Non-cooperative target trajectory prediction method fusing camera miss distance and encoder information

Abstract

The invention discloses a non-cooperative target trajectory prediction method fusing camera miss distance and encoder information. According to the method, the actual track sequence of the non-cooperative target is obtained by fusing the miss distance of the camera of the optical telescope and the grating encoder information, the motion track of the tracking target in the blind area time is predicted by using a prediction algorithm, the telescope is controlled to move directly through optical feedback at the camera sampling moment, and the motion track of the tracking target in the blind area time is obtained at the encoder sampling moment. The corresponding predicted value and the encoder measurement value are subtracted to obtain the tracking error in the blind area time, and the tracking error is input into the telescope motion control system, so that the control frequency and closed-loop bandwidth of the telescope motion control system are improved, and the error between the reference trajectory and the actual trajectory of the tracking target is reduced; therefore, the tracking capability of the optical telescope on the non-cooperative target is improved.

Description

Technical Field

[0001] The invention relates to the field of trajectory prediction of non-cooperative targets, and in particular to a non-cooperative target trajectory prediction method that integrates camera miss distance and encoder information. Background Art

[0002] With the development of astronomy, aerospace, and aviation, optical telescopes have become increasingly indispensable in applications such as astronomical observation, space target monitoring, and target tracking. Especially in the fields of space exploration and national defense security, optical telescopes play an important role in performing complex and diverse observation tasks. Among them, high-precision tracking of non-cooperative targets has become the focus and difficulty of current technical research.

[0003] Traditional optical telescopes use optical cameras for image feedback, that is, the tracking target is imaged by the optical camera, and its coordinate pixels in the field of view are identified and calculated by image processing algorithms to determine the target's miss distance (i.e., the deviation between the target position and the telescope's aiming axis). This miss distance is input into the telescope's servo control system as a feedback signal to form a closed-loop control, thereby achieving accurate tracking of the target.

[0004] However, the traditional image feedback system has the problems of slow response speed, poor real-time performance and easy loss of tracking targets when dealing with high-speed moving targets because it relies on the sampling frequency and image processing algorithm of the optical camera. This shortcoming makes the traditional optical telescope limited in performing certain specific tasks, especially when tracking non-cooperative targets. The so-called non-cooperative targets refer to those space objects that cannot provide trajectory prediction data and have no active cooperation capabilities, such as low-orbit satellites, drones, space debris, etc. that do not communicate and cooperate. Unlike cooperative targets, the trajectory prediction data of non-cooperative targets cannot be obtained before the tracking task, so the tracking can only rely on the image feedback in the optical telescope. When tracking fast-moving targets such as drones and low-orbit satellites, the existing optical cameras have problems such as low sampling frequency, slow feedback speed and even delayed response, which seriously restricts the telescope's tracking ability for non-cooperative targets. Summary of the invention

[0005] In view of the shortcomings of the existing non-cooperative target trajectory prediction technology, the purpose of the present invention is to provide a non-cooperative target trajectory prediction method that integrates camera miss distance and encoder information. By combining two different feedback signals, the tracking trajectory of the non-cooperative target can be accurately predicted in real time, thereby improving the tracking response speed and tracking accuracy of the optical telescope for the non-cooperative target.

[0006] To achieve the above object, the present invention provides a non-cooperative target trajectory prediction method that integrates camera miss distance and encoder information. The method comprises the following steps:

[0007] Step 1: System initialization, including acquisition of initial guidance information and initialization settings of the telescope;

[0008] Step 2: At the camera sampling moment, the optical camera exposes and images the tracking target, recognizes the image and calculates the miss distance, which is further converted into the angular position error of the telescope in the azimuth and pitch directions and input into the telescope motion control unit. At the same time, the encoder samples and reads the encoder measurement value at the current moment.

[0009] Step 3: After completing an optical imaging and miss distance reading, the telescope tracks the time of entering the blind area. The camera miss distance and grating encoder information obtained at the camera sampling time are integrated to calculate the actual position of the target and update the actual trajectory sequence. Based on the actual trajectory sequence, the trajectory prediction algorithm is used to predict the motion trajectory of the non-cooperative target before the next optical imaging is completed, and interpolation is performed according to the encoder sampling time to obtain the predicted trajectory sequence.

[0010] Step 4: At the encoder sampling moment, read the grating encoder measurement value. At the same time, read the predicted value corresponding to the current moment in the predicted trajectory sequence, subtract it from the encoder measurement value, and obtain the angular position error, which is input into the telescope motion control unit.

[0011] Step 5: Confirm whether to continue the tracking mission. If yes, repeat steps 2 to 4. If no, end the mission and turn off the telescope control system.

[0012] Furthermore, step 1 specifically includes:

[0013] Step 1.1, start the telescope motion control system and initialize the parameters. The optical telescope is equipped with an optical camera and a grating encoder, which are used to measure the target miss distance and the telescope rotation angle respectively. The optical camera sampling time is T c , the grating encoder sampling time is T e , the camera sampling time is an integer multiple of the encoder sampling time, that is, T c =mT e , m is a positive integer. The telescope control system sampling time T s The same as the encoder sampling time, that is, T s = T e The time between two imagings of the optical camera is called the blind time. During this time period, the system cannot obtain the camera image information, but the encoder and motion control system still measure and control at the above system sampling frequency.

[0014] Step 1.2, input guidance information, that is, the initial trajectory of the tracking target, and set the initial target actual trajectory sequence. The initial target actual trajectory sequence R of the system a (0), R e(0) selects the data corresponding to the last four camera sampling moments in the above guidance information, where R a Indicates azimuth data, R e Indicates pitch angle data.

[0015] Step 1.3, the telescope motion control system controls the telescope to track the target according to the guidance information so that it appears in the field of view of the optical camera.

[0016] Furthermore, step 2 specifically includes:

[0017] Step 2.1: At the kth optical camera sampling moment, the camera exposes and images the tracking target, identifies the image pixels and calculates the miss distance.

[0018] Step 2.2: Convert the miss distance into the angular error of the telescope in the azimuth and elevation directions. Suppose the miss distance on the imaging horizontal axis is Δx c , the miss distance on the vertical axis is Δy c , the focal length of the camera is f c , then at the kth optical camera sampling moment, that is, the kmth system sampling moment, the azimuth angle error e ca , pitch angle error e ce The calculations can be described by the following formulas:

[0019]

[0020]

[0021] The position error calculated according to the above formula is input into the telescope motion control unit.

[0022] Step 2.3, read the current grating encoder measurement value at the same time, that is, the rotation angle of the telescope in the azimuth and pitch directions, recorded as y ea (km), y ee (km).

[0023] Furthermore, step 3 specifically includes:

[0024] Step 3.1, after completing the kth optical imaging and miss distance reading, the telescope tracks the time when it enters the kth blind zone. The camera miss distance obtained at the camera sampling moment and the grating encoder information are integrated to calculate the actual position of the target, as shown in the following formula:

[0025]

[0026]

[0027] Furthermore, the actual trajectory sequence is updated, i.e.

[0028]

[0029]

[0030] Step 3.2: According to the updated actual trajectory sequence R a , R e , a polynomial fitting algorithm is used to predict the motion trajectory P of the non-cooperative target before the next optical imaging is completed a , P e , the polynomial model is shown in the following formula

[0031]

[0032] The actual trajectory sequence R a , R e Substitute the data points into the polynomial model to further obtain the matrix equation, as shown in the following formula

[0033]

[0034] Simplified to Xa = P. The least square method is used to solve the optimal coefficient a = (X T X) -1 X T P, and obtain the polynomial model parameters a0, a1, a2, and a3, that is, the motion trajectory model is predicted based on the actual trajectory sequence.

[0035] Step 3.3, substitute the model parameters updated in step 3.2 into the polynomial model to complete the update of the prediction model. Interpolation is performed according to the encoder sampling time. Specifically, select , substitute into the updated polynomial model, and calculate the interpolated predicted trajectory sequence according to the polynomial model formula.

[0036] Furthermore, step 4 specifically includes:

[0037] Step 4.1, at the nth encoder sampling moment (n = 1, 2, …, m) in the kth blind time, read the grating encoder measurement value.

[0038] Step 4.2, at the same time, the corresponding value of the interpolated predicted trajectory sequence at the current moment is subtracted from the current encoder measurement value to calculate the angular position error of the telescope in the blind time and input it into the telescope motion control unit.

[0039] Furthermore, step 5 specifically includes:

[0040] Confirm whether the tracking mission will continue. If yes, repeat steps 2 to 4. If no, end the mission and send a control command to shut down the telescope control system.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] The present invention effectively solves the problems of slow tracking speed and easy loss of tracking target in the existing optical feedback tracking technology, improves the trajectory prediction accuracy of the telescope for non-cooperative targets during the blind time, thereby significantly improving the trajectory tracking capability for non-cooperative targets. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a specific flow diagram of the present invention;

[0044] Figure 2 is a discrete timing schematic diagram of the present invention;

[0045] Figure 3 It is a schematic diagram of optical camera imaging and miss distance calculation;

[0046] Figure 4 It is a comparison diagram of reference trajectories of the traditional optical feedback tracking method and the method of the present invention;

[0047] Figure 5 3 and 4. The figures are comparison diagrams of non-cooperative target tracking effects; wherein, (a) is a non-cooperative target tracking effect diagram of a traditional optical feedback method, and (b) is a non-cooperative target tracking effect diagram of the method of the present invention. DETAILED DESCRIPTION

[0048] The specific implementation of the present invention will be described below with reference to the accompanying drawings and embodiments.

[0049] like Figures 1 to 5 As shown, the present invention provides a non-cooperative target trajectory prediction algorithm that integrates camera miss distance and grating encoder information, which specifically includes the following steps:

[0050] Step 1: System initialization, including acquisition of initial guidance information and initialization settings of the telescope;

[0051] Step 2: At the camera sampling moment, the optical camera exposes and images the tracking target, recognizes the image and calculates the miss distance, which is further converted into the angular position error of the telescope in the azimuth and pitch directions and input into the telescope motion control unit. At the same time, the encoder samples and reads the encoder measurement value at the current moment.

[0052] Step 3: After completing an optical imaging and miss distance reading, the telescope tracks the time of entering the blind spot. The camera miss distance and grating encoder information obtained at the camera sampling time are integrated to calculate the actual position of the target and update the actual trajectory sequence. Based on the actual trajectory sequence, a polynomial fitting algorithm is used to predict the motion trajectory of the non-cooperative target before the next optical imaging is completed, and interpolation is performed according to the encoder sampling time to obtain the predicted trajectory sequence.

[0053] Step 4: At the encoder sampling moment, read the grating encoder measurement value. At the same time, read the predicted value corresponding to the current moment in the predicted trajectory sequence, subtract it from the encoder measurement value, and obtain the angular position error, which is input into the telescope motion control unit.

[0054] Step 5: Confirm whether to continue the tracking mission. If yes, repeat steps 2 to 4. If no, end the mission and turn off the telescope control system.

[0055] It should be noted that the optical telescope is equipped with an optical camera and a grating encoder, which are used to measure the distance of the tracking target from missing and the rotation angle of the telescope respectively.

[0056] Figure 2 is a schematic diagram of the discrete timing of the present invention. The optical camera sampling time of the present invention is T c , the grating encoder sampling time is T e , the camera sampling time is an integer multiple of the encoder sampling time, that is, T c = mT e , m is a positive integer. The telescope control system sampling time T s The same as the encoder sampling time, that is, T s = T e The time between two imagings of the optical camera is called the blind time. During this time period, the system cannot obtain the camera image information, but the encoder and motion control system still measure and control at the above system sampling frequency. Therefore, the non-cooperative target trajectory prediction algorithm proposed by the present invention that integrates the camera miss distance and the grating encoder information can be used to calculate the actual trajectory sequence of the non-cooperative target, further predict the target motion trajectory, and interpolate the predicted trajectory in the blind time.

[0057] It should be noted that step 1 specifically includes starting the telescope motion control system and setting parameter initialization, inputting initial guidance information, and controlling the telescope to track the target so that it appears in the field of view of the optical camera. The actual trajectory sequence R of the initial target in step 1 a (0), R e (0) selects the data corresponding to the last four camera sampling moments in the guidance information, where R a Indicates azimuth data, R e Indicates the pitch angle data.

[0058] It should be noted that step 2 specifically includes, at the kth optical camera sampling moment, the camera exposes and images the tracking target, identifies the image pixels and calculates the miss amount. Further, the miss amount is converted into the angular error of the telescope in the azimuth and pitch directions and input into the telescope motion control unit. At the same time, the current grating encoder measurement value is read.

[0059] Figure 3 Schematic diagram of optical camera imaging and miss distance calculation. The specific implementation steps of the miss distance and rotation angle error conversion described in step 2 are as follows: Assume that the miss distance on the horizontal axis of the imaging is Δx c , the miss distance on the vertical axis is Δy c , the focal length of the camera is f c , then at the kth optical camera sampling moment, that is, the kmth system sampling moment, the azimuth angle error e ca , pitch angle error e ce The calculations can be described by the following formulas:

[0060]

[0061]

[0062] The position error calculated according to the above formula is input into the telescope motion control unit.

[0063] It should be noted that the current grating encoder measurement value is read in step 2, that is, the rotation angle of the telescope in the azimuth and elevation directions, which are recorded as y and ea (km), y ee (km).

[0064] It should be noted that step 3 specifically includes: after completing an optical imaging and miss distance reading, the telescope tracks the time of entering the blind spot. The camera miss distance and grating encoder information obtained at the camera sampling time are integrated to calculate the actual position of the target and update the actual trajectory sequence. Based on the actual trajectory sequence, a polynomial fitting algorithm is used to predict the motion trajectory of the non-cooperative target before the next optical imaging is completed, and interpolation is performed according to the encoder sampling time to obtain a predicted trajectory sequence.

[0065] It should be noted that the specific implementation of step 3 is as follows: After completing the kth optical imaging and miss distance reading, the telescope enters the kth blind time. The camera miss distance obtained at the camera sampling moment and the grating encoder information are integrated to calculate the actual position of the target, as shown in the following formula:

[0066]

[0067]

[0068] Furthermore, the actual trajectory sequence is updated, i.e.

[0069]

[0070]

[0071] According to the updated actual trajectory sequence R a , Re , a third-order polynomial fitting algorithm is used to predict the motion trajectory P of the non-cooperative target before the next optical imaging is completed a , P e , the third-order polynomial model is shown in the following formula

[0072]

[0073] The actual trajectory sequence R a , R e Substitute the data points into the polynomial model to further obtain the matrix equation, as shown in the following formula

[0074]

[0075] Simplified to Xa = P. The least square method is used to solve the optimal coefficient a = (X T X) -1 X T P , and obtain the third-order polynomial model parameters a0, a1, a2, and a3, that is, the motion trajectory model is predicted according to the actual trajectory sequence. Substitute the updated model parameters into the polynomial model to complete the update of the prediction model. Interpolate according to the encoder sampling time. Specifically, select , substitute into the updated polynomial model, and calculate the interpolated predicted trajectory sequence according to the third-order polynomial model formula.

[0076] It should be noted that step 4 specifically includes reading the grating encoder measurement value at the nth encoder sampling moment (n = 1, 2, …, m) in the kth blind time. At the same time, the corresponding value of the interpolated prediction trajectory sequence at the current moment is subtracted from the current encoder measurement value to calculate the angular position error of the telescope in the blind time and input it into the telescope motion control unit.

[0077] It should be noted that step 5 specifically includes: confirming whether the tracking task is to continue, if so, repeating steps 2 to 4; if not, ending the task and sending a control command to shut down the telescope control system.

[0078] This embodiment builds a telescope motion control system, and its sensors specifically include the optical camera and the grating encoder. The optical camera can image the non-cooperative target and calculate its miss distance and tracking error. The grating encoder can measure the telescope rotation angle. The calculated tracking error can be directly input into the telescope motion control system to control the telescope to rotate and track the non-cooperative target.

[0079] In this embodiment, the conventional optical feedback method is compared with the non-cooperative target trajectory prediction algorithm proposed in the present invention that integrates the camera miss distance and the grating encoder information. Figure 4By comparing the reference trajectory of direct optical camera sampling and the method proposed in the present invention, it can be seen that the target trajectory predicted by the method proposed in the present invention is closer to the actual trajectory of the non-cooperative target, and is also beneficial to improving the closed-loop control frequency of the system.

[0080] Figure 5 It is the tracking effect of the telescope on the non-cooperative target. It can be seen that compared with the traditional optical feedback tracking method, the method proposed in the present invention improves the closed-loop tracking accuracy and effectively reduces the tracking error, so that the tracking performance of the telescope on the non-cooperative target meets the expected requirements.

[0081] In summary, the present invention provides a non-cooperative target trajectory prediction method that integrates camera miss amount and encoder information. The present invention obtains the actual trajectory sequence of the non-cooperative target by integrating the camera miss amount and grating encoder information of the optical telescope, and uses a prediction algorithm to predict the motion trajectory of the tracking target within the blind time. At the camera sampling moment, the telescope movement is directly controlled by optical feedback, and at the encoder sampling moment, the corresponding predicted value is subtracted from the encoder measurement value to obtain the tracking error in the blind time, which is input into the telescope motion control system, thereby improving the control frequency and closed-loop bandwidth of the telescope motion control system, reducing the error between the reference trajectory and the actual trajectory of the tracking target, and thus improving the optical telescope's tracking ability for non-cooperative targets.

[0082] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A non-cooperative target trajectory prediction method integrating camera miss distance and encoder information, characterized in that: The steps include: Step 1: System initialization, including acquisition of initial guidance information and initialization settings of the telescope; Step 2: At the camera sampling moment, the optical camera exposes and images the tracking target, recognizes the image and calculates the miss distance, which is further converted into the angular position error of the telescope in the azimuth and elevation directions and input into the telescope motion control unit; at the same time, the grating encoder samples and reads the encoder measurement value at the current moment; Step 3: After completing an optical imaging and off-target amount reading, the telescope tracks the time of entering the blind zone; The camera miss distance and grating encoder information obtained at the camera sampling moment are integrated to calculate the actual position of the target and update the actual trajectory sequence. Based on the actual trajectory sequence, the trajectory prediction algorithm is used to predict the motion trajectory of the non-cooperative target before the next optical imaging is completed, and interpolation is performed according to the encoder sampling moment to obtain the predicted trajectory sequence. Step 4: At the encoder sampling moment, read the grating encoder measurement value; at the same time, read the predicted value corresponding to the current moment in the predicted trajectory sequence, subtract it from the encoder measurement value, and obtain the angular position error, which is input into the telescope motion control unit; Step 5: Confirm whether to continue the tracking mission. If yes, repeat steps 2 to 4. If no, end the mission and turn off the telescope control system.

2. The non-cooperative target trajectory prediction method of fusion camera miss distance and encoder information according to claim 1, characterized in that, Step 1 specifically includes: Step 1.1: Start the telescope motion control system and set parameters for initialization; the optical telescope is equipped with an optical camera and a grating encoder, which are used to measure the target miss distance and the telescope rotation angle respectively; the optical camera sampling time is T c , the grating encoder sampling time is T e , the camera sampling time is an integer multiple of the encoder sampling time; the telescope control system sampling time T s The same as the encoder sampling time; the time between two imaging of the optical camera is called the blind time, during which the system cannot obtain camera image information, but the encoder and motion control system still measure and control at the above system sampling frequency; Step 1.2: Input guidance information, i.e. the initial trajectory of the tracking target, and set the initial target actual trajectory sequence; the initial target actual trajectory sequence R of the system a (0), R e (0) selects the data corresponding to the last several camera sampling moments in the guidance information, where R a Indicates azimuth data, R e Indicates the pitch angle data; Step 1.3: The telescope motion control system controls the telescope to track the target based on the guidance information so that it appears in the field of view of the optical camera.

3. The non-cooperative target trajectory prediction method of fusion camera miss distance and encoder information according to claim 1, characterized in that: Step 2 specifically includes: Step 2.1: At the kth optical camera sampling moment, the camera exposes and images the tracking target, identifies the image pixels and calculates the miss distance; Step 2.2: Convert the miss distance into the angular error of the telescope in the azimuth and elevation directions; let the miss distance on the imaging horizontal axis be Δx c , the miss distance on the vertical axis is Δy c , the focal length of the camera is f c , then at the kth optical camera sampling moment, that is, the kmth system sampling moment, calculate the azimuth angle error e ca , pitch angle error e ce , and input the position error into the telescope motion control unit; Step 2.3: Simultaneously read the current grating encoder measurement value, that is, the rotation angle of the telescope in the azimuth and elevation directions.

4. The non-cooperative target trajectory prediction method of fusion camera miss distance and encoder information according to claim 1, characterized in that: Step 3 specifically includes: Step 3.1: After completing the k-th optical imaging and miss distance reading, the telescope tracks the time when it enters the k-th blind zone; the camera miss distance and grating encoder information obtained at the camera sampling moment are integrated to calculate the actual position of the target, and further update the actual trajectory sequence; Step 3.2: Based on the updated actual trajectory sequence, a polynomial fitting algorithm is used to predict the motion trajectory of the non-cooperative target before the next optical imaging is completed; the actual trajectory sequence data points are respectively substituted into the polynomial model to further obtain the matrix equation; the least squares method is used to solve the optimal coefficients and obtain the polynomial model parameters, that is, the motion trajectory model is predicted based on the actual trajectory sequence; Step 3.3: Substitute the model parameters updated in step 3.2 into the polynomial model to complete the update of the prediction model; interpolate according to the encoder sampling time, substitute it into the updated polynomial model, and calculate the interpolated prediction trajectory sequence.

5. The non-cooperative target trajectory prediction method of fusion camera miss distance and encoder information according to claim 1, characterized in that: Step 4 specifically includes: Step 4.1: At the nth encoder sampling moment (n = 1, 2, …, m) in the kth blind time, read the grating encoder measurement value; Step 4.2: At the same time, the corresponding value of the interpolated predicted trajectory sequence at the current moment is subtracted from the current encoder measurement value to calculate the angular position error of the telescope in the blind time and input it into the telescope motion control unit.

6. The non-cooperative target trajectory prediction method of fusion camera miss distance and encoder information according to claim 1, characterized in that: Step 5 specifically includes: confirming whether the tracking mission is to continue, if yes, repeating steps 2 to 4; if no, ending the mission and sending a control command to shut down the telescope control system.

Images