Machine-learning augmented synthetic aperture sonar
An autonomous underwater device includes one or more receiver arrays. Each receiver array includes a plurality of receiver elements, and each receiver element is configured to generate a signal responsive to detecting sound energy in an aquatic environment. The autonomous underwater device also includes one or more processors coupled to the one or more receiver arrays and configured to receive signals from the receiver elements, to generate input data based on the received signals, and to provide the input data to an on-board machine learning model to generate model output data. The model output data includes sonar image data based on the sound energy, a label associated with the sonar image data, or both.
1 . An autonomous underwater device comprising
one or more receiver arrays, each receiver array comprising a plurality of receiver elements, each receiver element configured to generate a signal responsive to detecting sound energy in an aquatic environment;
an on-board navigation control system coupled to a propulsion system and configured to send navigation control signals to the propulsion system to move the autonomous underwater device based on swarm operation data to cause the autonomous underwater device to follow a course that limits sonar system interference between devices of a swarm of devices; and
one or more processors coupled to the one or more receiver arrays and configured to receive signals from the receiver elements, to generate input data based on the received signals, and to provide the input data to an on-board machine learning model to generate model output data, wherein the model output data includes sonar image data based on the sound energy, a label associated with the sonar image data, or both.
2 . The autonomous underwater device of claim 1 , further comprising:
the propulsion system configured to propel the autonomous underwater device within the aquatic environment; and
wherein the navigation control signals are further based on synthetic aperture sonar settings data.
3 . The autonomous underwater device of claim 2 , wherein the one or more processors are configured to generate the synthetic aperture sonar settings data based on second model output of a second on-board machine learning model.
4 . The autonomous underwater device of claim 2 , wherein the synthetic aperture sonar settings data indicate a ping encoding scheme, a ping wavelength, a ping duration, a ping amplitude, a ping interval, ping timing, a navigation course, an along course speed, or a combination thereof.
5 . The autonomous underwater device of claim 2 , wherein the navigation control signals are further based on swarm operation data indicating a position or operation of the autonomous underwater device in the swarm of devices.
6 . The autonomous underwater device of claim 5 , wherein the course is selected to be non-parallel to courses of one or more other devices of the swarm of devices.
7 . The autonomous underwater device of claim 1 , further comprising a communications system configured to receive a communication signal from another device, and wherein the input data is further based on the communication signal.
8 . The autonomous underwater device of claim 7 , further comprising one or more sonar transducers configured to emit sound energy into the aquatic environment via one or more pings, wherein the communication signal includes calibration data indicating measurements of sound energy at the other device responsive to emission of the one or more pings by the one or more sonar transducers.
9 . The autonomous underwater device of claim 7 , wherein the communication signal includes second sonar image data generated by the other device, and wherein the model output data is based on multiple sonar images from multiple devices.
10 . The autonomous underwater device of claim 1 , further comprising one or more sonar transducers configured to emit one or more pings, wherein at least a portion of the sound energy detected by the one or more receiver arrays is derived from the one or more pings.
11 . The autonomous underwater device of claim 1 , wherein the one or more processors are configured to generate, based on the received signals, a plurality of sonar images using two or more distinct sonar image processing techniques, and wherein the on-board machine learning model includes an image generation model trained to generate the sonar image data based on the plurality of sonar images generated using the two or more distinct sonar image processing techniques.
12 . The autonomous underwater device of claim 1 , further comprising a notification system configured to generate an alert responsive to detection, in the sonar image data, of a particular type of object.
13 . A system comprising:
a plurality of autonomous underwater devices configured to operate cooperatively as a swarm, the plurality of autonomous underwater devices comprising:
a first autonomous underwater device comprising:
a propulsion system configured to propel the first autonomous underwater device within an aquatic environment;
an on-board navigation control system coupled to the propulsion system and configured to send navigation control signals to the propulsion system to move the first autonomous underwater device based on swarm operation data to cause the first autonomous underwater device to follow a course that limits sonar system interference between devices of the swarm;
one or more receiver arrays, each receiver array comprising a plurality of receiver elements, each receiver element configured to generate a signal responsive to detecting sound energy in the aquatic environment; and
one or more processors coupled to the one or more receiver arrays and configured to receive signals from the receiver elements, to generate input data based on the received signals, and to provide the input data to a machine learning model associated with the swarm to generate model output data, wherein the model output data includes sonar image data based on the sound energy, a label associated with the sonar image data, or both; and
a second autonomous underwater device comprising one or more sonar transducers configured to emit one or more pings into the aquatic environment.
14 . The system of claim 13 , wherein the swarm further comprises at least one stationary device configured to facilitate navigation, communications, or both, among the plurality of autonomous underwater devices of the swarm.
15 . The system of claim 13 , wherein the one or more receiver arrays of the first autonomous underwater device are configured to detect the one or more pings emitted by the one or more sonar transducers of the second autonomous underwater device.
16 . The system of claim 13 , wherein the first autonomous underwater device further comprises a first synthetic aperture sonar system that includes the one or more receiver arrays, the one or more processors, and at least one sonar transducer, wherein the second autonomous underwater device further comprises a second synthetic aperture sonar system that includes the one or more sonar transducers, at least one receiver array, and at least one processor, and wherein the first synthetic aperture sonar system and the second synthetic aperture sonar system are configured to operate cooperatively to generate the sonar image data.
17 . The system of claim 16 , wherein the machine learning model associated with the swarm comprises an on-board machine learning model of the first synthetic aperture sonar system and wherein the input data to the on-board machine learning model of the first synthetic aperture sonar system is further based on communication signals received from the second autonomous underwater device.
18 . The system of claim 13 , wherein the one or more processors of the first autonomous underwater device are further configured to generate synthetic aperture sonar settings data based on second model output of a second on-board machine learning model.
19 . The system of claim 13 , wherein the first autonomous underwater device further comprises a communications system configured to transmit, to the second autonomous underwater device, a communication signal including calibration data based on measurements of sound energy of one or more pings detected at the one or more receiver arrays of the first autonomous underwater device.
20 . The system of claim 13 , wherein the first autonomous underwater device further comprises a notification system configured to generate an alert responsive to detection, in the sonar image data, of a particular type of object.
21 . The system of claim 13 , wherein the navigation control signals are further based on synthetic aperture sonar settings data.
22 . A method comprising:
emitting, from a first autonomous underwater device, sound energy in an aquatic environment via one or more pings;
receiving, at the first autonomous underwater device from a second autonomous underwater device, an electromagnetic waveform encoding a communication signal;
based on the communication signal, determining synthetic aperture sonar settings data to limit interferences between synthetic aperture sonar systems of devices of a swarm of autonomous underwater devices;
detecting, at the first autonomous underwater device, sonar returns based on the one or more pings;
generating input data based on the sonar returns and the communication signal; and
providing the input data to a machine learning model on-board the first autonomous underwater device to generate model output data, wherein the model output data includes sonar image data based on the sonar returns, a label associated with the sonar image data, or both.
23 . The method of claim 22 , further comprising processing the sonar returns based on the synthetic aperture sonar settings data to generate the input data.
24 . The method of claim 22 , wherein the one or more pings include at least a first ping and a second ping, wherein the communication signal is received responsive to the first ping, and further comprising setting parameters of the second ping responsive to the communication signal, wherein the sonar returns are based on the second ping.
25 . The method of claim 22 , wherein the communication signal include second sonar image data, and further comprising determining first sonar image data based on the sonar returns, and wherein the sonar image data of the model output data is based on the first sonar image data and the second sonar image data.
26 . The method of claim 22 , further comprising sending the sonar image data, the label, or both, to another autonomous underwater device of the swarm of autonomous underwater devices, the swarm of autonomous underwater devices including the first autonomous underwater device and the second autonomous underwater device.
27 . A method comprising:
obtaining position data indicating locations of one or more underwater devices of a swarm of devices in an aquatic environment;
determining navigation data for at least one device of the swarm of devices based on output of a tactical machine learning model that is trained to predict adversarial actions during an engagement between an adversarial agent and one or more friendly agents, to select responses to detected adversarial actions, or both; and
sending the navigation data to a navigation control system onboard the at least one device to cause the at least one device to, in cooperation with one or more other devices of the swarm of devices, gather synthetic aperture sonar data in a particular area by selecting one or more synthetic aperture sonar settings to limit interferences between synthetic aperture sonar systems of devices of the swarm of devices.
28 . The method of claim 27 , wherein the one or more synthetic aperture sonar settings includes a ping encoding scheme, a ping wavelength, a ping duration, a ping amplitude, a ping interval, ping timing, a navigation course, an along course speed, or a combination thereof.
29 . The method of claim 27 , wherein the navigation data is determined at one or more processors onboard a first autonomous underwater device, and wherein the navigation data is sent to a navigation control system onboard the first autonomous underwater device, wherein the navigation control system onboard the first autonomous underwater device is configured to send navigation control signals to a propulsion system of the first autonomous underwater device to cause the first autonomous underwater device to move to the particular area.