construction method of claim 1 , wherein some of the WLAN fingerprints is provided with location information, and

wherein the arranging comprises, with respect to the WLAN fingerprints provided with the location information, determining a location state for arranging with reference to the location information.

11. The automated WLAN radio map construction method of claim 10 , wherein the location information is provided through at least one of GPS information, information of a tag of which installation location is known, and movement information.

12. The automated WLAN radio map construction method of claim 1 , wherein the learning model is a model where the WLAN fingerprints are arranged in the FSA, and which has transition probabilities between the location states and probabilities that the WLAN fingerprints are observed as parameters.

13. An automated Wireless Local Area Network (WLAN) radio map construction system, comprising:

a collection server comprising a memory and a processor configured to collect WLAN fingerprints; and

a learning server comprising a memory and a processor configured to generate a learning model based on Finite State Automata (FSA) in which divided areas of an indoor map are expressed by location states, and arrange the WLAN fingerprints collected by the collection server in corresponding location states through machine learning based on the learning model,

wherein the processor of the learning server is further configured to set representative WLAN fingerprints for the location states and arrange the collected WLAN fingerprints in the corresponding location states with reference to the representative WLAN fingerprints, and

wherein the processor of the learning server is further configured to:

find Access Points (APs) from the collected WLAN fingerprints,

randomly arrange the APs in the location states, and

assign the representative WLAN fingerprints to the location states with reference to an arraignment of the APs and a WLAN signal propagation model,

wherein a number of APs included in an initial value of the learning model is variable,

wherein the number of APs included in the initial value is dynamically determined by increasing the number of APs until a condition in which a location range determined by trilateration is not distributed and is determined to be a single range less than or equal to a predetermined size is satisfied, and

wherein an order of the APs included in the initial value is determined with reference to Information Gain (IG).

14. The automated WLAN radio map construction system of claim 13 , wherein the processor of the collection server is configured to collect WLAN fingerprints obtained by a mobile device in an indoor space.

15. The automated WLAN radio map construction system of claim 14 , wherein the WLAN fingerprints obtained by the mobile device do not comprise location information and movement information.

16. The automated WLAN radio map construction system of claim 13 , wherein the processor of the learning server is configured to change the representative WLAN fingerprints for the location states based on a result of the arranging.

17. The automated WLAN radio map construction system of claim 13 , wherein the processor of the learning server is configured to change the representative WLAN fingerprints for the location states with reference to all of the WLAN fingerprints arranged.