NETWORK SELECTION RECOMMENDER SYSTEM AND METHOD

• US 20150131483A1
• Filed: 11/12/2014
• Published: 05/14/2015
• Est. Priority Date: 11/12/2013
• Status: Active Grant

• Current Assignee: Taiwan Semiconductor Manufacturing Company Limited
• Original Assignee: Wi-LAN Labs, Inc. (Wi-LAN Inc.)
• Inventors: Colban, Erik, Gell, David
• Granted Patent: US 10,015,737 B2
• Time in Patent Office: Days
• US Class Current: 370/254
• CPC Class Codes: H04W 24/08 Testing, supervising or monitoring using real traffic
• CPC Class Codes: H04W 48/14 using user query or user detection
• CPC Class Codes: H04W 48/16 Discovering, processing access restriction or access information
• CPC Class Codes: H04W 48/17 Selecting a data network PoA [Point of Attachment]
• CPC Class Codes: H04W 48/18 Selecting a network or a communication service
• CPC Class Codes: H04W 48/20 Selecting an access point
• CPC Class Codes: H04W 84/12 WLAN [Wireless Local Area Networks]

Abstract

A method, network device and computer program product determine a recommended access point (AP) for a wireless device to access a wireless network. AP feature values associated with each one of a plurality of APs within an access range of the wireless device and user feature values associated with identified user features of a user of the wireless device are obtained via a wireless interface. A predicted score for each AP is determined based on the feature values and a recommended AP is determined based on the predicted scores. An identifier associated with the recommended access point is transmitted to the wireless device. AP feature values include AP characteristic, scheduling and payment values. User features include wireless device location and velocity, services-in-use, time of day and day of week. Optionally, circumstantial feature values may be obtained and used to determine the predicted scores.

Claims

1. A method of determining, in a network device having a network interface, a recommended access point for a wireless device to access a wireless network, the method comprising:

obtaining, via the network interface, access point feature values associated with each one of a plurality of access points within an access range of the wireless device;

obtaining, via the network interface, user feature values associated with identified user features of a user of the wireless device;

determining a predicted score for each access point of the plurality of access points based on at least one of the obtained access point feature values and the obtained user feature values;

determining a recommended access point of the plurality of access points based on the predicted scores associated with the plurality of access points; and

transmitting, via the network interface, an identifier associated with recommended access point to the wireless device.

2. The method of claim 1, wherein the access point feature values comprise at least one of:

an access point characteristic value, a scheduling value, and a payment value.

3. The method of claim 2, wherein the access point characteristic value is at least one of the group comprising:

an access point distance value, an access point distance differential value, an access point environment value, a coverage area radius value, a services offered value, a throughput value, a backhaul bandwidth value, an air interface channel utilization value, a round trip delay value, a packet data network reachability value, a direct internet access value, a local area network access value, a 3 GPP-trusted value, a security value, a mobility support value, a closed subscriber group value, a signal strength value, a noise and interference value, a signal to noise and interference ratio value, an air interface technology value and an operator policies value.

4. The method of claim 1, wherein the identified user features are one or more of the group comprising:

a wireless device location, a wireless device velocity, a services-in-use, a time of day and a day of week.

5. The method of claim 1, further comprising:

obtaining circumstantial feature values; and

wherein the predicted score for each access point is further based on at least one of the obtained circumstantial feature values.

6. The method of claim 1, wherein the predicted score for each access point is determined based on a sum of each feature value multiplied by a weighting factor associated with a particular user of the wireless device.

7. The method of claim 1, wherein the predicted score for each access point is determined based on a sum of products of each access point feature value multiplied by a weighting factor, and wherein each weighting factor is a function of the user feature values.

8. The method of claim 7, wherein the weighting factor is a linear function of the user feature values.

9. The method of claim 1, wherein the user of the wireless device is represented as a combination of two or more partial users, and the predicted score is determined by:

determining a partial predicted score for each partial user based on at least one of the obtained access point feature values;

determining a coefficient associated with each partial user based on at least one of the obtained user feature values;

multiplying the partial predicted score by the corresponding coefficient for each partial user to obtain a predicted score product for each partial user; and

summing all of the predicted score products to determine the predicted score.

10. The method of claim 1, wherein the determining a predicted score for each access point is conducted by applying a model to the at least one of the obtained access point feature values and the at least one of the obtained user feature values.

11. The method of claim 10, wherein the model is derived using collaborative filtering.

12. The method of claim 1, further comprising:

deriving a user satisfaction score associated with at least one access point of the plurality of access points;

deriving a sample for the at least one access point of the plurality of access points based on the user satisfaction score and one or more of at least one of the obtained access point feature values and at least one of the obtained user feature values; and

storing the sample for the at least one access point of the plurality of access points.

13. The method of claim 12, wherein the user satisfaction score is derived using at least one of:

a user experience metric associated with the at least one access point, a user access point preference setting, a user manual selection and a user-indicated satisfaction indication.

14. The method of claim 13, wherein the user experience metric associated with the at least one access point is one or more of:

a packet retransmissions indication, a download speed indication, a received signal strength indication, a frequency of playback buffer underflows indication, a dropped calls indication, a connection time indication, and a frequency of returns to respective access point.

15. The method of claim 13, wherein the user access point preference indication is one or more of a preferred access point identification and a bookmarked access point indication.

16. The method of claim 12, further comprising generating a model based on the derived samples corresponding to the at least one of the plurality of access points.

17. The method of claim 16, wherein the model is associated with a particular user and the model is generated based on the derived samples corresponding to the particular user.

18. The method of claim 16, wherein the model is based on collaborative filtering, and the derived samples correspond to a plurality of users.

19. The method of claim 16, further comprising normalizing all of the obtained access point feature values and the obtained user feature values.

20. The method of claim 16, further comprising training the model to determine feature weighting factors by minimizing a cost function.

21. A network device for determining a recommended access point to allow a wireless device to access a wireless network, the network device comprising:

a network interface; and

a processor, in communication with the network interface, that;

obtains, via the network interface, access point feature values associated with each one of a plurality of access points within an access range of the wireless device;

obtains, via the network interface, user feature values associated with identified user features of a user of the wireless device;

determines a predicted score for each access point of the plurality of access points based on at least one of the obtained access point feature values and the obtained user feature values; and

determines a recommended access point of the plurality of access points based on the predicted scores associated with the plurality of access points; and

the network interface transmits an identifier associated with the recommended access point to the wireless device.

22. The network device of claim 21, wherein the access point feature values comprise at least one of:

an access point characteristic value, a scheduling value, and a payment value.

23. The network device of claim 21, wherein the access point characteristic value is at least one of the group comprising:

an access point distance value, an access point distance differential value, an access point environment value, a coverage area radius value, a services offered value, a throughput value, a backhaul bandwidth value, an air interface channel utilization value, a round trip delay value, a packet data network reachability value, a direct internet access value, a local area network access value, a 3GPP-trusted value, a security value, a mobility support value, a closed subscriber group value, a signal strength value, a noise and interference value, a signal to noise and interference ratio value, an air interface technology value and an operator policies value.

24. The network device of claim 21, wherein the identified user features are one or more of the group comprising:

a wireless device location, a wireless device velocity, a services-in-use, a time of day and a time of week.

25. The network device of claim 21, wherein the access point selector further obtains circumstantial feature values and wherein the predicted score for each access point is further based on at least one of the obtained circumstantial feature values.

26. The network device of claim 21, wherein the processor determines the predicted score for each access point based on a sum of each feature value multiplied by a weighting factor associated with a particular user of the wireless device.

27. The network device of claim 21, wherein the processor determines the predicted score for each access point based on a sum of products of each access point feature value multiplied by a weighting factor, and wherein each weighting factor is a function of the user feature values.

28. The network device of claim 27, wherein the weighting factor is a linear function of the user feature values.

29. The network device of claim 21, wherein the user of the wireless device is represented as a combination of two or more partial users, and the processor determines the predicted score by:

determining a partial predicted score for each partial user based on at least one of the obtained access point feature values;

determining a coefficient associated with each partial user based on at least one of the obtained user feature values;

multiplying the partial predicted score by the corresponding coefficient for each partial user to obtain a predicted score product for each partial user; and

summing all of the predicted score products to determine the predicted score.

30. The network device of claim 21, wherein the processor determines a predicted score for each access point by applying a model to the at least one of the obtained access point feature values and the at least one of the obtained user feature values.

31. The network device of claim 30, wherein the processor further derives the model using collaborative filtering.

32. The network device of claim 21, further comprising a sample storage and wherein the processor further:

derives a user satisfaction score associated with at least one access point of the plurality of access points;

derives a sample for the at least one access point of the plurality of access points based on the evaluated user satisfaction score and one or more of at least one of the obtained access point feature values and at least one of the obtained user feature values; and

stores the sample for the at least one access point of the plurality of access points in the sample storage.

33. The network device of claim 32, wherein the processor evaluates the user satisfaction score using at least one of:

a user experience metric associated with the at least one access point, a user access point preference indication, a user manual selection and a user-indicated satisfaction indication.

34. The network device of claim 33, wherein the user experience metric associated with the at least one access point is one or more of:

a packet retransmissions indication, a download speed indication, a received signal strength indication, a frequency of playback buffer underflows indication, a dropped calls indication, a connection time indication, and a frequency of returns to respective access point.

35. The network device of claim 33, wherein the user access point preference indication is one or more of a preferred access point identification and a bookmarked access point indication.

36. The network device of claim 32, wherein the processor further generates a model based on the derived samples corresponding to the at least one of the plurality of access points.

37. The network device of claim 36, wherein the model is associated with a particular user and the processor generates the model based on the derived samples corresponding to the particular user.

38. The network device of claim 36, wherein the model is based on collaborative filtering, and the derived samples correspond to a plurality of users.

39. The network device of claim 36, wherein the processor further normalizes all of the obtained access point feature values and the obtained user feature values.

40. The network device of claim 36, wherein the processor further trains the model to determine feature weighting factors by minimizing a cost function.

41. A computer program product for determining a recommended access point for a wireless device having a network interface, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by the network device to cause the network device to perform a method comprising:

obtaining, via the network interface, access point feature values associated with each one of a plurality of access points within an access range of the wireless device;

obtaining, via the network interface, user feature values associated with identified user features of a user of the wireless device;

determining a predicted score for each access point of the plurality of access points based on at least one of the obtained access point feature values and the obtained user feature values;

determining a recommended access point of the plurality of access points based on the predicted scores associated with the plurality of access points; and

transmitting an identifier associated with recommended access point to the wireless device.