Payload delivery system utilizing machine vision algorithms

Abstract

A system for the payload delivery utilizing a machine vision model, the system including a camera-equipped component configured to capture images of a target array for pose calculations. The payload delivery system also includes at least one adaptive thresholding and neural network algorithm configured to recognize patterns of LEDs on the target array, and calculate navigation parameters based on the recognized patterns. The payload delivery system also includes a plurality of compensation mechanisms configured to accommodate errors induced by coupling of autonomous underwater vehicle platforms.

Description

PATENT COOPERATION TREATYTITLE: PAYLOAD DELIVERY SYSTEM UTILIZING MACHINE VISION ALGORITHMSINVENTORS:Daniel Simpkins; Matthew Rubai; Gregory Eckert; Nathan Gabriel; Michael Hagerman;David E. Pyle; and Mark LandyCROSS-REFERENCE TO RELATED PATENT APPLICATIONS

[0001] This application claims priority to U.S. Appl. Serial No. 63 / 587,098, filed September 30, 2023, entitled “Payload Delivery System Utilizing Machine Vision Algorithms,” which patent application is commonly owned by the owner of the present invention. This patent application is incorporated herein in its entirety.TECHNICAL FIELD

[0002] The present disclosure relates to the field of autonomous payload delivery and recovery systems. Specifically, relates to a payload delivery system that utilizes machine vision algorithms for accurate navigation and positioning of payloads in various environmental applications.STATEMENT OF FEDERALLY FUNDED RESEARCH

[0003] None.BACKGROUND

[0004] Autonomous payload delivery and recovery systems play a crucial role in exploration, research, and industrial applications. Traditional navigation methods, relying on acoustic-based technologies like sonar and USBL, have been effective in providing relative positioning of objects in moderate distances. However, their performance rapidly deteriorates in close-range scenarios due to signal reflections and acoustic complexities. These limitations have imposed significant challenges in achieving accurate and reliable navigation for autonomous vehicles during the critical phases of payload deployment and recovery.

[0005] Machine vision has emerged to address the shortcomings of acoustic-based navigation systems. Vision-based techniques, utilizing cameras and image analysis algorithms, offer the potential to enhance navigation precision, particularly in the close proximity of structures. Existing approaches in this field have often integrated multiple sensors alongside cameras to compensate for limitations. However, these hybrid systems can introduce complexities, synchronization challenges, and uncertainties stemming from the fusion of disparate data sources.

[0006] Furthermore, certain environments, such as underwater environments as a non-limiting example, pose unique optical challenges that can hinder traditional machine vision algorithms. Light scattering and absorption in environments, along with the presence of suspended particles, can limit the effective range and quality of images captured by cameras. Additionally, the inherent instability of vehicles in certain environmental settings, such as for example in underwater environments, and the potential for dynamic loading during coupling operations introduce new complexities in maintaining accurate positioning and trajectory control.

[0007] The limitations of conventional methods have highlighted the need for an integrated and robust system that overcomes the challenges of accurate navigation and payload operations in close-range underwater scenarios. Accordingly, there exists a clear demand for a solution that bridges the gap between long-range acoustic navigation and the intricacies of close-proximity maneuvers required for autonomous payload delivery and recovery.SUMMARY OF THE DISCLOSURE

[0008] The present disclosure addresses these challenges by introducing a payload delivery system that leverages machine vision algorithms to enable precise navigation, pose estimation, and autonomous operations in previously difficult to navigate environments. By utilizing camera-based vision systems and innovative algorithms, this invention opens the door to accurate and reliable navigation in the critical phase of coupling autonomous vehicle platforms and payloads, thus expanding the capabilities and applications of underwater autonomous systems.

[0009] To address the need disclosed above, the present disclosure provides a transformative solution for enhancing navigation accuracy and payload operations. By leveraging adaptive thresholding, neural network algorithms, and innovative compensation mechanisms, the present system revolutionizes autonomous payload delivery and recovery processses.

[0010] The present disclosure allows for precise autonomous navigation, extended range navigation, robustness and adaptability, accurate path error detection, a simplified and unified approach to payload delivery, enhanced close range navigation, cost-efficiency, and autonomy.

[0011] The payload delivery system utilizing machine vision algorithms disclosed and described herein represents a paradigm shift in autonomous navigation and payload operations. Its fusion of adaptive thresholding, neural network techniques, and compensation mechanisms ensures accurate navigation, robustness in challenging conditions, and the potential for unprecedented achievements in underwater autonomy.

[0012] In general, in one embodiment, the disclosure features a payload delivery system, the system includes a camera-equipped component configured to capture images of a target array for pose calculations. The payload delivery system also includes at least one adaptive thresholding and neural network algorithm. The at least one adaptive thresholding and neural network algorithm are configured to recognize patterns of LEDs on the target array. The at least one adaptive thresholding and neural network algorithm are also configured to calculate navigation parameters based on the recognized patterns. The payload delivery system also includes compensationmechanisms configured to accommodate errors induced by coupling of autonomous vehicle platforms.

[0013] In general, in another embodiment, the disclosure features a machine vision model for recognizing patterns of LEDs on a target array in environments. The machine vision model includes a trained neural network algorithm. The trained neural network algorithm is capable of detecting the target LEDs in the image. The machine vision model also includes sorting and filtering mechanisms configured to eliminate noisy predictions based on relative LED distances and prediction confidences. The machine vision model also includes ordering algorithms configured to arrange the unordered LED positions into a mesh of points. The machine visions model also includes calibration methods to transform camera-view positions to real-world coordinates relative to the camera.BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Other advantages of the present disclosure will be apparent from the following detailed description of the disclosure in conjunction with embodiments as illustrated in the accompanying drawings, in which:

[0015] FIG. 1 depicts a block diagram model of a payload delivery and recovery system utilizing a machine vision model, in accordance with certain embodiments of the present disclosure.

[0016] FIG. 2 depicts a block diagram of the payload delivery and recovery system operating with a payload, in accordance with certain embodiments of the present disclosure.NOTATION AND NOMENCLATURE

[0017] Various terms are used to refer to particular system components. Different companies may refer to a component by different names - this document does not intend to distinguish between components that differ in name but not function. In the following discussion and in the claims, the terms “including” and “comprising” are used in an open-ended fashion, and thus should be interpreted to mean “including, but not limited to . . . .” Also, the term “couple” or “couples” is intended to mean either an indirect or a direct connection. Thus, if a first device couples to a second device, that connection may be through a direct connection or through an indirect connection via other devices and connections.

[0018] The terminology used herein is for the purpose of describing particular example embodiments only, and is not intended to be limiting. Following long-standing patent law convention, the terms “a” and “an” mean “one or more” when used in this application, including the claims.

[0019] As used herein, the singular forms “a,” “an,” and “the” may be intended to include the plural forms as well, unless the context clearly indicates otherwise. The method steps, processes, and operations described herein are not to be construed as necessarily requiring their performance in the particular order discussed or illustrated, unless specifically identified as an order of performance. It is also to be understood that additional or alternative steps may be employed.

[0020] The terms first, second, third, etc. may be used herein to describe various elements, components, regions, layers and / or sections; however, these elements, components, regions, layers and / or sections should not be limited by these terms. These terms may be only used to distinguish one element, component, region, layer, or section from another region, layer, or section. Terms such as “first,” “second,” and other numerical terms, when used herein, do not imply a sequenceor order unless clearly indicated by the context. Thus, a first element, component, region, layer, or section discussed below could be termed a second element, component, region, layer, or section without departing from the teachings of the example embodiments. The phrase “at least one of,” when used with a list of items, means that different combinations of one or more of the listed items may be used, and only one item in the list may be needed. As used herein, the term “and / or” when used in the context of a listing of entities, refers to the entities being present singly or in combination. Thus, for example, the phrase “A, B, C, and / or D” includes A, B, C, and D individually, but also includes any and all combinations and subcombinations of A, B, C, and D. Accordingly, as an example, “at least one of: A, B, and C” includes any of the following combinations: A, B, C, A and B, A and C, B and C, and A and B and C. In another example, the phrase “one or more” when used with a list of items means there may be one item or any suitable number of items exceeding one.

[0021] Spatially relative terms, such as “inner,” “outer,” “beneath,” “below,” “lower,” “above,” “upper,” “top,” “bottom,” and the like, may be used herein. These spatially relative terms can be used for ease of description to describe one element’s or feature’s relationship to another element(s) or feature(s) as illustrated in the figures. The spatially relative terms may also be intended to encompass different orientations of the device in use, or operation, in addition to the orientation depicted in the figures. For example, if the device in the figures is turned over, elements described as “below” or “beneath” other elements or features would then be oriented “above” the other elements or features. Thus, the example term “below” can encompass both an orientation of above and below. The device may be otherwise oriented (rotated 90 degrees or at other orientations) and the spatially relative descriptions used herein interpreted accordingly.

[0022] Unless otherwise indicated, all numbers expressing quantities of ingredients, reaction conditions, and so forth used in the specification are to be understood as being modified in all instances by the term “about.” Accordingly, unless indicated to the contrary, the numerical parameters set forth in this specification are approximations that can vary depending upon the desired properties sought to be obtained by the presently disclosed subject matter.DETAILED DESCRIPTION OF THE DISCLOSURE

[0023] The present disclosure is directed towards a payload delivery system that employs machine vision algorithms to address the shortcomings of conventional acoustic-based navigation and enhance autonomous payload operations in underwater environments. The system integrates a camera-equipped component capable of capturing images of a target array for pose calculations. This system overcomes the limitations of acoustic methods and provides robust navigation even in close-range situations.

[0024] The present disclosure allows for precise autonomous navigation. The payload delivery system overcomes the limitations of traditional acoustic-based navigation methods by introducing machine vision algorithms capable of recognizing patterns of LEDs on a target array. This recognition process, based on neural network techniques, allows for accurate pose calculations and navigation even in close-range scenarios where acoustic techniques falter.

[0025] The present disclosure allows for extended range navigation. The incorporation of active light patterns, such as specifically selected LED arrays, allows for navigation at extended ranges, surpassing the capabilities of conventional vision-based systems. Additionally, the system's ability to employ multiple LED patterns in different color bands, including blue LEDs for longer range and red LEDs for close proximity, enhances its versatility in various underwater environments.

[0026] The present disclosure allows for robustness and adaptability. The system's robustness is evident in its capacity to tolerate missing or additional pattern features, providing resilience in challenging conditions. The integration of adaptive thresholding and neural network techniques ensures accurate recognition of target patterns, even when faced with optical challenges posed by underwater environments.

[0027] The present disclosure allows for accurate path error detection. The invention's compensation mechanisms effectively account for errors induced by the coupling of autonomous underwater vehicle platforms. This capability allows the system to detect deviations from the intended path and autonomously adjust, ensuring successful coupling and payload operations.

[0028] The present disclosure allows for a simplified and unified approach. Unlike traditional methods that often rely on a combination of external sensors, the proposed system employs camera footage as the sole input for navigation and pattern recognition. This simplification streamlines the system's architecture while eliminating synchronization and integration complexities.

[0029] The present disclosure allows for enhanced close-range navigation. The system's ability to navigate in environments where acoustic techniques are typically limited is a groundbreaking feature. The integration of machine vision techniques directly addresses the challenges posed by reflections and signal interference in close-range scenarios.

[0030] The present disclosure allows for cost-efficiency and autonomy. By enabling accurate autonomous operations, the invention reduces the reliance on costly manned interventions for payload delivery and recovery tasks. This enhancement is poised to revolutionize underwater exploration, research, and industrial applications.

[0031] FIG. 1 depicts a payload delivery and recovery system utilizing a machine vision model, in accordance with certain embodiments of the present disclosure.

[0032] As shown in FIG. 1, the payload delivery and recovery system 100, also referred to herein as the payload delivery system, can contain a flying lead component 102 that includes a camera system 104. The camera system 104, in certain embodiments, may include a camera, one or more lenses, a power system, and a communications system. Additionally, the camera system 104 may include or be operatively coupled to a vision system 106 that includes both target models 108 and machine vision algorithms 110. In some embodiments, and as depicted in FIG. 1, the machine vision algorithms 110 may be configured for video capture 112, image pre-processing 114, pattern locating 116, pattern identification 118, image post-processing 120, and image transformation 122 to provide coordinates of the payload or payload delivery area in an observed environment.

[0033] In some embodiments, the payload delivery system employs adaptive thresholding and neural network techniques to recognize patterns of LEDs on the target array. These patterns, generated with specifically selected LED arrays, can be utilized for extended-range navigation. Different color bands, such as blue LEDs for longer range and red LEDs for close proximity, can be employed for diverse navigational requirements. The system compensates for external loading and errors induced by the coupling of autonomous underwater vehicle platforms, allowing for accurate navigation and path error detection.

[0034] In certain embodiments, the payload delivery system combines an innovative architecture for payload deployment and recovery with machine vision algorithms for enhanced navigation. The architecture includes a camera-equipped component that aids in finding and navigating to the payload. Adaptive thresholding and neural network techniques are employed to recognize patterns of LEDs on the target array. This allows for accurate pose calculations and navigation.

[0035] The architecture of the autonomous payload delivery system may include a specialized component, known as a “flying lead” component, equipped with a camera system. In someembodiments, this flying lead component may be used to accurately locate and navigate to the target payload. In some embodiments, the flying lead component may be autonomous. The camera system can be configured with suitable resolution and optics to capture clear images of the target array for subsequent pose calculations.

[0036] The flying lead component, in some embodiments, can operate in tandem with another autonomous vehicle, allowing the system to perform coupled operations. In certain embodiments, while the autonomous vehicle maintains station-keeping during the operation, the may have its own inherent motions. The flying lead component, in such an embodiment, is responsible for navigating to the payload and directing its specific maneuvers. The challenges posed by the coupling of two autonomous vehicle platforms can be addressed by the system's compensation mechanisms, which can anticipate and accommodate errors induced by this coupling process. Accordingly, such an embodiment allows for accurate navigation and path error detection, ensuring successful coupling and payload operations.

[0037] In certain embodiments, the payload delivery system includes pattern recognition and navigation algorithms. The machine vision algorithms integrated into the payload delivery system can be central to the navigation and pose estimation capabilities. For example, in certain embodiments, the machine vision algorithms can enable the system to recognize patterns of LEDs on the target array and calculate navigation parameters based on the identified patterns. The recognition process leverages both adaptive thresholding and advanced neural network techniques to ensure robust and accurate pattern detection.

[0038] Specifically, in some embodiments, the machine vision algorithms focus on recognizing patterns of LEDs on the target array. These patterns can be generated by employing specifically selected LED arrays that emit active, non-symmetric light patterns optimized for transmission, forexample but not limited to underwater transmission. The choice of LED colors, such as blue LEDs for longer range navigation and red LEDs for close-in navigation, can allow for versatility in the payload delivery system's capabilities in different environments.

[0039] The machine vision algorithms may apply adaptive thresholding techniques to filter excessive signals and isolate the centroids of LEDs on the target array. This process can be utilized in some embodiments to remove noise and enhance the quality of detected patterns. The neural network component of the algorithms can be utilized for pattern recognition. In some embodiments the payload delivery system may employ convolutional neural networks (CNNs) that have multiple layers, including convolutional, pooling, and fully connected layers. These layers, in certain embodiments, process the image data captured by the camera system and identify LED patterns, enabling accurate recognition and navigation.

[0040] The algorithms may also exhibit adaptability to tolerate missing or additional pattern features. In some embodiments where there are imperfect conditions or partial visibility, the robustness can enhance the payload delivery system's performance.

[0041] The payload delivery system can significantly enhance close-range navigation, a critical capability for successful payload deployment and recovery operations. Conventional acousticbased navigation systems encounter limitations in close proximity due to signal reflections and acoustic complexities in certain environments. In the payload delivery system, however, the machine vision algorithms employed can bridge this gap by directly addressing the challenges of accurate navigation in such environments.

[0042] By utilizing camera-based vision systems and innovative algorithms, in some embodiments, the payload delivery system can provide accurate navigation and payload operations in close-range scenarios. The fusion of adaptive thresholding and neural network techniques candirectly counter the limitations of acoustic techniques, which in some embodiments allows the system to navigate and position itself with unprecedented accuracy in the immediate vicinity of the target payload.

[0043] In some embodiments, the payload delivery system may use a simplified and unified approach to navigation and payload operations. Unlike traditional methods that often rely on a combination of external sensors, the payload delivery system may employ camera footage as the sole input for navigation and pattern recognition. In the embodiment of this approach, the need for synchronization and integration complexities associated with hybrid sensor systems can be reduced or entirely eliminated.

[0044] In some embodiments, by relying on camera-based vision, the payload delivery system may have a streamlined architecture, reducing both cost and complexity. This simplicity, combined with the system's adaptability to varying underwater conditions can allow for efficient and effective autonomous payload operations.

[0045] In some embodiments, the payload delivery system can include components such as a camera-equipped component designed to capture images of a target array, facilitating subsequent pose calculations. In certain embodiments, the system can incorporate adaptive thresholding and neural network algorithms that, for example, jointly fulfill two main functions: recognizing patterns of LEDs on the target array and calculating navigation parameters based on the recognized patterns. Furthermore, in some embodiments, a plurality of compensation mechanisms are integrated into the system, adept at accommodating errors arising from the coupling of autonomous underwater vehicle platforms.

[0046] In certain embodiments, the payload delivery system can leverage adaptive thresholding techniques inherent within the camera-equipped component. This, for example, enables the systemto effectively filter excessive signals and precisely isolate the centroids of LEDs present on the target array.

[0047] In some embodiments, the camera-equipped component takes advantage of a neural network algorithm, enhancing pattern recognition capabilities and, consequently, navigation accuracy within the system. Similarly, in certain embodiments, the target array utilized in the payload delivery system is equipped with specifically selected LED arrays. These arrays are engineered, for example, to generate active light patterns optimized for transmission underwater.

[0048] In some embodiments, the target array's multiple LED patterns, for example, may differ in color bands. These configurations are specifically intended to, for example, facilitate navigation across varying distances within the underwater environment.

[0049] In some embodiments, the payload delivery system employs multi-frequency patterns. More specifically, various LED arrangements can be operated or identified in a sequence or based on specific conditions. For example, the payload delivery system is configured to modulate LEDs at diverse rates within a pattern. This modulation not only differentiates LEDs but also facilitates the simultaneous transmission of multiple signals. Additionally, the payload delivery system incorporates LEDs of varying colors, each corresponding to a distinct frequency on the electromagnetic spectrum. Moreover, the payload delivery system’s use of machine vision enables that layered patterns, be it through physical arrangement and / or modulation, are accurately recognized, allowing for precise navigation and pattern differentiation in complex scenarios and environments.

[0050] In certain embodiments, the payload delivery system incorporates compensation mechanisms adept at accounting for external loading and deviations stemming from the coupling of autonomous vehicle platforms. This ensures precise navigation and the ability to, for example,detect path errors and autonomously restart operations upon coupling failure. The compensation mechanisms, in some embodiments, are designed not only to accommodate external loading but also to recognize deviations induced by the coupling of autonomous vehicle platforms. This capability enables, for example, path error detection and facilitates an autonomous restart in the event of a coupling failure.

[0051] In some embodiments, the machine vision model may be tailored to recognize patterns of LEDs on a target array within specific environments. This model comprises, in certain embodiments, a trained neural network algorithm responsible for detecting LEDs in an image, even when they are located in unordered LED positions.

[0052] Additionally, in some embodiments, the model integrates sorting and filtering mechanisms, eliminating predictions based on relative LED distances and prediction confidences. In certain embodiments, ordering algorithms are also present within the model, arranged to organize unordered LED positions into a mesh of points. Furthermore, the model incorporates, for example, at least one calibration method that transforms camera-view positions into real-world coordinates relative to the camera.

[0053] Elaborating on the machine vision model, the trained neural network algorithm can possess the capability to distinguish individual LED predictions and associate them with their respective matrix predictions.

[0054] In certain embodiments, the machine vision model may include a plurality of ordering algorithms that are responsible for rotating the mesh of target points. This rotation, for example, seeks to identify the orientation that minimizes distance errors between grid points and the LEDs present in the target array.

[0055] Further, in certain embodiments, the calibration methods incorporated within iteratively fine-tune real-world coordinates and rotations. The iterative optimization process in such an embodiment can leverage the known dimensions of the target, further enhancing the precision of pose estimation.

[0056] The machine vision model can be utilized as an integral part of a payload delivery system. In this role, it contributes to accurate navigation and reliable pose estimation, thus facilitating, for example, the smooth execution of autonomous payload operations. In certain embodiments, the machine vision model exclusively utilizes captured camera footage as its primary input for pattern recognition and navigation calculations. This design choice eliminates the need for external sensors in the pattern recognition process.

[0057] The machine vision model employs, in some embodiments, convolutional neural networks (CNNs). This utilization enhances the detection of target LEDs, resulting in improved accuracy and operational efficiency. The machine vision model integrates, in certain embodiments, convolutional neural networks (CNNs), featuring multiple layers including convolutional, pooling, and fully connected layers. These layers collectively process image data and effectively identify patterns of LEDs on the target array.

[0058] In certain embodiments, the plurality of ordering algorithms responsible for arranging unordered LED positions into a mesh of points can utilize optimization techniques, which may be geared towards minimizing distance errors and subsequently enhancing accuracy in the mesh arrangement.

[0059] Expanding upon the machine vision model, in some embodiments, the calibration methods employed within encompass a multi-step refinement process utilizing iterative optimizationalgorithms. This complex process leads to heightened precision in pose estimation, a crucial aspect of the model's functionality.

[0060] FIG. 2 depicts a block diagram of the payload delivery and recovery system operating with a payload, in accordance with certain embodiments of the present disclosure.

[0061] As shown in FIG. 2, the payload delivery and recovery system 100 may be utilized based on the configurations and components detailed above to identify a payload package module 200 situated on, in, around, or adjacent to a package for recovery or package area for delivery. Through the use of the flying lead 102 containing the vision system 110, the payload delivery and recovery system 100 can identify an active optical target 202 and a passive optical target 204 in the payload package module.

[0062] The techniques introduced here may be implemented by, for example, programmable circuitry (e.g., one or more microprocessors), programmed with software and / or firmware, entirely in special-purpose hardwired (i.e., non-programmable) circuitry, or in a combination or such forms. Special -purpose circuitry can be in the form of, for example, one or more applicationspecific integrated circuits (ASICs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), etc.

[0063] Note that any of the embodiments described above can be combined with another embodiment, except to the extent that it may be stated otherwise above or to the extent that any such embodiments might be mutually exclusive in function and / or structure.

[0064] Although the present invention has been described with reference to specific exemplary embodiments, it will be recognized that the invention is not limited to the embodiments described, but can be practiced with modification and alteration within the spirit and scope of the appendedclaims. The specification and drawings are to be regarded in an illustrative sense rather than a restrictive sense.

[0065] The foregoing description, for purposes of explanation, used specific nomenclature to provide a thorough understanding of the described embodiments. However, it should be apparent to one skilled in the art that the specific details are not required in order to practice the described embodiments. Thus, the foregoing descriptions of specific embodiments are presented for purposes of illustration and description. They are not intended to be exhaustive or to limit the described embodiments to the precise forms disclosed. It should be apparent to one of ordinary skill in the art that many modifications and variations are possible in view of the above teachings.

[0066] While embodiments of the disclosure have been shown and described, modifications thereof can be made by one skilled in the art without departing from the spirit and teachings of the disclosure. The embodiments described and the examples provided herein are exemplary only, and are not intended to be limiting. Many variations and modifications of the disclosure disclosed herein are possible and are within the scope of the disclosure. The scope of protection is not limited by the description set out above, but is only limited by the claims which follow, that scope including all equivalents of the subject matter of the claims.

[0067] Amounts and other numerical data may be presented herein in a range format. It is to be understood that such range format is used merely for convenience and brevity and should be interpreted flexibly to include not only the numerical values explicitly recited as the limits of the range, but also to include all the individual numerical values or sub-ranges encompassed within that range as if each numerical value and sub-range is explicitly recited. For example, a numerical range of approximately 1 to approximately 4.5 should be interpreted to include not only the explicitly recited limits of 1 to approximately 4.5, but also to include individual numerals such as2, 3, 4, and sub-ranges such as 1 to 3, 2 to 4, etc. The same principle applies to ranges reciting only one numerical value, such as “less than approximately 4.5,” which should be interpreted to include all of the above-recited values and ranges. Further, such an interpretation should apply regardless of the breadth of the range or the characteristic being described. The symbolis the same as “approximately”.

[0068] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood to one of ordinary skill in the art to which the presently disclosed subject matter belongs. Although any methods, devices, and materials similar or equivalent to those described herein can be used in the practice or testing of the presently disclosed subject matter, representative methods, devices, and materials are now described.

[0069] The above discussion is meant to be illustrative of the principles and various embodiments of the present disclosure. Numerous variations and modifications will become apparent to those skilled in the art once the above disclosure is fully appreciated. It is intended that the following claims be interpreted to embrace all such variations and modifications.

[0070] Those skilled in the art will appreciate that the steps described herein may be carried out in a variety ways and that no particular ordering is required. It will be further understood from the foregoing description that modifications and changes may be made in various embodiments of the present disclosure without departing from its true spirit. The descriptions in this specification are for purposes of illustration only and are not to be construed in a limiting sense.

[0071] Consistent with the above disclosure, the examples of systems and methods enumerated in the following clauses are specifically contemplated and are intended as a non-limiting set of examples.

[0072] Clause 1. A payload delivery system including a camera-equipped component configured to capture images of a target array for pose calculations; at least one adaptive thresholding and neural network algorithm configured to: recognize patterns of LEDs on the target array, and calculate navigation parameters based on the recognized patterns; and a plurality of compensation mechanisms configured to accommodate errors induced by coupling of autonomous underwater vehicle platforms.

[0073] Clause 2. The payload delivery system of any foregoing clause, where the camera-equipped component is a camera-equipped, autonomous flying lead component.

[0074] Clause 3. The payload delivery system of any foregoing clause, where the camera-equipped component utilizes adaptive thresholding techniques to filter excessive signals and isolate centroids of LEDs on the target array.

[0075] Clause 4. The payload delivery system of any foregoing clause, where the camera-equipped component incorporates a neural network algorithm to enhance pattern recognition and navigation accuracy.

[0076] Clause 5. The payload delivery system of any foregoing clause, where the target array includes specifically selected LED arrays capable of generating active light patterns optimized for underwater transmission.

[0077] Clause 6. The payload delivery system of any foregoing clause, where the target array includes multiple LED patterns.

[0078] Clause 7. The payload delivery system of any foregoing clause, where the multiple LED patterns are in different color bands configured to enable navigation across varying distances.

[0079] Clause 8. The payload delivery system of any foregoing clause, where the compensation mechanisms account for external loading and deviations induced by the coupling of autonomous vehicle platforms.

[0080] Clause 9. The payload delivery system of any foregoing clause, where the machine vision algorithms further enable path error detection and autonomous restart upon coupling failure.

[0081] Clause 10. A machine vision model for recognizing patterns of LEDs on a target array in environments including: a trained neural network algorithm configured to detecting the LEDs in an image, wherein the LEDs are present in unordered LED positions; a plurality of sorting and filtering mechanisms configured to eliminate predictions, wherein the elimination of predications is based on one or more relative LED distances and one or more prediction confidences; a plurality of ordering algorithms configured to arrange the unordered LED positions into a mesh of points; and at least one calibration method to transform camera-view positions to real-world coordinates relative to the camera.

[0082] Clause 11. The machine vision model of any foregoing clause, where the trained neural network algorithm is capable of distinguishing individual LED predictions and associating them with respective matrix predictions.

[0083] Clause 12. The machine vision model of any foregoing clause, where the plurality of ordering algorithms rotate the mesh of target points to find the orientation that minimizes distance errors between grid points and LEDs.

[0084] Clause 13. The machine vision model of any foregoing clause, where the at least one calibration method iteratively fine-tunes real-world coordinates and rotations using the known dimensions of the target.

[0085] Clause 14. The machine vision model of any foregoing clause, where the machine vision model is employed as part of a payload delivery system to enable accurate navigation and pose estimation for autonomous payload operations.

[0086] Clause 15. The machine vision model of any foregoing clause, where the machine vision model utilizes the captured camera footage as the input for pattern recognition and navigation calculations.

[0087] Clause 16. The machine vision model of any foregoing clause, where the trained neural network algorithm employs convolutional neural networks (CNNs) to detect the target LEDs with improved accuracy and efficiency.

[0088] Clause 17. The machine vision model of any foregoing clause, where the convolutional neural networks (CNNs) consist of multiple layers, including convolutional, pooling, and fully connected layers, to process the image data and identify LED patterns.

[0089] Clause 18. The machine vision model of any foregoing clause, where the plurality of ordering algorithms for arranging the unordered LED positions into the mesh of points is configured to utilize optimization techniques to minimize distance errors and enhance accuracy.

[0090] Clause 19. The machine vision model of any foregoing clause, where the calibration methods for transforming camera-view positions to real-world coordinates incorporate a multi- step refinement process using iterative optimization algorithms, resulting in increased precision in pose estimation.

Claims

CLAIMSWHAT IS CLAIMED IS:

1. A payload delivery system comprising:(a) a camera-equipped component configured to capture images of a target array for pose calculations;(b) at least one adaptive thresholding and neural network algorithm configured to:(i) recognize patterns of LEDs on the target array, and(ii) calculate navigation parameters based on the recognized patterns; and(c) a plurality of compensation mechanisms configured to accommodate errors induced by coupling of autonomous underwater vehicle platforms.

2. The payload delivery system of Claim 1, wherein the camera-equipped component is a camera-equipped, autonomous flying lead component.

3. The payload delivery system of Claim 1, wherein the camera-equipped component utilizes adaptive thresholding techniques to filter excessive signals and isolate centroids of LEDs on the target array.

4. The payload delivery system of Claim 1, wherein the camera-equipped component incorporates a neural network algorithm to enhance pattern recognition and navigation accuracy.

5. The payload delivery system of Claim 1, wherein the target array includes specifically selected LED arrays capable of generating active light patterns optimized for underwater transmission.

6. The payload delivery system of claim 1, wherein the target array includes multiple LED patterns.

7. The payload delivery system of Claim 6, wherein the multiple LED patterns are in different color bands configured to enable navigation across varying distances.

8. The payload delivery system of Claim 1, wherein the compensation mechanisms account for external loading and deviations induced by the coupling of autonomous vehicle platforms.

9. The payload delivery system of Claim 1, wherein the machine vision algorithms further enable path error detection and autonomous restart upon coupling failure.

10. A machine vision model for recognizing patterns of LEDs on a target array in environments, comprising:(a) a trained neural network algorithm configured to detecting the LEDs in an image, wherein the LEDs are present in unordered LED positions;(b) a plurality of sorting and filtering mechanisms configured to eliminate predictions, wherein the elimination of predications is based on one or more relative LED distances and one or more prediction confidences;(c) a plurality of ordering algorithms configured to arrange the unordered LED positions into a mesh of points; and(d) at least one calibration method to transform camera-view positions to real-world coordinates relative to the camera.

11. The machine vision model of Claim 10, wherein the trained neural network algorithm is capable of distinguishing individual LED predictions and associating them with respective matrix predictions.

12. The machine vision model of Claim 10, wherein the plurality of ordering algorithms rotate the mesh of target points to find the orientation that minimizes distance errors between grid points and LEDs.

13. The machine vision model of Claim 10, wherein the at least one calibration method iteratively fine-tunes real-world coordinates and rotations using the known dimensions of the target.

14. The machine vision model of Claim 10, wherein the machine vision model is employed as part of a payload delivery system to enable accurate navigation and pose estimation for autonomous payload operations.

15. The machine vision model of Claim 10, wherein the machine vision model utilizes the captured camera footage as the input for pattern recognition and navigation calculations.

16. The machine vision model of Claim 10, wherein the trained neural network algorithm employs convolutional neural networks (CNNs) to detect the target LEDs with improved accuracy and efficiency.

17. The machine vision model of Claim 16, wherein the convolutional neural networks (CNNs) consist of multiple layers, including convolutional, pooling, and fully connected layers, to process the image data and identify LED patterns.

18. The machine vision model of Claim 10, wherein the plurality of ordering algorithms for arranging the unordered LED positions into the mesh of points is configured to utilize optimization techniques to minimize distance errors and enhance accuracy.

19. The machine vision model of Claim 10, wherein the calibration methods for transforming camera-view positions to real-world coordinates incorporate a multi-step refinement process using iterative optimization algorithms, resulting in increased precision in pose estimation.

Images