Classifications

• • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
  • F41—WEAPONS
  • F41G—WEAPON SIGHTS; AIMING
  • F41G3/00—Aiming or laying means
  • F41G3/14—Indirect aiming means
  • F41G3/147—Indirect aiming means based on detection of a firing weapon
• • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
  • F41—WEAPONS
  • F41H—ARMOUR; ARMOURED TURRETS; ARMOURED OR ARMED VEHICLES; MEANS OF ATTACK OR DEFENCE, e.g. CAMOUFLAGE, IN GENERAL
  • F41H11/00—Defence installations; Defence devices
• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
  • G01C23/00—Combined instruments indicating more than one navigational value, e.g. for aircraft; Combined measuring devices for measuring two or more variables of movement, e.g. distance, speed or acceleration
• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
  • G01S3/00—Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received
  • G01S3/78—Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received using electromagnetic waves other than radio waves
  • G01S3/782—Systems for determining direction or deviation from predetermined direction
  • G01S3/783—Systems for determining direction or deviation from predetermined direction using amplitude comparison of signals derived from static detectors or detector systems
  • G01S3/784—Systems for determining direction or deviation from predetermined direction using amplitude comparison of signals derived from static detectors or detector systems using a mosaic of detectors
• • G—PHYSICS
  • G01—MEASURING; TESTING
  • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
  • G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
  • G01S5/18—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using ultrasonic, sonic, or infrasonic waves
  • G01S5/22—Position of source determined by co-ordinating a plurality of position lines defined by path-difference measurements

Definitions

• a system operative to estimate a location of a projectile launch, the system comprising one or more processing circuitries configured to obtain, from an imaging device, data D ⁇ ash informative of an optical event associated with the projectile launch, obtain, from acoustic sensors, acoustic data, wherein the acoustic data includes data Dshockwaves informative of shock waves generated by a motion of the projectile, and use Dfiash anc l D shockwaves to estimate the location of the projectile launch.
• the device can optionally comprise one or more of features (i) to (xxix) below, in any technically possible combination or permutation: i. the system is configured to estimate the two-dimensional or the three-dimensional location of the projectile launch; ii. the system is configured to use D shockwaves to estimate a time t shockwaves at which at least some of the shock waves are sensed by the acoustic sensors, and use tshockwaves to estimate the location of the projectile launch; iii.
• the system uses data D ⁇ ash to obtain a time tf iash which is an estimate of a time t 0 at which the projectile has been launched, and vi. the system is configured to use the time tf iash to estimate the location of the projectile launch; vii. the system is configured to use data Df iash to estimate a direction DRf iash at which the projectile has been launched, and use the direction DRf iash to estimate the location of the projectile launch; viii.
• the system is operative to determine the location of the projectile launch based on f in at least one of (i) or (ii) is met: (i) a scenario in which a blast generated by the projectile launch is sensed by the acoustic sensors with a frequency which is below a detection threshold; (ii) a scenario in which a blast generated by the projectile launch is sensed by the acoustic sensors with a signal to noise ratio which is below a detection threshold; ix. the detection threshold is equal to 1 kHz; x.
• the system is configured to use a model to estimate, for a plurality of candidate locations of the projectile launch, a time t shocliwaves and a direction DR shocliwaves at which the shock waves are sensed by the acoustic sensors, and use: determined for the plurality of candidate locations, to determine a given candidate location of the projectile which meets an optimization criterion; xv.
• the one or more processing circuitries are configured to implement a first machine learning model and a second machine learning model, wherein the first machine learning model has been trained to detect, in acoustic data, an event corresponding to shock waves generated by a motion of a projectile of a first type, and the second machine learning model has been trained to detect, in acoustic data, an event corresponding to shock waves generated by a motion of a projectile of a second type, different from the first type; xix. the optical event is a muzzle flash associated with the projectile launch; xx. the projectile is a supersonic projectile; xxi. the system comprises an electronic card embedding the imaging device, the acoustic sensors, and the one or more processing circuitries; xxii.
• the system comprises the imaging device and/or the acoustic sensors; xxiii. the acoustic data includes data D biast informative of a blast generated by the projectile launch, wherein the system is configured to use D tash , D shockwaves and Dbiast t° estimate the location of the projectile launch, and a firing line direction of the projectile; xxiv. use D tash and D biast to estimate the location P launch °f the projectile launch, xxv.
• the system is configured to use a model to estimate, for the location P launch °f the projectile launch and one or more candidate firing line directions of the projectile, a time t shockwaves an d a direction DR shockwaves at which the shock waves are sensed by the acoustic sensors, use to determine a time and a direction DR ⁇ ; hnr " k -waves at which at least some of the shock waves are sensed by the acoustic sensors, use DR shockwaves , t shockwaves , and DR shockwaves and t shockwaves determined for the one or more candidate firing line directions of the projectile, to estimate the firing line direction of the projectile; xxvi.
• the system is configured to perform a determination of a content of the acoustic data, wherein said determination enables to identify whether the acoustic data includes (i) or (ii): (i) data D shockwaves informative of shock waves generated by a motion of the projectile shock waves, wherein a blast generated by the projectile launch cannot be identified in the acoustic data, (ii) data D shockwaves informative of shock waves generated by a motion of the projectile shock waves and data D b iast informative of a blast generated by the projectile launch, and trigger an estimation method of the location of the projectile launch which depends on this determination; xxvii.
• the system is operative to be mounted on a moving platform, wherein the system comprises, or is operatively coupled to a sensor operative to provide data D inertiai informative of a direction of the moving platform or of the system over time; xxviii. the system is configured to use D ftash , D shockwaves and D inerticd to estimate the location of the projectile launch; and xxix. the system is mounted on at least one of: a military vehicle, a tank, an aircraft.
• a platform comprising the system.
• a method to estimate a location of a projectile launch comprising, by one or more processing circuitries: obtaining, from an imaging device, data Df iash informative of an optical event associated with the projectile launch, obtaining, from acoustic sensors, acoustic data, wherein the acoustic data includes data D shockwaves informative of shock waves generated by a motion of the projectile, and using Df i asll and D shockwaves to estimate the location of the projectile launch.
• the method can include the additional features described above with respect to the system.
• a non-transitory computer readable medium comprising instructions that, when executed by one or more processing circuitries, cause the one or more processing circuitries to perform the operations of the method.
• a system operative to estimate a location of projectile launch, the system comprising one or more processing circuitries configured to: obtain, from an imaging device, data h informative of an optical event associated with the projectile launch, obtain, from acoustic sensors, acoustic data, wherein the acoustic data includes: o data informative of shock waves generated by a motion of the projectile, and o data informative of a blast generated by the projectile launch, use to estimate. o the location of the projectile launch, and o a firing line direction of the projectile.
• system can optionally comprise one or more of features (xxx) to (xxxii) below (and, in some embodiments, features i to xxix described above): xxx.
• the system is configured to use a model which estimates, for a given location of the projectile launch and a given firing line direction of the projectile, a time a r
• the one or more processing circuitries are configured to implement at least one machine learning model, wherein the one or more processing circuitries are configured to feed the acoustic data, or data informative thereof, to the machine learning model and use the machine learning model to identify, in the acoustic data, data s informative of shock waves generated by a motion of the projectile and data D informative of a blast generated by the projectile launch; xxxii.
• a method to estimate a location of a projectile launch comprising, by one or more processing circuitries: obtaining, from an imaging device, data Df iash informative of an optical event associated with the projectile launch, obtaining, from acoustic sensors, acoustic data, wherein the acoustic data includes data D acoustic i mpac t informative of a sound generated by an impact of the projectile, and
• the method can include the additional features described above with respect to the system.
• a method to estimate a location of a projectile launch comprising, by one or more processing circuitries: obtaining, from acoustic sensors, acoustic data, wherein the acoustic data includes data D acoustic i mpac t informative of a sound generated by an impact of the projectile, and data D biast informative of a blast generated by the projectile launch,
• the method can include the additional features described above with respect to the system.
• a non-transitory computer readable medium comprising instructions that, when executed by one or more processing circuitries, cause the one or more processing circuitries to perform the operations of the method.
• a system operative to estimate a location of projectile launch, the system comprising one or more processing circuitries configured to: obtain, from an imaging device, data D ⁇ ash informative of an optical event associated with the projectile launch, obtain, from acoustic sensors, acoustic data; perform a determination of a content of the acoustic data, wherein: o when the determination indicates that the acoustic data includes data Dshockwaves informative of shock waves generated by a motion of the projectile, but a blast associated with the projectile launch cannot be identified in the acoustic data, use Df iash and D shocliwaves to estimate the location of the projectile launch; o when the determination indicates that the acoustic data includes data Dacoustic impact informative of a sound generated by an impact of the projectile, use data D fiash and data D acoustic impact to estimate the location of the projectile launch; o when the determination indicates that the acoustic
• the projectile when the determination indicates that the acoustic data D shockwaves informative of shock waves generated by a motion of the projectile the projectile and data D biast informative of a blast generated by the projectile launch, use D ⁇ ash , D shockwaves and D biast to estimate the location of the projectile launch and a firing line direction of the projectile.
• a method to estimate a location of projectile launch comprising, by one or more processing circuitries: obtaining, from an imaging device, data Df iash informative of an optical event associated with the projectile launch, obtaining, from acoustic sensors, acoustic data; performing a determination of a content of the acoustic data, wherein: o when the determination indicates that the acoustic data includes data Dshockwaves informative of shock waves generated by a motion of the projectile, but a blast associated with the projectile launch cannot be identified in the acoustic data, use D ⁇ ash and D shocliwaves to estimate the location of the projectile launch; o when the determination indicates that the acoustic data includes data Dacoustic impact informative of a sound generated by an impact of the projectile, use data D fiash and data D acoustic impact to estimate the location of the projectile launch; o when the determination indicates that the acoustic
• the method can include the additional features described above with respect to the system.
• a non-transitory computer readable medium comprising instructions that, when executed by one or more processing circuitries, cause the one or more processing circuitries to perform the operations of the method.
• the proposed solution allows determining the location of a projectile launch in adverse conditions (in particular, in conditions which include background noise).
• the proposed solution allows determining the location of a projectile launch in conditions in which at least some of the prior art systems are inoperative.
• the proposed solution provides a system for determining the location of a projectile launch which can be mounted on a moving platform.
• localization of the projectile launch can be performed while the moving platform is in motion.
• the proposed solution allows determining accurately the firing direction of the projectile (in contradiction to prior art methods which could not calculate the firing direction).
• Fig. 1 illustrates an embodiment of a system which can be used to estimate the location of a projectile launch
• Fig. 2A illustrates a non-limitative example of a launch of a projectile from an offensive system
• - Fig. 2D illustrates a non-limitative example of a reference pattern of shock waves
• Fig. 2E illustrates an embodiment of a method of training a machine learning model to detect shock waves in acoustic data
• Fig. 2F illustrates schematically the time flow of the events which occur in the method of Fig. 2A;
• Fig. 2G illustrates an embodiment of a method of determining the direction at which at least some of the shock waves are sensed by the acoustic sensors
• FIG. 2H illustrates operations which can be performed to estimate the location of the projectile launch in the method of Fig. 2A;
• Fig. 21 illustrates a model which can be used to to estimate the location of the projectile launch in the method of Fig. 2A;
• Fig. 3A illustrates another embodiment of a method of estimating the location of a projectile launch and the firing line direction of the projectile
• Fig. 3B illustrates a non-limitative example of identification of a blast within acoustic data by a trained machine learning model
• Fig. 3C illustrates an embodiment of a method of selecting an estimation process of the location of the projectile launch based on the content of the acoustic data
• FIG. 3D illustrates an embodiment of a method of determining the direction at which the blast generated by the projectile launch is sensed by the acoustic sensors
• Fig. 3E illustrates a particular implementation of operations of the method of Fig. 3A;
• Fig. 4A illustrates a non-limitative example of a launch of a projectile from an offensive system, in which the projectile hits a platform on which the system is mounted;
• Fig. 4B illustrates a non-limitative example of a launch of a projectile from an offensive system, in which the projectile hits an obstacle located in the vicinity of the platform on which the system is mounted;
• FIG. 4C illustrates an embodiment of a method of estimating the location of a projectile launch, which is operative in the context of Figs. 4A and 4B;
• Fig. 4D illustrates an embodiment of a method of selecting an estimation process of the location of the projectile launch based on the content of the acoustic data
• Fig. 4E illustrates a non-limitative example of identification of a projectile hit within acoustic data by a trained machine learning model
• Fig. 4F illustrates an embodiment of a method of training a machine learning model to detect an impact of a projectile in acoustic data
• Fig. 4G illustrates schematically the time flow of the events which occur in the method of Fig. 4C;
• FIG. 4H illustrates another embodiment of a method of estimating the location of a projectile launch
• launch of a projectile is known in the art, and includes the launch, projection, ejection, or firing of the projectile.
• Non-limitative examples of projectiles include e.g., a bullet, a rocket, a missile, a shell.
• the projectile can be a supersonic projectile (ammunition, bullet, etc. - these examples are not limitative), or a subsonic projectile (a rocket, a missile, a shell - these examples are not limitative).
• the projectile can be powered or not, and that the projectile can be guided or unguided.
• Fig. 1 is a schematic representation of an embodiment of a system 100.
• system 100 can be used to determine the location of the launch of a projectile 105.
• the location of the launch of the projectile is the location of the device (launcher - such as a weapon) from which the projectile has been launched.
• system 100 can be used also to determine additional (or alternative) data, such as data informative of the path of the projectile, impact point of the projectile, firing line direction of the projectile, etc.
• the system 100 can be mounted on a mobile platform (mobile system).
• the system 100 can be mounted on a vehicle traveling on land (e.g., a car, a truck, etc.), at sea (e.g., a ship), in the air (e.g., a helicopter, an aircraft, etc.).
• the imaging device 110 can include a detector that utilizes potassium (K) doublet emission lines (e.g., ⁇ 760nm) or Sodium (N) emission lines detection with contrast to adjacent optical spectrum narrow bands which lack this emission from flash, plume, and fire radiance; imaging device 110 can include a detector that utilizes the contrast between "Red-band” (within MWIR 4.4um - 4.9um spectral band) and other MWIR subbands as evident in flash, plume, and fire radiance; the imaging device 110 can include a detector that utilizes solar blind UV (SBUV) emission from flash, plume and/or fire radiance - with contrast to natural and/or artificial scene background which lack significant radiance in the SBUV band; the imaging device 110 can include a detector module that utilizes LWIR spect
• the system 100 includes (or is operatively coupled to) an array of acoustic sensors, including a plurality of acoustic sensors 120.
• the array of acoustic sensors 120 can include an array of at least two microphones (or more).
• the acoustic sensors 120 are configured to record acoustic signals from the same scene captured by the imaging device 110 (or from a scene which at least partially overlaps with the scene captured by the imaging device 110).
• the acoustic sensors 120 of the array are located at a similar position (each acoustic sensor can be located in the vicinity of the other acoustic sensors of the array).
• the acoustic sensors 120 can include (or are operatively coupled to) a processing circuitry, operative to process acoustic signal(s) provided by the acoustic sensors 120, as a result of their detection.
• This processing circuitry can correspond to the processing circuitry 130 described hereinafter, or to a different processing circuitry.
• system 100 includes a sensor 150 operative to provide data informative of a direction over time (heading) of the system 100 and/or of a platform on which the system 100 is mounted.
• the sensor 150 can include for example at least one of: a GPS sensor, one or more accelerometers, a gyroscope, an IMU (Inertial Measurement Unit).
• the senor 150 can provide additional data, such as data informative of a position and/or a velocity and/or an acceleration of the system 100 and/or of a platform on which the system 100 is mounted.
• the system 100 further includes at least one (or more) processing circuitry 130, comprising one or more processors and one or more memories.
• the processing circuitry 130 is operatively coupled to the imaging device 110 and/or to the acoustic sensors 120. In particular, it can process signals (optical signals) provided by the imaging device 110 and/or signals (acoustic signals) provided by the acoustic sensors 120.
• the machine learning model 135 can include a neural network (NN). In some embodiments, the machine learning model 135 can include a deep neural network (DNN).
• NN neural network
• DNN deep neural network
• the processor can execute several computer-readable instructions implemented on a computer-readable memory comprised in the processing circuitry 130, wherein execution of the computer-readable instructions enables data processing by the machine learning model 135, such as processing of acoustic data, in order to detect shock waves generated by the motion of the projectile.
• the machine learning model 135 is able to detect, in an acoustic signal, a blast (muzzle blast) generated by the launch of the projectile. In some embodiments, the machine learning model 135 is able to detect, in an acoustic signal, both shock waves generated by the motion of the projectile, and the blast generated by the launch of the projectile.
• the machine learning model 135 is able to detect, in an acoustic signal, a sound generated by an impact (hit) of a projectile.
• a different machine learning model can be used for detecting the shock waves, the blast, and the impact of the projectile.
• the same machine learning model can be trained to detect two of these events (in practice, when there is an impact of the projectile, the shock waves are not sensed by the acoustic sensors), or all of these events.
• the imaging device 110 and the acoustic sensor(s) 120 are located in close vicinity (e.g., on the same platform).
• the imaging device 110 and the acoustic sensor(s) 120 are mounted on the same electronic circuit (e.g., PCB).
• the imaging device 110, the acoustic sensor(s) 120 and also the processing circuitry 130 are mounted on the same electronic circuit (e.g., PCB).
• the imaging device 110 and the acoustic sensor 120 have a high sampling rate (e.g., of at least 64.000 samples per sec). This is not limitative.
• the imaging device 110 is mounted on a platform, and the acoustic sensors 120 are not mounted on the same platform, but rather in the vicinity of this platform.
• Fig. 2A illustrates a launch of a projectile 200 from an offensive system 205 (e.g., a weapon, a launcher, etc.).
• an offensive system 205 e.g., a weapon, a launcher, etc.
• the launch of the projectile 200 generates a muzzle flash 210 (optical event).
• the launch of the projectile 200 can also induce the generation of a blast (launch/detonation blast), which is an acoustic event.
• the blast includes acoustic waves, schematically represented as (spherical/circular) wavefront 220.
• a velocity of the projectile 200 can be higher than the sound velocity in the air (at least during part of its motion, or during all of its motion up to the impact point).
• the projectile 200 can be a supersonic projectile.
• shock waves which correspond to a type of propagating disturbance that moves faster than the local speed of sound in the air. These shock waves generate a sound in the air. Shock waves are schematically illustrated as reference 225 in Fig. 2A. They propagate along the path of the projectile 200 (also called firing direction or firing line direction).
• the blast cannot be identified in the acoustic data recorded by the acoustic sensors 120.
• Fig. 2B describes a method of estimating the location of a projectile launch, in particular in adverse conditions as described with reference to Fig. 2A.
• Fig. 2B can use the system 100 as described with reference to Fig. 1. Assume that a projectile is launched from a certain location. For example, a sniper shoots a bullet using a rifle.
• the method includes obtaining (operation 230) data Df iash informative of an optical event associated with the projectile launch.
• data D ⁇ ash is obtained by the processing circuitry 130 from the imaging device 110.
• Df iash can correspond to the output of the imaging device 110, or to a signal derived from the output of the imaging device 110 (for example, the output of the imaging device 110 can undergo some intermediate processing (e.g., removal of noise, etc.)) before it is sent to the processing circuitry 130.
• a muzzle flash (optical event) is generated, which is detected by the imaging device 110.
• a signal informative of the muzzle flash detected by the imaging device 110 can be obtained by the processing circuitry 130 from the imaging device 110.
• the signal can include e.g., a distribution of pixel intensities and/or a sequence of images acquired by the imaging device 110.
• the projectile is launched at time t 0 .
• shock waves are generated along the path of the projectile.
• the method further includes obtaining (operation 240) acoustic data from the array of acoustic sensors 120.
• acoustic data Assume a scenario in which the motion of the projectile generates shock waves, which are sensed by the acoustic sensors 120. When the projectile passes the acoustic sensors 120, the shock waves propagate towards the acoustic sensors 120 and therefore can be sensed by the acoustic sensors 120. This is visible in the non- limitative example of Fig. 2A, in which the shock waves 225 are sensed by the acoustic sensors 120 when the projectile 200 passes the acoustic sensors 120.
• the acoustic data provided by the acoustic sensors 120 include data Dshockwaves informative of shock waves generated by a motion of the projectile. Methods for identifying shock waves within the acoustic data will be provided hereinafter.
• a blast generated by the projectile launch is sensed by the acoustic sensors 120 with a frequency below a detection threshold (and/or with a signal to noise ratio below a detection threshold).
• the blast cannot be identified in the acoustic data provided by the acoustic sensors 120.
• the blast cannot be used to determine a direction and/or a location of the projectile launch using the methods described in the prior art.
• Operation 240 can include identifying the shock waves in the acoustic data.
• a database stores a plurality of acoustic patterns which are characteristics of shock waves generated by projectiles of interest (e.g., bullets, etc.). Note that a typical pattern (time signal) of shock waves (this example is not limitative) is visible in reference 279 of Fig. 2C.
• the processing circuitry 130 performs a comparison between these patterns and the acoustic data provided by the acoustic sensors 120 in order to identify the shock waves in the acoustic data.
• the database stores the type of projectile associated with each acoustic pattern. Therefore, this comparison can be used to identify the type of the projectile.
• a machine learning model (see reference 135 in Fig. 1) can be used to identify the shock waves. This is illustrated in the non-limitative example of Fig. 2D.
• the machine learning model 135 has been trained to detect, in an acoustic signal, presence of shock waves within the acoustic signal.
• the machine learning model 135 is therefore used to identify, in the acoustic data, an event corresponding to shock waves sensed by the acoustic sensors 120.
• the machine learning model 135 can therefore output data D shocliwaves informative of shock waves generated by a motion of the projectile.
• data D shockwaves includes the portion of the acoustic data (acoustic signal over time) corresponding to the sensing of the shock waves by the acoustic sensors 120.
• data D shockwaves can include the time t shockwaves at which the acoustic sensors 120 have sensed the shock waves.
• Fig. 2E describes a method of training the machine learning model 135 to detect shock waves in acoustic data.
• the method can be performed by a processing circuitry (such as processing circuitry 130).
• the method includes obtaining (operation 290) a training set comprising a plurality of acoustic data.
• Some of the acoustic data includes an acoustic event corresponding to shock waves generated by the supersonic motion of a projectile (“positive samples”).
• the training set is then used (operation 291) to train the machine learning model 135 to identify, based on an input including acoustic data, an acoustic event corresponding to shock waves generated by the supersonic motion of a projectile.
• the training can rely on techniques such as Backpropagation. This is however not limitative.
• the training set includes acoustic events corresponding to shock waves generated by the supersonic motion of different types of projectiles (e.g., different calibers, different weights, different shapes, etc.).
• the machine learning model 135 is trained to detect the shock waves in the acoustic data, and to determine the type of projectile.
• the training set includes: acoustic data including events corresponding to shock waves generated by the supersonic motion of different types of projectiles, wherein a label indicates presence of the shock waves (and time location in the acoustic data) and the type of projectile which generated these shock waves (the label can be provided by an operator who annotates the data); acoustic data which do not include acoustic events corresponding to shock waves generated by the supersonic motion of a projectile.
• a label indicates that these acoustic data do not include any shock waves generated by the supersonic motion of a projectile.
• the training set is then used to train the machine learning model 135 to both identify the shock waves, and to determine the type of projectile which generated these shock waves.
• a plurality of machine learning models is used. Each machine learning model is trained to detect an acoustic event corresponding to shock waves generated by the supersonic motion of a different type of projectile.
• the first machine learning model is trained to detect an acoustic event corresponding to shock waves generated by the supersonic motion of a bullet of a first caliber
• the second machine learning model is trained to detect an acoustic event corresponding to shock waves generated by the supersonic motion of a bullet of a second caliber (different from the first caliber), etc.
• this set of machine learning models can be used as explained hereinafter.
• D shocliwaves can be used also to determine the direction DR shocliwaves (see Fig. 2A) at which at least some of the shock waves are sensed by the acoustic sensors 120, as explained with reference to the method of Fig. 2G.
• the acoustic sensors 120 sense the shock waves with a different delay between them. For example, for a direction of the shock waves corresponding to an angle of degrees, the delay between the first acoustic sensor and the second sensor is equal to tq, for a direction of the shock waves corresponding to an angle of ? 2 degrees, the delay between the first acoustic sensor and the second sensor is equal to At 2 , etc.
• a corresponding delay can be stored between the sensors (delay At x between sensor Si and sensor S2, delay At 2 between sensor S2 and sensor S3, etc).
• the different time delays for the different candidate direction of the shock waves can be determined in advance and stored in a database.
• the method further includes, for a candidate value of the direction DR shocliwaves , multiplying (operation 271), for each given acoustic sensor of the array: a portion of the acoustic data provided by the given acoustic sensor (in particular, the portion D shockwaves of the acoustic data corresponding to the shock waves), with the reference acoustic pattern delayed by the time delay corresponding to this candidate value of the direction DR shockwaves .
• a signal is therefore obtained for each acoustic sensor.
• An aggregated signal can be obtained by summing the different signals obtained for the different acoustic sensors of the array (operation 272).
• This process can be repeated for a plurality of different candidate values of the direction DR shockwaves (operation 273).
• the candidate value for which the aggregated signal is the strongest can be selected as the estimate of the direction DR shockwaves (operation 274).
• the method of Fig. 2B further includes using (operation 250) D ⁇ ash and Dshockwaves to estimate the location of the projectile launch.
• the two-dimensional or the three-dimensional location of the projectile launch is estimated.
• Fig. 2B can be performed with a system 100 mounted on a moving platform.
• Operation 250 can involve using: time tf cLsh- which is an estimate of the time t 0 of the projectile launch (which can be determined using Df iash ) the direction DRf iash , which is an estimate of the direction of the projectile launch (which can be determined using Df iash ) a modelled kinematic behavior of the projectile.
• the modelled kinematic behavior takes into account various parameters which enable describing the flight of the projectile, such as drag, mass of the projectile, muzzle velocity, etc. Note that for parameters which are unknown, estimated values can be used, which can be updated using an iterative optimization process; and
• operation 250 comprises using a model operative to estimate, for a given location of the projectile launch, a time t shockwaves an d a direction DR shockwaves at which the shock waves are sensed by the acoustic sensors.
• operation 250 can include (see Fig. 2G) using (operation 292) the model to estimate, for each of one or more candidate locations of the projectile launch (which can be initialized with a guess value), a time t shockwaves ar
• the model can rely on a physical model of shock waves propagation: since all parameters of the flight of the projectile are known or estimated, the model can predict the time and direction at which the shock waves will be sensed by the acoustic sensors.
• the method can include (operation 293) using D shockwaves to determine: a direction DR shockwaves at which at least some of the shock waves are sensed by the acoustic sensors 120 (see a corresponding method in Fig. 2F); a time t sh0C kwaves at which at least some of the shock waves are sensed by the acoustic sensors 120.
• the method can further include using (operation 294) DR shockwaves , t shockwaves , DRshockwaves and tshockwaves to estimate the location Pi aunch of the projectile launch.
• operation 294 DR shockwaves , t shockwaves , DRshockwaves and tshockwaves to estimate the location Pi aunch of the projectile launch.
• an iterative process of optimization can be used, as explained hereinafter.
• DR shockwaves and t shockwaves are determined for a plurality of candidate locations of the projectile launch using the model, and D are used to determine a given candidate location of the projectile which meets an optimization criterion.
• Fig. 2H illustrates a particular method of implementing operation 250, in order to estimate the location of the projectile launch. Note that this method is only an example and is not limitative.
• the method includes obtaining: a modelled kinematic behavior of the projectile; an estimate Piaunch °f the location of the projectile launch; an estimate FRD of a firing line direction of the projectile; a time t ⁇ ash , which is an (accurate) estimate of a time t 0 at which the projectile has been launched (t ⁇ iash is determined using data Df iash ) a direction DR ⁇ ash , which is an estimate of the direction of the projectile launch (DRf iash is determined using data Df iash ).
• Piaunch i s initially unknown and can be initialized with a guess or a random value.
• FRD is unknown and can be estimated with a guess or a random value.
• the firing direction (see angle ⁇ p) can be (initially) considered as small and can be neglected.
• the method can include using a cost function which is informative of: a difference between t shockwaves aa d t shockwaves? aa d a difference between D Rshockwaves aa d D Rshockwaves •
• the method can include attempting to minimize the cost function for different candidate values of Piaunch located along the direction DRf iash .
• Piaunch which minimizes the cost function can be used as the estimate of the location of the projectile launch.
• the method can include using D shocliwaves to determine a time t S hockwaves an d a direction DR shocliwaves at which at least some of the shock waves are sensed by the acoustic sensors.
• the acoustic data provided by the acoustic sensors include not only data D shocliwaves informative of shock waves generated by a motion of the projectile, but also data D biast informative of the blast generated by the projectile launch.
• the blast generated by the projectile launch is sensed by the acoustic sensors 120 with a frequency above the detection threshold (and with a signal to noise ratio above the detection threshold).
• the method of Fig. 3A includes obtaining (operation 330) data Df iash informative of an optical event associated with the projectile launch.
• data Df iash is obtained by the processing circuitry 130 from the imaging device 110.
• Operation 330 is similar to operation 230 and is therefore not detailed again.
• the method of Fig. 3A further includes (operation 340) obtaining acoustic data from the acoustic sensors 120.
• the acoustic data includes both data Dshockwaves informative of shock waves generated by a motion of the projectile and data D b iast informative of a blast generated by the projectile launch.
• DR biast can be also used.
• DR biast can be used instead of DR ⁇ ash to estimate the direction of the projectile launch.
• a correlation between DR biast and DR ⁇ ash can be performed to validate that the blast and the flash originates from the same projectile launch.
• Fig. 3A can be performed with a system 100 mounted on a moving platform.
• the location of the projectile launch can be determined using D f h and D (using e.g., the method described in WO 2006/096208), or using ⁇ and s •
• a plurality of candidate firing line directions are tested, and the candidate firing line direction which meets an optimization criterion is selected as the estimate.
• the method can include using a cost function which is informative of: a difference between a ⁇ d a difference between
• the method can include attempting to minimize the cost function for different candidate values of the firing line direction FDR.
• the launch of the projectile 400 induces a muzzle flash 410 (optical event).
• the launch of the projectile 400 can also induce the generation of a blast (launch/detonation blast), which is an acoustic event.
• the blast includes acoustic waves, schematically represented as (spherical/circular) wavefront 420.
• a velocity of the projectile 400 can be higher than the sound velocity in the air (at least during part of its motion, or all of its motion up to the impact point).
• the projectile 400 can be a supersonic projectile.
• Operation 470 can include identifying, in the acoustic data, data D acoustic i mpact informative of an event corresponding to an impact (hit) of the projectile. Note that, as visible in Fig. 4D, it is not always known in advance which data will be present in the acoustic data provided by the acoustic sensors. The system can therefore rely for example on the smart process of Fig. 4D. The method of Fig. 4D first identifies which data is present in the acoustic data. If the acoustic data includes data informative of shock waves, but does not include data informative of a blast (the blast is below the detection threshold), then the processing circuitry 130 instructs to perform the method of Fig. 2B.
• the acoustic data provided by the acoustic sensors 120, or data informative thereof, is fed to the machine learning model 135.
• the acoustic data sensed by the acoustic sensors 120 over time can be fed in real time or quasi real time.
• the machine learning model 135 can output a prospect (a probability) that the piece of data includes an acoustic event corresponding to an impact of the projectile.
• the probability is above a threshold, this indicates that the acoustic event can be considered as an impact of the projectile, which corresponds to D acoustic i mpac t ⁇
• Some of the acoustic data includes an acoustic event corresponding to an impact of a projectile on an obstacle (or target).
• the obstacle is located at a distance from the acoustic sensors which meets the proximity criterion. This corresponds to the “positive examples”.
• the training set is then used (operation 491) to train the machine learning model 135 to predict, based on an input including acoustic data, whether it includes an acoustic event corresponding to an impact of a projectile.
• the training set includes acoustic events corresponding to an impact of a projectile, for different types of projectiles (e.g., different calibers, different weight, different shapes, etc.).
• the training set includes acoustic events corresponding to an impact of a projectile on an obstacle, for different types of materials of the obstacle (e.g., impact of the projectile on a metallic obstacle, impact of the projectile on an obstacle including glass, etc.).
• the acoustic data can be fed to each of the plurality of machine learning models.
• the output of the different machine learning models can be aggregated (using various methods such as averaging, voting method, etc.) to generate a decision of whether an impact of a projectile is present in the acoustic data.
• the output of the different machine learning models can be used to estimate the type of material on which the impact has occurred.
• the method further includes using (operation 480) D fiash and D acoustic impact to estimate the location of the projectile launch.
• Operation 480 can include (once the impact of the projectile has been identified), determining a corresponding time t impact in the acoustic data. This time t impact is an estimate of the time at which the impact of the projectile has occurred.
• time tf iash is an estimate of the time t 0 of the projectile launch.
• tf iash can be used to estimate the location of the projectile launch.
• M is the bullet mass
• p is the air density
• C cL is the drag coefficient of the projectile
• A is the projectile cross-section
• v 0 is an estimate of the velocity of the projectile (estimate of the muzzle velocity, which depends on the type of projectile).
• Fig. 41 can be used also when the shock waves cannot be detected by the acoustic sensors 120.
• Operation 481 can include using t impact and t biast (time at which the blast has occurred or has been sensed by the acoustic sensors) to determine the location of the projectile launch.
• k 0 is an assumption on a constant velocity of the projectile (estimate of the muzzle velocity, which depends on the type of projectile).
• system 100 can be mounted on a moving platform 508, such as a vehicle.
• a projectile 500 is launched at time t 0 .
• the imaging device senses the muzzle flash when the moving platform 508 is located at position Po, at time t ⁇ ash .
• the acoustic sensors sense the shock waves with a time delay (since the speed of sound is far slower than the speed of light).
• the moving platform 508 is located at position Pi, different from Po.
• the sensor 150 (see Fig. 1) can be used to determine the variation in the orientation/direction of the moving platform 508 (and in turn of the system 100) between the detection of the muzzle flash 510 and detection of the shock waves 525.
• the sensor 150 can be used to determine the change in the heading/direction (see 0) of the vehicle. Note that the distance between Po and Pi can be generally neglected for the purpose of angular calculations.
• the functionalities/operations can be performed by the one or more processors of the processing circuitry 130 in various ways.
• the operations described hereinafter can be performed by a specific processor, or by a combination of processors.
• the operations described hereinafter can thus be performed by respective processors (or processor combinations) in the processing circuitry 130 (or other processing circuitries), while, optionally, at least some of these operations may be performed by the same processor.
• the present disclosure should not be limited to be construed as one single processor always performing all the operations.

Landscapes

• Engineering & Computer Science (AREA)
• Radar, Positioning & Navigation (AREA)
• Remote Sensing (AREA)
• Physics & Mathematics (AREA)
• General Physics & Mathematics (AREA)
• General Engineering & Computer Science (AREA)
• Aviation & Aerospace Engineering (AREA)
• Electromagnetism (AREA)
• Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
• Compositions Of Macromolecular Compounds (AREA)
• Electrophonic Musical Instruments (AREA)

Applications Claiming Priority (2)

Publications (1)

Family

ID=90274372

Family Applications (1)

Country Status (5)

Families Citing this family (1)

Citations (9)

Family Cites Families (3)

• 2022
  • 2022-09-13 IL IL296482A patent/IL296482A/en unknown
• 2023
  • 2023-09-13 WO PCT/IL2023/050991 patent/WO2024057314A1/en not_active Ceased
  • 2023-09-13 IL IL319569A patent/IL319569A/en unknown
  • 2023-09-13 EP EP23864914.9A patent/EP4587779A1/de active Pending
  • 2023-09-13 KR KR1020257009914A patent/KR20250099108A/ko active Pending
  • 2023-09-13 AU AU2023340742A patent/AU2023340742A1/en active Pending

Patent Citations (9)

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events