Object discrimination based on a swarm of agents

• US 10,353,052 B2
• Filed: 09/15/2016
• Issued: 07/16/2019
• Est. Priority Date: 09/15/2016
• Status: Active Grant

• Current Assignee: Lawrence Livermore National Security LLC (Government of the United States of America)
• Original Assignee: Lawrence Livermore National Security LLC (Government of the United States of America)
• Inventors: Beer, N. Reginald, Chambers, David, Chandrasekaran, Hema
• Primary Examiner(s): Bythrow, Peter M
• Application Number:
  US15/267,066
• Publication Number:
  US 20180074174A1
• Time in Patent Office: 1,034 Days
• CPC Class Codes: B63G 9/02 Means for protecting vessels against torpedo attack armoured hulls B63B3/10
• CPC Class Codes: B64C 39/024 of the remote controlled vehicle type, i.e. RPV
• CPC Class Codes: B64U 2101/00 UAVs specially adapted for particular uses or applications
• CPC Class Codes: G01S 13/885 for ground probing prospecting or detecting using electromagnetic waves G01V3/12
• CPC Class Codes: G01S 7/411 Identification of targets based on measurements of radar reflectivity G01S7/415 takes precedence
• CPC Class Codes: G01S 7/539 using analysis of echo signal for target characterisation; Target signature; Target cross-section

Abstract

A configuration system for identifying a target configuration of agents for acquiring data is provided. In some embodiments, the configuration system generates sample configurations, each of which has a location within the space for each agent. The configuration system selects as the target configuration a sample configuration that is suitable to acquire data. The configuration system selects the target configuration by, for at least one or more sample configurations, directing that the agents assume the sample configuration and then acquire data when in the sample configuration. The configuration system then stores, for each sample configuration, an indication of suitability for acquiring data based on the sample configuration. The configuration system selects a sample configuration that is deemed to be suitable as the target configuration.

Claims

1. A method performed by a computing system for identifying a target configuration of agents for acquiring data, the method comprising:

generating sample configurations, each sample configuration having a location within a space for each agent; and

selecting as the target configuration a sample configuration that is suitable to acquire data by, for at least one or more sample configurations,

directing that the agents assume the sample configuration and then acquire data when in the sample configuration; and

storing an indication of suitability for acquiring data based on the sample configuration.

2. The method of claim 1 wherein each agent includes a transmitter and a receiver for transmitting a signal and receiving a return signal.

3. The method of claim 2 wherein the transmitted signal and the return signal are selected from a group consisting of an electromagnetic signal, an acoustic signal, and a sonar signal.

4. The method of claim 1 wherein the directing and storing for a sample configuration are performed to acquire first data for a first object of a first object type and are performed to acquire second data for a second object of a second object type and wherein the indication of suitability for the sample configuration is based on how the first data and the second data for the sample configuration discriminate the first object from the second object.

5. The method of claim 4 wherein when an object is detected, directing that the agents assume the target configuration and then acquire target data, wherein the target data is used to identify whether the detected object is of the first object type or the second object type.

6. The method of claim 1 wherein the agents form a communications network and further comprising directing that the agents assume the target configuration and provide communication services from those locations.

7. The method of claim 1 wherein the agents are selected from a group consisting of an unmanned aerial vehicle, unmanned underwater vehicle, unmanned ground vehicle, and unmanned space vehicle.

8. The method of claim 1 wherein the generating of the sample configurations is based on a design-of-experiments technique.

9. The method of claim 8 wherein the design-of-experiments technique is a Latin hypercube technique.

10. The method of claim 8 wherein the design-of-experiments technique generates candidate configurations and the sample configurations are candidate configurations whose locations satisfy a separation criterion.

11. The method of claim 1 wherein the sample configurations are in a three-dimensional space.

12. A method performed by a computing system for determining whether a target object at a target location is of a first object type or a second object type, the method comprising:

generating sample configurations, each sample configuration having a location within a space for each of a plurality of agents;

selecting as a target configuration a sample configuration based on suitability of first sensor data for a first object of a first object type and second sensor data for a second object of a second object type to discriminate the first object from the second object by, for one or more sample configurations,

directing that the agents acquire the first sensor data while in the sample configuration; and

directing that the agents acquire the second sensor data while in the sample configuration;

directing that the agents assume the target configuration relative to a target location and then acquire target sensor data; and

indicating whether the target object is of the first object type or the second object type based on the target sensor data.

13. The method of claim 12 wherein a plurality of sample configurations are selected as target configurations and further comprising:

for each target configuration, directing that the agents assume the target configuration relative to the target location and then acquire target sensor data; and

indicating whether the target object is of the first object type or the second object type based on the target sensor data acquired for each of the target configurations.

14. The method of claim 13 wherein the agents communicate with each other to coordinate selecting of a next target configuration, positioning at the locations of the next target configuration, and acquiring of the target sensor data while in the next target configuration.

15. The method of claim 12 wherein the generating of the sample configurations is based on a design-of-experiments technique.

16. The method of claim 15 wherein the design-of-experiments technique is a Latin hypercube technique.

17. The method of claim 15 wherein the design-of-experiments technique generates candidate configurations and the sample configurations are candidate configurations whose locations satisfy a separation criterion.

18. The method of claim 12 wherein the sensor data is acquired by the agents physically moving to the locations and collecting physical signals.

19. The method of claim 12 wherein the sensor data that would be acquired by the agents physically moving to the locations is estimated by a computer program.

20. The method of claim 19 wherein the computer program simulates transmitting signals from the agents in the target configuration and then receiving return signals at the agents in the target configuration.

21. The method of claim 12 wherein the sensor data is acquired by at least some of the agents transmitting a transmitted signal and at least some of the agents receiving the return signal.

22. The method of claim 21 further comprising determining suitability to discriminate by, for the first sensor data and the second sensor data, converting the return signals into a frequency domain signal, and, for each of a plurality of frequencies, performing a singular value decomposition on the frequency domain signal for that frequency to identify a singular value and then analyzing the singular values derived from the first sensor data and the singular values derived from the second sensor data.

23. The method of claim 21 comprising determining suitability to discriminate based on analysis of singular values derived from the frequencies of the return signals.

24. A method performed by a computing system for determining whether a target object is of a first object type or a second object type, the method comprising:

providing a target configuration having a location for each of a plurality of agents, the target configuration being identified based on suitability of sensor data for the target configuration to discriminate an object of a first object type from an object of a second object type;

directing that the agents assume the target configuration relative to the location of the target object and acquire target sensor data; and

indicating whether the target object is of the first object type or the second object type based on the target sensor data.

25. The method of claim 24 wherein a plurality of target configurations are provided, each target configuration being identified based on suitability of the sensor data for the target configuration to discriminate between objects of different object types and further comprising:

for each target configuration, directing that the agents assume the target configuration relative to the location of the target object and acquire target sensor data; and

indicating the target object is a certain object type based on the target sensor data acquired based on the target configurations.

26. The method of claim 25 wherein the agents communicate with each other to coordinate selecting of a next target configuration, positioning at the locations of the next target configuration, and acquiring of the target sensor data while in the next target configuration.

27. The method of claim 24 wherein the target configuration is identified from sample configurations generated based on a design-of-experiments technique.

28. The method of claim 27 wherein the design-of-experiments technique is a Latin Hypercube technique.

29. The method of claim 27 wherein the design-of-experiments technique generates candidate configurations and the sample configurations are candidate configurations whose locations satisfy a separation criterion.

30. The method of claim 24 wherein the computing system is a distributed computing system with each agent being a node of the distributed computing system.

31. The method of claim 24 wherein the target configuration specifies a characteristic of a sensor for an agent and the directing that that agent assume the target configuration includes directing that agent to use the specified characteristic of the sensor.

32. The method of claim 31 wherein the characteristic is polarization of the sensor.

33. The method of claim 31 wherein the characteristic is frequency bandwidth.

34. The method of claim 31 wherein the characteristic is modality.

35. The method of claim 31 wherein the target configuration specifies multiple characteristics of the sensor.

36. The method of claim 31 wherein the target configuration specifies a characteristic of a sensor for each of a plurality of agents.

37. A method performed by a computing system for establishing a communications network to support communications devices, the method comprising:

generating sample configurations, each sample configuration having a location for each agent; and

selecting a target configuration that is suitable for communications by, for one or more sample configurations;

directing that the agents assume the sample configuration and then communicate with the communications devices and with other agents; and

storing an indication of suitability of the agents in the sample configuration for communications.

38. The method of claim 37 wherein if the agents are not already in the target configuration when the target configuration is selected, directing the agents to assume the target configuration.

39. The method of claim 37 further comprising:

when an agent detects a change in the communications devices, reselecting a target configuration; and

if the agents are not already in the reselected target configuration when it is reselected, directing the agents to assume the reselected target configuration.

40. The method of claim 39 wherein a change in the communications devices may result from a change in the number of communications devices, a change in the locations of the communication, or a change in a signal strength of a communications device.

41. The method of claim 37 wherein the agents are unmanned aerial vehicles (“

UAVs”

).

42. The method of claim 37 wherein the agents are unmanned aerial vehicles (“

UAVs”

) and at least one agent is a non-UAV.

43. The method of claim 42 wherein an agent that is a non-UAV is an unmanned ground vehicle (“

UGV”

).

44. The method of claim 42 wherein an agent that is a non-UAV is an unmanned underwater vehicle (“

UUV”

).

45. The method of claim 37 wherein the agents are heterogeneous.

46. The method of claim 37 wherein the generating of the sample configurations is based on a design-of-experiments technique.

47. The method of claim 37 wherein the design-of-experiments technique is a Latin Hypercube technique.

48. The method of claim 37 wherein the design-of-experiments technique generates candidate configurations and the sample configurations are candidate configurations whose locations satisfy a separation criterion.

49. The method of claim 37 wherein a sample configuration having a characteristic of a sensor for an agent and the directing that that agent assume the sample configuration includes directing that agent to use the specified characteristic of the sensor.

50. The method of claim 49 wherein the sample configuration having a characteristic of a sensor for each of a plurality of agents.

51. One or more computing systems for identifying a target configuration of agents for acquiring data, the one or more computing systems comprising:

one or more processors; and

one or more computer-readable storage mediums storing computer-executable instructions that, when executed by the one or more processors, cause the one or more computing systems to perform;

generating sample configurations, each sample configuration having a location within a space for each agent; and

selecting as the target configuration a sample configuration that is suitable to acquire data by, for at least one or more sample configurations,

directing that the agents assume the sample configuration and then acquire data when in the sample configuration; and

storing an indication of suitability for acquiring data based on the sample configuration.

52. The one or more computing systems of claim 51 wherein each agent includes a transmitter and a receiver for transmitting a signal and receiving a return signal.

53. The one or more computing systems of claim 52 wherein the transmitted signal and the return signal are selected from a group consisting of an electromagnetic signal, an acoustic signal, and a sonar signal.

54. The one or more computing systems of claim 51 wherein the directing and storing for a sample configuration are performed to acquire first data for a first object of a first object type and are performed to acquire second data for a second object of a second object type and wherein the indication of suitability for the sample configuration is based on how the first data and the second data for the sample configuration discriminate the first object from the second object.

55. The one or more computing systems of claim 54 wherein the computer-executable instructions, when executed by the one or more processors, cause the one or more computing systems to perform when an object is detected, directing that the agents assume the target configuration and then acquire target data, wherein the target data is used to identify whether the detected object is of the first object type or the second object type.

56. The one or more computing systems of claim 51 wherein the agents form a communications network and wherein the computer-executable instructions, when executed by the one or more processors, cause the one or more computing systems to perform, when executed by the one or more processors, cause the one or more computing systems to perform directing that the agents assume the target configuration and provide communication services from those locations.

57. The one or more computing systems of claim 51 wherein the agents are selected from a group consisting of an unmanned aerial vehicle, unmanned underwater vehicle, unmanned ground vehicle, and unmanned space vehicle.

58. The one or more computing systems of claim 51 wherein the generating of the sample configurations is based on a design-of-experiments technique.

59. The one or more computing systems of claim 58 wherein the design-of-experiments technique is a Latin hypercube technique.

60. The one or more computing systems of claim 58 wherein the design-of-experiments technique generates candidate configurations and the sample configurations are candidate configurations whose locations satisfy a separation criterion.

61. The one or more computing systems of claim 51 wherein the sample configurations are in a three-dimensional space.

62. One or more computing systems for determining whether a target object at a target location is of a first object type or a second object type, the one or more computing systems comprising:

one or more processors; and

one or more computer-readable storage mediums storing computer-executable instructions that, when executed by the one or more processors, cause the one or more computing systems to perform;

generating sample configurations, each sample configuration having a location within a space for each of a plurality of agents;

selecting as a target configuration a sample configuration based on suitability of first sensor data for a first object of a first object type and second sensor data for a second object of a second object type to discriminate the first object from the second object by, for one or more sample configurations,

directing that the agents acquire the first sensor data while in the sample configuration; and

directing that the agents acquire the second sensor data while in the sample configuration;

directing that the agents assume the target configuration relative to a target location and then acquire target sensor data; and

indicating whether the target object is of the first object type or the second object type based on the target sensor data.

63. The one or more computing systems of claim 62 wherein a plurality of sample configurations are selected as target configurations and wherein the computer-executable instructions, when executed by the one or more processors, cause the one or more computing systems to perform:

for each target configuration, directing that the agents assume the target configuration relative to the target location and then acquire target sensor data; and

indicating whether the target object is of the first object type or the second object type based on the target sensor data acquired for each of the target configurations.

64. The one or more computing systems of claim 63 wherein the agents communicate with each other to coordinate selecting of a next target configuration, positioning at the locations of the next target configuration, and acquiring of the target sensor data while in the next target configuration.

65. The one or more computing systems of claim 62 wherein the generating of the sample configurations is based on a design-of-experiments technique.

66. The one or more computing systems of claim 65 wherein the design-of-experiments technique is a Latin hypercube technique.

67. The one or more computing systems of claim 65 wherein the design-of-experiments technique generates candidate configurations and the sample configurations are candidate configurations whose locations satisfy a separation criterion.

68. The one or more computing systems of claim 62 wherein the sensor data is acquired by the agents physically moving to the locations and collecting physical signals.

69. The one or more computing systems of claim 62 wherein the sensor data that would be acquired by the agents physically moving to the locations is estimated by a computer program.

70. The one or more computing systems of claim 69 wherein the computer program simulates transmitting signals from the agents in the target configuration and then receiving return signals at the agents in the target configuration.

71. The one or more computing systems of claim 62 wherein the sensor data is acquired by at least some of the agents transmitting a transmitted signal and at least some of the agents receiving the return signal.

72. The one or more computing systems of claim 71 wherein the computer-executable instructions, when executed by the one or more processors, cause the one or more computing systems to perform determining suitability to discriminate by, for the first sensor data and the second sensor data, converting the return signals into a frequency domain signal, and, for each of a plurality of frequencies, performing a singular value decomposition on the frequency domain signal for that frequency to identify a singular value and then analyzing the singular values derived from the first sensor data and the singular values derived from the second sensor data.

73. The one or more computing systems of claim 71 wherein the computer-executable instructions, when executed by the one or more processors, cause the one or more computing systems to perform determining suitability to discriminate based on analysis of singular values derived from the frequencies of the return signals.

74. One or more computing systems for determining whether a target object is of a first object type or a second object type, the one or more computing systems comprising:

one or more processors; and

one or more computer-readable storage mediums storing computer-executable instructions that, when executed by the one or more processors, cause the one or more computing systems to perform;

providing a target configuration having a location for each of a plurality of agents, the target configuration being identified based on suitability of sensor data for the target configuration to discriminate an object of a first object type from an object of a second object type;

directing that the agents assume the target configuration relative to the location of the target object and acquire target sensor data; and

indicating whether the target object is of the first object type or the second object type based on the target sensor data.

75. The one or more computing systems of claim 74 wherein a plurality of target configurations are provided, each target configuration being identified based on suitability of the sensor data for the target configuration to discriminate between objects of different object types and wherein the computer-executable instructions, when executed by the one or more processors, cause the one or more computing systems to perform:

for each target configuration, directing that the agents assume the target configuration relative to the location of the target object and acquire target sensor data; and

indicating the target object is a certain object type based on the target sensor data acquired based on the target configurations.

76. The one or more computing systems of claim 75 wherein the agents communicate with each other to coordinate selecting of a next target configuration, positioning at the locations of the next target configuration, and acquiring of the target sensor data while in the next target configuration.

77. The one or more computing systems of claim 74 wherein the target configuration is identified from sample configurations generated based on a design-of-experiments technique.

78. The one or more computing systems of claim 77 wherein the design-of-experiments technique is a Latin Hypercube technique.

79. The one or more computing systems of claim 77 wherein the design-of-experiments technique generates candidate configurations and the sample configurations are candidate configurations whose locations satisfy a separation criterion.

80. The one or more computing systems of claim 74 wherein the computing system is a distributed computing system with each agent being a node of the distributed computing system.

81. The one or more computing systems of claim 74 wherein the target configuration specifies a characteristic of a sensor for an agent and the directing that that agent assume the target configuration includes directing that agent to use the specified characteristic of the sensor.

82. The one or more computing systems of claim 81 wherein the characteristic is polarization of the sensor.

83. The one or more computing systems of claim 81 wherein the characteristic is frequency bandwidth.

84. The one or more computing systems of claim 81 wherein the characteristic is modality.

85. The one or more computing systems of claim 81 wherein the target configuration specifies multiple characteristics of the sensor.

86. The one or more computing systems of claim 81 wherein the target configuration specifies a characteristic of a sensor for each of a plurality of agents.

87. One or more computing systems for establishing a communications network to support communications devices, the one or more computing systems comprising:

one or more processors; and

one or more computer-readable storage mediums storing computer-executable instructions that, when executed by the one or more processors, cause the one or more computing systems to perform;

generating sample configurations, each sample configuration having a location for each agent; and

selecting a target configuration that is suitable for communications by, for one or more sample configurations;

directing that the agents assume the sample configuration and then communicate with the communications devices and with other agents; and

storing an indication of suitability of the agents in the sample configuration for communications.

88. The one or more computing systems of claim 87 wherein the computer-executable instructions, when executed by the one or more processors, cause the one or more computing systems to perform if the agents are not already in the target configuration when the target configuration is selected, directing the agents to assume the target configuration.

89. The one or more computing systems of claim 87 wherein the computer-executable instructions, when executed by the one or more processors, cause the one or more computing systems to perform:

when an agent detects a change in the communications devices, reselecting a target configuration; and

if the agents are not already in the reselected target configuration when it is reselected, directing the agents to assume the reselected target configuration.

90. The one or more computing systems of claim 89 wherein a change in the communications devices may result from a change in the number of communications devices, a change in the locations of the communication, or a change in a signal strength of a communications device.

91. The one or more computing systems of claim 87 wherein the agents are unmanned aerial vehicles (“

UAVs”

).

92. The one or more computing systems of claim 87 wherein the agents are unmanned aerial vehicles (“

UAVs”

) and at least one agent is a non-UAV.

93. The one or more computing systems of claim 92 wherein an agent that is a non-UAV is an unmanned ground vehicle (“

UGV”

).

94. The one or more computing systems of claim 92 wherein an agent that is a non-UAV is an unmanned underwater vehicle (“

UUV”

).

95. The one or more computing systems of claim 87 wherein the agents are heterogeneous.

96. The one or more computing systems of claim 87 wherein the generating of the sample configurations is based on a design-of-experiments technique.

97. The one or more computing systems of claim 87 wherein the design-of-experiments technique is a Latin Hypercube technique.

98. The one or more computing systems of claim 87 wherein the design-of-experiments technique generates candidate configurations and the sample configurations are candidate configurations whose locations satisfy a separation criterion.

99. The one or more computing systems of claim 87 wherein a sample configuration having a characteristic of a sensor for an agent and the directing that that agent assume the sample configuration includes directing that agent to use the specified characteristic of the sensor.

100. The one or more computing systems of claim 99 wherein the sample configuration having a characteristic of a sensor for each of a plurality of agents.