KR102631310B1 - Trajectory prediction method and apparatus - Google Patents

Abstract

A trajectory prediction method and apparatus are disclosed. A trajectory prediction method according to an embodiment includes detecting inertial data of a projectile, obtaining initial firing conditions of the projectile based on the inertial data, and obtaining a reference trajectory of the projectile. and predicting the trajectory of the projectile by inputting the initial launch conditions and the reference trajectory into at least one neural network.

Description

Trajectory prediction method and apparatus {TRAJECTORY PREDICTION METHOD AND APPARATUS}

Embodiments relate to a trajectory prediction method and device using a neural network.

There are various techniques for estimating the trajectory of a flying object or projectile. The estimation method using an inertial sensor mathematically calculates and provides the position and attitude information of the projectile at each point in time using acceleration and angular velocity information under the assumption that the initial state of the aircraft or projectile is known.

GNSS (Global Navigation Satellite System) uses satellite signals to mathematically determine the user's location, and motion capture attaches a distinguishing device to the object to be measured and analyzes images taken with a camera to obtain the object's location and posture information. This is the way to obtain it.

However, these trajectory estimation methods have the problem of requiring expensive equipment and having very poor prediction performance for limited environments or objects with rapid movement.

For example, projectiles such as discs, baseballs, and artillery shells are small in size and have a very large movement at the beginning of launch, so the above-described trajectory estimation methods cannot accurately estimate the trajectory due to problems such as exceeding the measurement range, missing signals, and miniaturization. impossible.

Unlike aircraft, drones, etc., the above-mentioned projectiles have the characteristic that their flight trajectories are determined depending on initial conditions. Therefore, if an accurate aerodynamic model of the projectile and sufficient initial launch conditions exist, the subsequent flight trajectory of the projectile can be predicted.

However, in the case of the above-mentioned projectiles, the aerodynamic model is different depending on the shape, and the process for calculating the aerodynamic model is very complicated and consuming.

Embodiments may provide technology for predicting a trajectory using a neural network. However, technical challenges are not limited to the above-mentioned technical challenges, and other technical challenges may exist.

A trajectory prediction method according to an embodiment includes detecting inertial data of a projectile, obtaining initial firing conditions of the projectile based on the inertial data, and obtaining a reference trajectory of the projectile. and predicting the trajectory of the projectile by inputting the initial launch conditions and the reference trajectory into at least one neural network.

The initial firing conditions may include the posture of the projectile, the speed of the projectile, the rotation speed of the projectile, and the initial position of the projectile.

Obtaining the reference trajectory may include obtaining a function corresponding to the reference trajectory based on the position of the projectile.

The function may include polynomials.

The number of the at least one neural network may be determined based on the degree of the polynomial.

The predicting step may include extracting a trajectory coefficient representing the trajectory based on the reference trajectory, and predicting the trajectory based on the initial launch condition and the trajectory coefficient.

The trajectory prediction method may further include training the at least one neural network based on the initial firing condition and the reference trajectory.

The training step may include training the at least one neural network to input the initial firing condition and output trajectory coefficients based on the reference trajectory.

A trajectory prediction device according to an embodiment includes a sensor that detects inertial data of a projectile, obtains initial firing conditions of the projectile based on the inertial data, obtains a reference trajectory of the projectile, and obtains the initial launch condition of the projectile. and a processor that predicts the trajectory of the projectile by inputting launch conditions and the reference trajectory into at least one neural network.

The initial firing conditions may include the posture of the projectile, the speed of the projectile, the rotation speed of the projectile, and the initial position of the projectile.

The processor may obtain a function corresponding to the reference trajectory based on the location of the projectile.

The function may include polynomials.

The number of the at least one neural network may be determined based on the degree of the polynomial.

The processor may extract a trajectory coefficient representing the trajectory based on the reference trajectory and predict the trajectory based on the initial launch condition and the trajectory coefficient.

The processor may train the at least one neural network based on the initial firing condition and the reference trajectory.

The processor may train the at least one neural network to input the initial firing conditions and output trajectory coefficients based on the reference trajectory.

1 shows a schematic block diagram of a trajectory prediction device in one embodiment.
Figure 2 shows a schematic block diagram of the processor shown in Figure 1.
Figure 3 shows an example of a neural network used for trajectory prediction.
Figure 4 shows an example of a reference trajectory.
Figure 5 shows the process of predicting a trajectory through multiple neural networks.
Figure 6 shows an example of a predicted trajectory.
FIG. 7 shows the operation sequence of the trajectory prediction device shown in FIG. 1.

Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only and may be changed and implemented in various forms. Accordingly, the actual implementation form is not limited to the specific disclosed embodiments, and the scope of the present specification includes changes, equivalents, or substitutes included in the technical idea described in the embodiments.

Terms such as first or second may be used to describe various components, but these terms should be interpreted only for the purpose of distinguishing one component from another component. For example, a first component may be named a second component, and similarly, the second component may also be named a first component.

When a component is referred to as being “connected” to another component, it should be understood that it may be directly connected or connected to the other component, but that other components may exist in between.

Singular expressions include plural expressions unless the context clearly dictates otherwise. In this specification, terms such as “comprise” or “have” are intended to designate the presence of the described features, numbers, steps, operations, components, parts, or combinations thereof, and are intended to indicate the presence of one or more other features or numbers, It should be understood that this does not exclude in advance the possibility of the presence or addition of steps, operations, components, parts, or combinations thereof.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by a person of ordinary skill in the art. Terms as defined in commonly used dictionaries should be interpreted as having meanings consistent with the meanings they have in the context of the related technology, and unless clearly defined in this specification, should not be interpreted in an idealized or overly formal sense. No.

Hereinafter, embodiments will be described in detail with reference to the attached drawings. In the description with reference to the accompanying drawings, identical components will be assigned the same reference numerals regardless of the reference numerals, and overlapping descriptions thereof will be omitted.

Figure 1 shows a schematic block diagram of a trajectory prediction device in one embodiment, and Figure 2 shows a schematic block diagram of the processor shown in Figure 1.

Referring to FIGS. 1 and 2 , the trajectory prediction device 10 can predict the trajectory of a projectile or flying object. The trajectory prediction device 10 can predict the trajectory of a projectile using a neural network.

Neural networks (or artificial neural networks) can include statistical learning algorithms that mimic neurons in biology in machine learning and cognitive science. A neural network can refer to an overall model in which artificial neurons (nodes), which form a network through the combination of synapses, change the strength of the synapse connection through learning and have problem-solving capabilities.

Neural networks may include deep neural networks. Neural networks include CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), perceptron, FF (Feed Forward), RBF (Radial Basis Network), DFF (Deep Feed Forward), LSTM (Long Short Term Memory), GRU (Gated Recurrent Unit), AE (Auto Encoder), VAE (Variational Auto Encoder), DAE (Denoising Auto Encoder), SAE (Sparse Auto Encoder), MC (Markov Chain), HN (Hopfield Network), BM (Boltzmann Machine) ), Restricted Boltzmann Machine (RBM), Depp Belief Network (DBN), Deep Convolutional Network (DCN), Deconvolutional Network (DN), Deep Convolutional Inverse Graphics Network (DCIGN), Generative Adversarial Network (GAN), Liquid State Machine (LSM) ), Extreme Learning Machine (ELM), Echo State Network (ESN), Deep Residual Network (DRN), Differential Neural Computer (DNC), Neural Turning Machine (NTM), Capsule Network (CN), Kohonen Network (KN), and AN. (Attention Network) may be included.

The machine reading device 10 may be implemented in the form of a mobile terminal. For example, the machine reading device 10 may be implemented as an IoT device, a machine-type communication device, or a portable electronic device.

Portable electronic devices include laptop computers, mobile phones, smart phones, tablet PCs, mobile internet devices (MIDs), personal digital assistants (PDAs), and enterprise digital assistants (EDAs). ), digital still camera, digital video camera, PMP (portable multimedia player), PND (personal navigation device or portable navigation device), handheld game console, e-book (e-book), can be implemented as a smart device. For example, a smart device may be implemented as a smart watch or smart band.

The trajectory prediction device 10 includes a processor 200. The trajectory prediction device 10 may further include a sensor 100 and a memory 300.

The sensor 100 may be located inside or outside the trajectory prediction device 10. The sensor 100 can detect the movement of a projectile. The sensor 100 can detect inertial data of a projectile. Inertial data may include acceleration and angular velocity. For example, sensor 100 may include an acceleration sensor, an angular velocity sensor, and a magnetic sensor. The sensor 100 may output the sensed inertial data to the processor 200.

The processor 200 may process data stored in the memory 300. The processor 200 may execute computer-readable code (eg, software) stored in the memory 300 and instructions triggered by the processor 200 .

The “processor 200” may be a data processing device implemented in hardware that has a circuit with a physical structure for executing desired operations. For example, the intended operations may include code or instructions included in the program.

For example, data processing devices implemented in hardware include microprocessors, central processing units, processor cores, multi-core processors, and multiprocessors. , ASIC (Application-Specific Integrated Circuit), and FPGA (Field Programmable Gate Array).

The processor 200 may predict the trajectory of the projectile based on inertial data. The processor 200 may obtain the initial launch conditions of the projectile based on inertial data. The initial firing conditions of the projectile may include the pose of the projectile, the velocity of the projectile, the rotational velocity of the projectile, and the initial position of the projectile.

The processor 200 may obtain a reference trajectory of the projectile. The processor 200 may obtain a function corresponding to the reference trajectory based on the location of the projectile. The function corresponding to the reference trajectory may include a polynomial.

The processor 200 may predict the trajectory of the projectile by inputting the initial launch conditions and reference trajectory into at least one neural network. In other words, the processor 200 can predict the trajectory of a projectile using one or more neural networks.

The number of at least one neural network may be determined based on the degree of the polynomial. The processor 200 may extract a trajectory coefficient indicating the trajectory of the projectile based on the reference trajectory. The processor 200 may predict the trajectory of the projectile based on the initial launch conditions and trajectory coefficients. The process by which the processor 200 extracts trajectory coefficients will be described in detail with reference to FIGS. 3 to 5.

The processor 200 may train at least one neural network. The processor 200 may train at least one neural network based on the initial firing conditions and the reference trajectory. The processor 200 may input initial firing conditions and train at least one neural network to output trajectory coefficients based on the reference trajectory.

The processor 200 may include a reference trajectory measurer 210, a reference trajectory preprocessor 220, an inertial data preprocessor 230, a learner 240, and a trajectory predictor 250. The sensor 100 may detect inertial data and output it to the inertial data preprocessor 220.

The reference trajectory measurer 210 can obtain a reference trajectory for training a neural network. The reference trajectory preprocessor 220 may extract trajectory coefficients based on the obtained reference trajectory. The reference trajectory preprocessor 220 may extract trajectory coefficients by processing the reference trajectory of the projectile.

The inertial data preprocessor 220 may obtain initial launch conditions based on the inertial data. The inertial data preprocessor 220 can use the received inertial data to obtain the attitude, speed, rotation speed, and initial position of the projectile at a specific point in time.

The learner 240 can train a neural network. The learner 240 may learn a neural network based on the initial launch conditions and trajectory coefficients input from the reference trajectory preprocessor 220 and the inertial data preprocessor 230.

In a neural network, the number of hidden layers and the shape of neurons may vary depending on the type of initial firing condition, the size of the input data, and the characteristics of the projectile. The learner 240 can generate an input/output model for predicting a trajectory by training a neural network.

The trajectory predictor 250 can predict the trajectory of the projectile using the input and output model generated by the learner 240.

The memory 300 may store instructions (or programs) executable by the processor 200. For example, the instructions may include instructions for executing the operation of the processor and/or the operation of each component of the processor.

The memory 300 may be implemented as a volatile memory device or a non-volatile memory device.

Volatile memory devices may be implemented as dynamic random access memory (DRAM), static random access memory (SRAM), thyristor RAM (T-RAM), zero capacitor RAM (Z-RAM), or twin transistor RAM (TTRAM).

Non-volatile memory devices include EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory, MRAM (Magnetic RAM), Spin-Transfer Torque (STT)-MRAM (MRAM), and Conductive Bridging RAM (CBRAM). , FeRAM (Ferroelectric RAM), PRAM (Phase change RAM), Resistive RAM (RRAM), Nanotube RRAM (Nanotube RRAM), Polymer RAM (PoRAM), Nano Floating Gate Memory (NFGM), holographic memory, molecular electronic memory device, or insulator resistance change memory.

Below, with reference to FIGS. 3 to 6, the process by which the processor 200 predicts the trajectory of a projectile will be described in detail.

Figure 3 shows an example of a neural network used for trajectory prediction, and Figure 4 shows an example of a reference trajectory.

Figure 5 shows the process of predicting a trajectory through a plurality of neural networks, and Figure 6 shows an example of a predicted trajectory.

Referring to FIGS. 3 to 6 , the processor 200 can predict the trajectory of a projectile using a neural network. The processor 200 may extract trajectory coefficients representing the trajectory of the projectile through a neural network.

The neural network may include a hidden layer 310 and an output layer 330. The neural network may include a plurality of hidden layers 310.

The processor 200 may input the initial launch conditions of the projectile obtained from inertial data into the neural network. For example, the processor 200 may input the posture, speed, rotation speed, and initial position of the projectile into a neural network to extract trajectory coefficients representing the trajectory of the projectile.

The processor 200 may input the initial launch conditions of the projectile into the neural network. As described above, the initial launch conditions may include the attitude, speed, rotation speed, and initial position of the projectile obtained based on inertial data.

The processor 200 may change the number of hidden layers 310 of the neural network depending on the embodiment. The hidden layer 310 may include neurons in a sigmoid shape.

The output layer 330 may include linear neurons. The output layer 330 can extract trajectory coefficients using the calculation results of the hidden layer 310. After learning of the neural network is completed, the processor 200 can predict the trajectory of the projectile using only the initial launch conditions obtained from the inertial data preprocessor 220.

The reference trajectory measurer 210 and the reference trajectory preprocessor 220 may generate learning data for learning a neural network. Reference trajectory finder 210 may be implemented using a platform such as motion capture or GNSS (Global Navigation Satellite System).

The reference trajectory generated by the reference trajectory measuring device 210 may be expressed as coordinate values. For example, if the reference trajectory appears on a two-dimensional coordinate space, the coordinate value can be expressed as (x ₁ , y ₁ ), and if the reference trajectory appears on a three-dimensional coordinate space, (x ₁ , y) ₁ , z ₁ ).

The predicted trajectory of the projectile may have the same dimensions as the reference trajectory. The example in FIG. 4 shows an example of a reference trajectory displayed on a two-dimensional coordinate space.

The reference trajectory measuring device 210 can acquire coordinates given as points, as shown in the left graph of FIG. 4. For example, the reference trajectory measuring device 210 measures (x ₁ , y ₂ ), (x ₂ , y ₂ )… You can obtain coordinates such as (x _n ,y _n ).

The reference trajectory preprocessor 220 can obtain a reference trajectory as shown in the right graph of FIG. 4 using coordinates given as points. The reference trajectory preprocessor 220 can convert the coordinate values obtained by the reference trajectory measuring device 210 into a curve equation through curve fitting and then extract the trajectory coefficients.

The reference trajectory preprocessor 220 may obtain the reference trajectory in the form of a function corresponding to the reference trajectory. For example, the reference trajectory preprocessor 220 may obtain trajectory coefficients from the equation or polynomial of the transformed curve.

For example, the polynomial may have the form y=a ₁ x ² +a ₂ x+b, and the reference trajectory preprocessor 220 may obtain the trajectory coefficients a1 and a2 using curve fitting.

The dimension of the polynomial may change depending on the complexity of the trajectory. In the case of a second-order polynomial, it can have two trajectory coefficients, and in the case of a third-order polynomial, it can have three trajectory coefficients. In other words, as the complexity of the trajectory increases, the number of trajectory coefficients may increase.

The number of neural networks that the processor 200 uses to predict a trajectory may vary depending on the number of trajectory coefficients. As in the example of Figure 5, the number of neural networks may be equal to the number of trajectory coefficients.

In other words, when there are two trajectory coefficients to extract, two neural networks of FIG. 3 may be needed, and when there are three trajectory coefficients, three neural networks may be needed.

The processor 200 can predict the trajectory of a projectile using a plurality of extracted trajectory coefficients. After the trajectory coefficient is determined through learning, the processor 200 can predict the trajectory of the projectile by inputting only inertial data.

The example in FIG. 6 shows an example of the trajectory of a projectile predicted by the trajectory prediction device 10. When using a neural network that has completed learning, the trajectory prediction device 10 can predict the flight trajectory of a projectile by using only the inertial sensor as an input.

Since the trajectory prediction device 10 uses only inertial sensors, it is not limited by the surrounding environment, can be miniaturized, and can be implemented as a low-cost platform.

When using measurement equipment using a camera, such as a bicon or motion capture, to measure the flight trajectory of a projectile, the flight trajectory can be estimated only in a special environment where multiple cameras are installed, but the trajectory prediction device 10 has completed learning. By predicting the trajectory using a neural network, the trajectory of a projectile equipped with only an inertial sensor can be predicted without constraints from the surrounding environment.

FIG. 7 shows the operation sequence of the trajectory prediction device shown in FIG. 1.

The sensor 100 may detect inertial data of a projectile (710).

The processor 200 may obtain the initial launch conditions of the projectile based on the inertial data (730). Initial launch conditions may include the attitude of the projectile, the speed of the projectile, the rotation speed of the projectile, and the initial position of the projectile.

The processor 200 may obtain the reference trajectory of the projectile. The processor 200 may obtain a function corresponding to the reference trajectory based on the location of the projectile. Functions can contain polynomials.

The processor 200 may predict the trajectory of the projectile by inputting the initial launch conditions and reference trajectory into at least one neural network. The number of at least one neural network may be determined based on the degree of the polynomial.

The processor 200 may extract trajectory coefficients representing the trajectory based on the reference trajectory. The processor 200 may predict the trajectory based on initial launch conditions and trajectory coefficients.

The processor 200 may train at least one neural network based on the initial firing conditions and the reference trajectory. The processor 200 may input initial firing conditions and train at least one neural network to output trajectory coefficients based on the reference trajectory.

The embodiments described above may be implemented with hardware components, software components, and/or a combination of hardware components and software components. For example, the devices, methods, and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, and a field programmable gate (FPGA). It may be implemented using a general-purpose computer or a special-purpose computer, such as an array, programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and software applications running on the operating system. Additionally, a processing device may access, store, manipulate, process, and generate data in response to the execution of software. For ease of understanding, a single processing device may be described as being used; however, those skilled in the art will understand that a processing device includes multiple processing elements and/or multiple types of processing elements. It can be seen that it may include. For example, a processing device may include multiple processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are possible.

Software may include a computer program, code, instructions, or a combination of one or more of these, which may configure a processing unit to operate as desired, or may be processed independently or collectively. You can command the device. Software and/or data may be used on any type of machine, component, physical device, virtual equipment, computer storage medium or device to be interpreted by or to provide instructions or data to a processing device. , or may be permanently or temporarily embodied in a transmitted signal wave. Software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored on a computer-readable recording medium.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. A computer-readable medium may include program instructions, data files, data structures, etc., singly or in combination, and the program instructions recorded on the medium may be specially designed and constructed for the embodiment or may be known and available to those skilled in the art of computer software. It may be possible. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks. -Includes optical media (magneto-optical media) and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, etc. Examples of program instructions include machine language code, such as that produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc.

The hardware devices described above may be configured to operate as one or multiple software modules to perform the operations of the embodiments, and vice versa.

Although the embodiments have been described with limited drawings as described above, those skilled in the art can apply various technical modifications and variations based on this. For example, the described techniques are performed in a different order than the described method, and/or components of the described system, structure, device, circuit, etc. are combined or combined in a different form than the described method, or other components are used. Alternatively, appropriate results may be achieved even if substituted or substituted by an equivalent.

Therefore, other implementations, other embodiments, and equivalents of the claims also fall within the scope of the claims described below.

Claims (17)

A trajectory prediction device acquiring initial launch conditions of the projectile based on inertial data of the projectile; and
A trajectory prediction device predicting the trajectory of the projectile by inputting the initial launch conditions into at least one neural network learned based on a reference trajectory associated with the projectile.
Including,
The at least one neural network is,
A trajectory prediction method that is learned by using the trajectory coefficients obtained by curve fitting the reference trajectory as ground truth.
According to paragraph 1,
The initial launch conditions are,
The attitude of the projectile, the speed of the projectile, the rotational speed of the projectile and the initial position of the projectile
Trajectory prediction method including.
delete delete delete delete According to paragraph 1,
Training the at least one neural network based on the initial firing conditions of the projectile and the reference trajectory.
A trajectory prediction method further comprising:
In clause 7,
The learning step is,
training the at least one neural network to input the initial firing conditions of the projectile and output the trajectory coefficient
Trajectory prediction method including.
A computer program combined with hardware and stored on a medium to execute the method of any one of claims 1, 2, 7, and 8.
Obtaining initial firing conditions of the projectile based on inertial data of the projectile, and determining the trajectory of the projectile by inputting the initial firing conditions into at least one neural network learned based on a reference trajectory associated with the projectile. processor that predicts
Including,
The at least one neural network is:
A trajectory prediction device that is learned using trajectory coefficients obtained by curve fitting the reference trajectory as ground truth.
According to clause 10,
The initial launch conditions are,
The attitude of the projectile, the speed of the projectile, the rotational speed of the projectile and the initial position of the projectile
A trajectory prediction device comprising:
delete delete delete delete According to clause 10,
The processor,
training the at least one neural network based on the initial firing conditions of the projectile and the reference trajectory.
Trajectory prediction device.
According to clause 16,
The processor,
training the at least one neural network to input the initial firing conditions of the projectile and output the trajectory coefficient
Trajectory prediction device.

Images