Machine learning of known or unknown motion states with sensor fusion
원문보기
IPC분류정보
국가/구분
United States(US) Patent
등록
국제특허분류(IPC7판)
G06F-017/20
G01D-015/00
G06N-099/00
출원번호
US-0269513
(2011-10-07)
등록번호
US-8756173
(2014-06-17)
발명자
/ 주소
Hunzinger, Jason Frank
Sarah, Anthony
Baheti, Pawan K.
출원인 / 주소
Qualcomm Incorporated
대리인 / 주소
Kilpatrick Townsend & Stockton LLP
인용정보
피인용 횟수 :
7인용 특허 :
1
초록
Example methods, apparatuses, or articles of manufacture are disclosed herein that may be utilized, in whole or in part, to facilitate or support one or more operations or techniques for machine learning of known or unknown motion states with sensor fusion.
대표청구항▼
1. A method comprising: extracting, at a mobile device, at least one feature from at least one sensor signal;recognizing at least one primitive based, at least in part, on said at least one extracted feature, said at least one primitive being indicative of a pattern of movement of said mobile device
1. A method comprising: extracting, at a mobile device, at least one feature from at least one sensor signal;recognizing at least one primitive based, at least in part, on said at least one extracted feature, said at least one primitive being indicative of a pattern of movement of said mobile device; andidentifying a new motion state based, at least in part, on said at least one recognized primitive. 2. The method of claim 1, wherein said new motion state is distinguishable from a previously known motion state based, at least in part, on a measure of confidence or consistency. 3. The method of claim 2, wherein said measure of confidence or consistency comprises a cumulative statistical probability of a sample distribution within a cluster. 4. The method of claim 2, wherein said measure of confidence or consistency comprises an individual iteration-based statistical probability. 5. The method of claim 1, and further comprising initiating discovery of at least one additional primitive in response to said new motion state not being distinguishable from at least one of the following: a previously known motion state, said new motion state, a new distinguishable motion state, or any combination thereof. 6. The method of claim 5, wherein said discovery of said at least one additional primitive is initiated based, at least in part, on a measure of confidence or consistency. 7. The method of claim 5, wherein said discovery of said at least one additional primitive is performed in connection with an application of a genetic algorithm-type process. 8. The method of claim 5, and further comprising storing at least one digital signal representing said new motion state as a known motion state defined, at least in part, by said at least one additional primitive. 9. The method of claim 1, and further comprising identifying a previously known motion state based, at least in part, on said at least one recognized primitive. 10. The method of claim 1, wherein said recognizing said at least one primitive is based, at least in part, on a probabilistic logical determination. 11. The method of claim 1, wherein said at least one feature comprises at least one pre-defined feature. 12. The method of claim 1, wherein said identifying said new motion state comprises identifying said state while said mobile device is in a power-saving mode. 13. The method of claim 12, wherein said power-saving mode comprises an operating mode sufficient to identify said new motion state based, at least in part, on a minimum number of primitives. 14. The method of claim 13, wherein while operating in said power-saving mode said mobile device uses, at least in part, sensor configuration with variable sensor activation parameters. 15. The method of claim 1, wherein said new motion state comprises at least one of the following: an unknown distinguishable motion state, an unknown indistinguishable motion state, or any combination thereof. 16. The method of claim 15, wherein said unknown indistinguishable motion state is identified based, at least in part, on a primitive discovery. 17. The method of claim 15, wherein said unknown distinguishable motion state is identified based, at least in part, on at least one new logical inference. 18. The method of claim 1, wherein said new motion state is identified in response to a user interaction with said mobile device. 19. The method of claim 1, wherein said at least one sensor signal comprises at least one of the following: an inertial sensor signal, an ambient environment sensor signal, or any combination thereof. 20. The method of claim 1, wherein said at least one sensor signal is generated, at least in part, by at least one of the following disposed in said mobile device: an accelerometer, a gyroscope, a magnetometer, a temperature sensor, an ambient light detector, a proximity sensor, a barometric pressure sensor, or any combination thereof. 21. An apparatus comprising: a mobile device comprising at least one processor configured to:extract at least one feature from at least one sensor signal;recognize at least one primitive based, at least in part, on said at least one extracted feature, said at least one primitive being indicative of a pattern of movement of said mobile device; andidentify a new motion state based, at least in part, on said at least one recognized primitive. 22. The apparatus of claim 21, wherein said new motion state is distinguishable from a previously known motion state based, at least in part, on a measure of confidence or consistency. 23. The apparatus of claim 22, wherein said measure of confidence or consistency comprises a cumulative statistical probability of a sample distribution within a cluster. 24. The apparatus of claim 21, wherein said at least one processor is further configured to: initiate discovery of at least one additional primitive in response to said new motion state not being distinguishable from at least one of the following: a previously known motion state, said new motion state, a new distinguishable motion state, or any combination thereof. 25. The apparatus of claim 24, wherein said discovery of said at least one additional primitive is performed in connection with an application of a genetic algorithm-type process. 26. The apparatus of claim 24, wherein said at least one processor is further configured to: store at least one digital signal representing said new motion state as a known motion state defined, at least in part, by said at least one additional primitive. 27. The apparatus of claim 21, wherein said at least one processor is further configured to: recognize said new motion state responsive to a user interaction with said mobile device. 28. The apparatus of claim 21, wherein said at least one processor is further configured to: identify a previously known motion state based, at least in part, on said at least one recognized primitive. 29. The apparatus of claim 21, wherein said at least one processor is configured to recognize said at least one primitive based, at least in part, on a probabilistic logical determination. 30. The apparatus of claim 21, wherein said at least one processor is configured to identify said new motion state by identifying said state while said mobile device is in a power-saving mode sufficient to identify said new motion state based, at least in part, on a minimum number of primitives. 31. The apparatus of claim 21, wherein said new motion state comprises at least one of the following: an unknown distinguishable motion state, an unknown indistinguishable motion state, or any combination thereof. 32. The apparatus of claim 31, wherein said at least one processor is configured to identify said unknown indistinguishable motion state based, at least in part, on a primitive discovery. 33. The apparatus of claim 31, wherein said at least one processor is configured to identify said unknown distinguishable motion state based, at least in part, on at least one new logical inference. 34. The apparatus of claim 21, wherein said at least one processor is configured to identify said new motion state in response to a user interaction with said mobile device. 35. The apparatus of claim 21, wherein said at least one sensor signal comprises at least one of the following: an inertial sensor signal, an ambient environment sensor signal, or any combination thereof. 36. An apparatus comprising: means for extracting, at a mobile device, at least one feature from at least one sensor signal; means for recognizing at least one primitive based, at least in part, on said at least one extracted feature, said at least one primitive being indicative of a pattern of movement of said mobile device; andmeans for identifying a new motion state based, at least in part, on said at least one recognized primitive. 37. The apparatus of claim 36, said new motion state is distinguishable from a previously known motion state based, at least in part, on a measure of confidence or consistency. 38. The apparatus of claim 37, wherein said measure of confidence or consistency comprises a cumulative statistical probability of a sample distribution within a cluster. 39. The apparatus of claim 36, and further comprising means for initiating discovery of at least one additional primitive in response to said new motion state not being distinguishable from at least one of the following: a previously known motion state, said new motion state, a new distinguishable motion state, or any combination thereof. 40. The apparatus of claim 39, wherein initiating discovery of said at least one additional primitive is initiated based, at least in part, on a measure of confidence or consistency. 41. The apparatus of claim 39, wherein said discovery of said at least one additional primitive is performed in connection with an application of a genetic algorithm-type process. 42. The apparatus of claim 39, and further comprising means for storing at least one digital signal representing said new motion state as a known motion state defined, at least in part, by said at least one additional primitive. 43. The apparatus of claim 36, and further comprising means for identifying a previously known motion state based, at least in part, on said at least one recognized primitive. 44. The apparatus of claim 36, wherein said means recognize said at least one primitive based, at least in part, on a probabilistic logical determination. 45. The apparatus of claim 36, wherein said means identify said new motion state while said mobile device is in a power-saving mode. 46. The apparatus of claim 45, wherein said power-saving mode comprises an operating mode sufficient to identify said new motion state based, at least in part, on a minimum number of primitives. 47. The apparatus of claim 46, wherein, while operating in said power-saving mode, said mobile device uses, at least in part, sensor configuration with variable sensor activation parameters. 48. The apparatus of claim 36, wherein said new motion state comprises at least one of the following: an unknown distinguishable motion state, an unknown indistinguishable motion state, or any combination thereof. 49. The apparatus of claim 48, wherein said unknown indistinguishable motion state is identified based, at least in part, on a primitive discovery. 50. The apparatus of claim 48, wherein said unknown distinguishable motion state is identified based, at least in part, on at least one new logical inference. 51. The apparatus of claim 36, wherein said new motion state is identified in response to a user interaction with said mobile device. 52. The apparatus of claim 36, wherein said at least one sensor signal comprises at least one of the following: an inertial sensor signal; an ambient environment sensor signal; or any combination thereof. 53. An article comprising: a non-transitory storage medium having instructions stored thereon executable by a special purpose computing platform at a mobile device to:extract at least one feature from at least one sensor signal;recognize at least one primitive based, at least in part, on said at least one extracted feature, said at least one primitive being indicative of a pattern of movement of said mobile device; andidentify a new motion state based, at least in part, on said at least one recognized primitive. 54. The article of claim 53, wherein said new motion state is distinguishable from a previously known motion state based, at least in part, on a measure of confidence or consistency. 55. The article of claim 54, wherein said measure of confidence or consistency comprises a cumulative statistical probability of a sample distribution within a cluster. 56. The article of claim 53, wherein said storage medium further includes instructions to: initiate discovery of at least one additional primitive in response to said new motion state not being distinguishable from at least one of the following: a previously known motion state, said new motion state, a new distinguishable motion state, or any combination thereof. 57. The article of claim 53, wherein said storage medium further includes instructions to: identify a previously known motion state based, at least in part, on said at least one recognized primitive. 58. The article of claim 53, wherein said instructions to said recognize said at least one primitive is based, at least in part, on a probabilistic logical determination. 59. The article of claim 53, wherein said instructions to identify said new motion state comprise instructions to identify said state while said mobile device is in a power-saving mode sufficient to said identify said new motion state based, at least in part, on a minimum number of primitives. 60. The article of claim 53, wherein said new motion state comprises at least one of the following: an unknown distinguishable motion state identified based, at least in part, on a primitive discovery; an unknown indistinguishable motion state identified based, at least in part, on at least one new logical inference; or any combination thereof.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (1)
Sorvari, Antti; Kähäri, Markus; Toivonen, Hannu; Mannila, Heikki; Salmenkaita, Jukka Pekka, System and method for providing context sensitive recommendations to digital services.
Hunzinger, Jason Frank; Chan, Victor Hokkiu; Levin, Jeffrey Alexander, Method and apparatus for structural delay plasticity in spiking neural networks.
Hunzinger, Jason Frank; Chan, Victor Hokkiu, Method and apparatus of robust neural temporal coding, learning and cell recruitments for memory using oscillation.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.