Method and system for compensating for deviations in multiple working postures of a manipulator in an autonomous charging device

Abstract

A method for minimizing the deviation between a target attitude and an actual attitude of an autonomous charging device manipulator, wherein the manipulator is adapted to connect a charging connector to an electric vehicle socket, the method comprising obtaining a compensation function that covers at least a portion of the working space of the manipulator, the compensation function utilizing a relationship between at least one motion command and a corresponding actual attitude, and operating the manipulator based on the obtained compensation function such that the resulting actual attitude is minimized from the target attitude, wherein the relationship between at least one motion command and a corresponding actual attitude is obtained by a method comprising: attaching an image acquisition device to the movable part of the manipulator to acquire an image including a visible reference marker positioned in a reference attitude; commanding the manipulator to move and acquiring an image of the reference marker; using the image to determine the relative attitude of the reference marker with respect to the image acquisition device; calculating the actual attitude of the movable part of the manipulator using the reference attitude and the determined relative attitude of the reference marker with respect to the image acquisition device; and fitting a template function that associates at least one motion command with the calculated actual attitude.

Description

Detailed Description of the Invention

[0001] [Summary] The present invention relates to a method for compensating for deviations in multiple working postures of a manipulator of an autonomous charging device (ACD), the ACD being configured to connect a charging connector into a vehicle's charging socket.

[0002] [Technical Field of the Invention] The present invention generally relates to an electric autonomous charging device, related methods and systems that enable autonomous charging of electric vehicles, and the system and method utilize artificial intelligence and machine learning. The methods and systems according to the present invention aim to improve the performance and reliability of an autonomous charging device (ACD). In particular, the methods and systems of the present invention enable at least compensating for deviations in one or more working postures of a manipulator of the ACD, the ACD being configured to connect a charging connector into a vehicle's charging socket.

[0003] [Background of the Invention] As the adoption of electric vehicles increases, the demand for efficient and autonomous charging solutions has significantly increased. Autonomous robot systems for automating the charging process, eliminating the need for human intervention, have been developed. These systems typically utilize computer vision techniques to detect the charging port of an electric vehicle and navigate towards it.

[0004] To ensure a successful connection between the connector and socket, the robot must accurately determine the position of the EV and / or EV socket and align the connector in a continuous and reliable manner. Several components may contribute to unsuccessful plugging and undesirable operational performance, and therefore may require further improvement. Various solutions have been proposed to ensure a precise connection between the connector and socket, including the use of magnetic coupling. However, many known techniques require modifications to either the EV or the connector, which can affect the scalability and availability of solutions in the market.

[0005] Recent advances in ACD technology have brought improvements to identifying the location of electric vehicle sockets and guiding charger connectors to charging ports without requiring human intervention or modifications to the vehicle or its charging ports. These improvements rely on the use of compliant electric robotic mechanisms, computer vision, and neural networks to accurately identify the vehicle or its charging ports, enabling ACD to complete plug-in and plug-out processes with greater reliability.

[0006] Devices for this purpose are known in the art from, for example, the international patent applications PCT / NL2020 / 050266, PCT / NL2021 / 050115, PCT / NL2021 / 050410, PCT / NL2021 / 050495, PCT / NL2021 / 05061, and PCT / EP2022 / 062233 from the same applicant, all of which are incorporated herein by reference. The ACDs according to this disclosure may include compliance mechanisms to ensure a safer plug-in process and are configured to support several connectors, including, but not limited to, Type 1, Type 2, CCS-1, CCS-2, CHAdeMO, Tesla, and MCS connectors.

[0007] Ensuring precise alignment within the socket before plugging in the connector is a persistent challenge in achieving successful mating. The accuracy required for this alignment is within a few millimeters, and this accuracy must be met by the Autonomous Charger (ACD) rather than relying on vehicle alignment, as vehicles may not automatically align to such precision. The Autonomous Charger (ACD) comprises various components and may have inherent inaccuracies due to the nature of robotic kinematics. These inaccuracies can arise from either the hardware or software components of the ACD.

[0008] Autonomous charging devices, such as those described in the current state of technology, include a manipulation system. The ACD may include at least one image acquisition device, such as a camera, and a physical manipulator, such as an electric robotic arm, capable of reaching into a socket in the vehicle and performing mating of the connector with the socket. The ability of the robotic arm to manipulate the connector with sufficient precision is important to the overall performance of the ACD. In particular, the robotic arm should preferably be able to perform with high precision throughout its entire workspace. This precision may depend on a proper correlation between the coordinate system of the camera and the coordinate system of the robotic arm.

[0009] Existing autonomous robot systems face the challenge of precisely aligning the connector of the robot arm manipulator with the charging port on an electric vehicle. Misalignments between the connector and the charging port can occur due to various factors, including imprecise positioning and inherent system uncertainties. These inaccuracies can result in connection attempts failing, interrupted charging sessions, and potentially damage to the charging infrastructure or the electric vehicle.

[0010] Existing autonomous robot systems often face challenges in achieving the desired posture of the robot arm manipulator during the charging process. Inaccuracies in the kinematic control of the robot arm, such as joint position errors, calibration uncertainties, and mechanical tolerances, can lead to discrepancies between the intended and actual positions of the connector. These discrepancies interfere with the successful engagement of the connector with the charging port and prevent the establishment of a secure and reliable electrical connection.

[0011] Because the charging environment is subject to dynamic changes, external factors also contribute to this technical problem. Factors such as vibration, ambient temperature fluctuations, and physical obstacles can further impair the accuracy of the robot arm's movements. These external influences introduce further uncertainty, affecting connector alignment and positioning, and subsequently hindering the success of the charging process.

[0012] The software controlling the autonomous robot system plays a crucial role in addressing this technical challenge. However, limitations in the software algorithms used for trajectory planning, motion control, and path optimization can lead to inaccuracies in the robot arm's movements. Improper sensor fusion, insufficient feedback control, or delays in data processing can result in deviations from the desired target posture or trajectory, potentially affecting alignment with the connector's charging port.

[0013] Furthermore, the computer vision algorithms used to recognize the charging environment and detect charging ports may be sensitive to inaccuracies. Challenges such as obstruction, reflection, variations in lighting conditions, and the presence of other objects can hinder reliable detection and recognition of charging ports, potentially affecting the precise positioning of the robotic arm and connectors.

[0014] Therefore, there is a need for improved autonomous robotic systems that address technical problems by mitigating inaccuracies related to the robotic arm, its kinematic mode, external factors, software, and computer vision, among other things.

[0015] Subsequent sections of this patent application provide a detailed description of features and mechanisms developed to overcome these challenges and ensure accurate alignment and successful plugging of the connector into the charging port of an electric vehicle.

[0016] [Overview of the prefecture] This invention provides a method that overcomes the limitations of previous methods and offers additional advantages.

[0017] In one embodiment, relating to a method for minimizing the deviation between the target and actual attitudes of the ACD manipulator, the manipulator is adapted to connect the charging connector into a socket in an electric vehicle.

[0018] In one embodiment, relating to a system for minimizing the deviation between the target attitude and the actual attitude of an ACD manipulator, the manipulator is adapted to connect a charging connector into a socket in an electric vehicle. [Brief explanation of the drawing]

[0019] [Figure 1] This is a reference diagram showing the ACD (1) that supports the connection of the charger connector (2) to the socket (4) of the electric vehicle (5). [Figure 2] This is a schematic diagram of the autonomous charging device (1). A robotic arm (10), motion control unit (20), computer vision unit (30), control system (40), and image acquisition device (50) are depicted. [Figure 3] This is a flowchart showing the steps of the method according to the first embodiment of the present invention. [Modes for carrying out the invention]

[0020] [A brief explanation of conventional technology] Some existing systems have various shortcomings for specific applications. Therefore, further contributions in this field are still needed.

[0021] International patent application WO2020142496A1 describes a method for training a robot coupled with a camera, the process of setting parameters for the robot or camera; using the robot or camera parameters to capture training images of a training object using the camera; changing the settings to capture different training images and repeating such settings and captures to obtain multiple training images based on different settings; training the system to recognize a training object based on the multiple training images; and evaluating the system using pre-selected test images. This document describes training an industrial robot intended to perform a manufacturing process and does not address methods for compensating for deviations in multiple working postures of a manipulator in an autonomous charging device.

[0022] [Detailed description of the invention] In the following description, certain details, such as specific methods, steps, apparatus, and components, are described for illustrative purposes only, not limiting purposes, in order to provide a complete understanding of the invention. However, it will be apparent to those skilled in the art that the invention can be practiced in other embodiments departing from these specific details.

[0023] In one embodiment, the present invention relates to a method for compensating for deviations in multiple working postures of a manipulator of an autonomous charging device (ACD), wherein the ACD is configured to connect a charging connector to a charging socket in a vehicle.

[0024] autonomous charging device An autonomous charging device (or simply a robot) as referred to herein may comprise several hardware components and software components. In particular, the ACD may comprise a robot arm (10) for supporting and moving a vehicle charging connector, a motion control unit (20), and a computer vision unit (30) for pose estimation that includes or communicates with at least a neural network. The motion control unit is configured to control the movement of the robot arm to plug the connector into the vehicle socket using the pose estimation from the computer vision unit. The terms robot arm and manipulator may be used interchangeably throughout the present disclosure. The ACD may include a control system (40) and may further comprise or communicate with an image acquisition device (50), such as a camera, for acquiring an image of the vehicle socket as an input to the computer vision unit. Preferably, the computer vision unit (30) is configured to determine the pose of a reference marker in the image with respect to the camera that acquired the image. The robot arm (10) may be a controllable electromechanical mechanism comprising means for supporting the vehicle charger connector in a releasable or non-releasable manner and it is possible to activate such an autonomous charging process.

[0025] Various types of robot arms are particularly suitable for the autonomous charging device according to embodiments of the present disclosure, including, among others, SCARA robot arms, parallel robot arms, articulated robot arms, and Cartesian robot arms, although other types may also be utilized. The SCARA robot arm according to the present invention may comprise at least two parallel rotational joints and a vertical linear joint (prismatic joint) or a linear joint. The parallel robot arm according to the present invention includes a parallel kinematic structure in which a plurality of arms are connected to a common base. The articulated robot arm according to the present invention comprises a plurality of interconnected rotational joints and provides a wide range of motion similar to that of a human arm.

[0026] The ACD further includes one or more actuators that serve to move the various components of the automatic charging device (ACD). The actuators enable the robotic arm (10) to perform precise and controlled movements and facilitate the positioning and operation of the vehicle charging connector. The motion control unit (20) utilizes inputs from a computer vision unit (30) that incorporates or communicates with at least a neural network to estimate the socket pose and provide the necessary commands to the actuators. In addition to the robotic arm of the type mentioned, the robotic arm may be equipped with other components for imparting compliant motion. The additional components may include a six-legged mechanism, a Stewart platform, or other similar mechanisms that can provide additional degrees of freedom to the robotic arm, as well as flexibility and compliance in the movement of the robotic arm. These mechanisms enable the robotic arm to perform tasks that require compliant interactions, such as adapting to variations in conditions, handling uncertainties, and safely plugging the vehicle charging connector into the socket.

[0027] The selection of an appropriate robotic arm (10) may depend on various factors, including the specific characteristics and requirements of the ACD. Additionally, considerations such as the desired degrees of freedom, maximum payload, workspace constraints, and task specifications may play a role in determining the appropriate robotic arm for the ACD.

[0028] The degrees of freedom (DOF) determine the types of movement and orientation that a robot arm can achieve. An ACD according to this invention may be configured to have 1, 2, 3, or more DOFs. In an ACD configured to have 1 DOF, there is a single axis of rotation or linear motion. The arm can rotate or extend / contract along this axis. An ACD configured to have 2 DOFs has two independent axes of motion, which can be rotational or translational. An ACD configured to have 3 DOFs includes another independent axis of motion for a 2DOF arm, which allows the arm to move in three-dimensional space. In a robot configured to have 4 or more degrees of freedom, each additional degree of freedom introduces an additional axis of motion, which is either rotational or translational motion around or along the X, Y, and Z axes. This configuration allows for positional adjustment and is suitable for a limited, predictable workspace. An ACD with 6DOF provides a full range of translational and rotational motion, allowing for more precise positioning in all three translational directions (X, Y, and Z) and three additional rotational degrees of freedom around each of these axes.

[0029] In the context of the present invention, the image acquisition device (50) means a hardware component or system integrated with or mounted on the movable part of a robotic arm, designed to capture visual data or images of the surrounding environment during arm movement, operation, or in the course of the method according to the independent claim. This may include, but is not limited to, a camera, a sensor, or other optical device capable of capturing visual information. The image acquisition device is strategically positioned and oriented on the movable part to provide a desired field of view for capturing images or video footage. The captured visual data can be used for a variety of purposes, such as object recognition, pose estimation, navigation, or other computer vision applications related to the operation of the robotic arm. By mounting the image acquisition device on the movable part, it is ensured that the captured visual information corresponds to the arm's viewpoint, facilitating real-time visual recognition and analysis during arm movement or manipulation tasks.

[0030] According to the present invention, sensors may be mounted on or around the ACD and may be used to capture data from the vehicle, vehicle socket or its surroundings. Multiple sensors may be mounted on the same ACD and may be of various types to collect information on several types of parts / objects. Sensors include sensors that collect data such as force, light, wind, geographical location and orientation, temperature sensors and pressure sensors, and any combination thereof. Sensors according to the present invention may also provide the date and time the data was captured.

[0031] Autonomous charging process The ACD connection process according to the present invention includes at least the steps of: determining the orientation of a vehicle charging port (the vehicle charging port includes a socket cover, socket feature portion, socket pins, reference feature portion, marker, or a combination thereof); moving a vehicle charging connector toward the vehicle charging port; connecting a charger into the vehicle socket; enabling the vehicle to be charged; optionally, releasing the connector from the ACD; and optionally, connecting and disconnecting the charging connector. The orientation of the vehicle socket is preferably determined by a computer vision unit of the ACD (or communicating with such an ACD) based on a neural network, the neural network utilizing image data from the vehicle socket to enable the computer vision unit to estimate the orientation, and the orientation includes at least one of the position and orientation of the socket. In this disclosure, the terms socket, charging port, or simply port may be used interchangeably.

[0032] Due to the nature of the kinematic model, the ACD may be affected by several intrinsic and / or extrinsic factors that may result in inaccurate motion of the robot end effector, inaccurate orientation estimation of the socket, and / or failure of the connector to mate into the socket. The aforementioned inaccuracies may be the result of hardware and / or software components of the ACD. Other ACD components that may result in inaccuracies include, among others, inaccuracies related to the image acquisition device (50) hardware inaccuracies, image acquisition device (50) calibration, ACD kinematics, looseness in hardware components, components that provide physical compliance, motion control units, and computer vision units, as well as hardware wear and degradation.

[0033] The ACD according to the present invention is intended to be deployed in a field, such as an EV charging station, where one or more charging devices are preferably positioned to facilitate autonomous charging of such EVs. The ACD may require one or more calibration processes to achieve appropriate performance in the field, and such calibration processes may be performed in different circumstances. Such calibration may include i) internal calibration of the camera (e.g., to compensate for lens distortion), which may be performed before assembly of the ACD or after installation in the field; ii) external calibration of the camera, so-called hand-to-eye calibration, to find an accurate reference coordinate system for the rest of the kinematic chain of the system, which may be performed during assembly or after installation in the field; and iii) calibration of the motion control unit, which may be performed during or after assembly, during commissioning of the robot, in a later stage, or even during operation, with the aim of calibrating motion behavior (such as control accuracy and / or model accuracy).

[0034] To ensure that any deviations that were not previously compensated through internal, external, and / or motion control unit calibrations are ultimately compensated for during or prior to the deployment of the ACD, it is desirable to compensate for any existing or remaining deviations in the attitude of the ACD manipulator across its entire working space.

[0035] The present invention provides a method and system for minimizing the discrepancy between the target and actual orientations of the movable parts of an autonomous charging device (ACD) manipulator.

[0036] In particular, a method for minimizing the discrepancy between the target and actual orientations of the movable parts of an Autonomous Charge Device (ACD) manipulator, wherein the manipulator is adapted to connect a charging connector to an electric vehicle socket, and this method The objective is to obtain a compensation function that covers at least a portion of the manipulator's workspace, wherein the compensation function utilizes the relationship between at least one motion command and the corresponding actual posture. The manipulator is operated based on the obtained compensation function so that the resulting actual posture deviates from the target posture to the minimum. Includes, The relationship between at least one motor command and the corresponding actual posture is, The image acquisition device is attached to the movable part of the manipulator, and the device is positioned to acquire an image including a visible reference marker placed in a reference position. Command the manipulator to move using at least one motion command and acquire an image of the reference marker, The computer vision unit and the above image are used to determine the relative orientation of the reference marker to the image acquisition device, Using the reference orientation of the reference marker and the determined relative orientation of the reference marker with respect to the image acquisition device, the actual orientation of the manipulator's movable parts is calculated. Fitting a template function that associates at least one motor command with the calculated actual posture, A method obtained by

[0037] Furthermore, it is an autonomous charging device (ACD) system, A manipulator adapted to connect the charging connector to an electric vehicle socket, An image acquisition device attached to the movable part of the manipulator, Computer vision module and Motion control unit and Equipped with, The motion control unit device performs the following actions, namely: The objective is to obtain a compensation function that covers at least a portion of the manipulator's workspace, wherein the compensation function utilizes the relationship between at least one motion command and the corresponding actual posture. The manipulator is operated based on the obtained compensation function so that the resulting actual posture deviates from the target posture to the minimum. It is configured to perform the following actions: The relationship between at least one motor command and the corresponding actual posture is, The image acquisition device is attached to the movable part of the manipulator, and the device is positioned to acquire an image including a visible reference marker placed in a reference position. Command the manipulator to move using at least one motion command and acquire an image of the reference marker, The computer vision unit and the above image are used to determine the relative orientation of the reference marker to the image acquisition device, Using the reference orientation of the reference marker and the determined relative orientation of the reference marker with respect to the image acquisition device, the actual orientation of the manipulator's movable parts is calculated. Fitting a template function that associates at least one motor command with the calculated actual posture, An autonomous charging device (ACD) system obtained by this method is disclosed.

[0038] posture In the context of this invention, the term “attitude” refers to the spatial configuration of an object with respect to a given coordinate system. Attitudes can be expressed mathematically in various forms, each of which is interchangeable. One such representation is a six-element vector consisting of the translational displacement of the object along the x, y, and z axes, as well as its orientation on rotation around the roll, pitch, and yaw axes. As used herein, the term “actual attitude” refers, for example, to the real-time position and orientation of a robotic arm manipulator, including its end effector or connector holder, with respect to a given coordinate system. The actual attitude may represent the physical configuration and alignment of the robotic arm at any given moment during the charging process. The target attitude represents the intended configuration and alignment that the robotic arm should achieve to accomplish a particular objective, such as accurately plugging in the charging connector. In the context of this invention, the actual attitude represents the real-time physical state of the robotic arm, while the target attitude serves as a reference configuration that the arm aims to achieve to successfully complete the charging task.

[0039] Movable parts of a robotic arm or manipulator In the context of the present invention, the term "movable part of a robot arm or manipulator" refers to a specific component or part of a robot arm that exhibits mobility and is capable of receiving controlled motion relative to the robot arm and / or other components or parts of the ACD. This may include segments, links, or joints of the robot arm that enable its jointing and manipulative functions. The movable part can be driven by actuators, motors, or other mechanisms to achieve various degrees of freedom and enable the robot arm to perform desired motions, positions, and orientations necessary to carry out a particular autonomous charging process. Preferably, the movable part represents a part of the robot arm that is configured to support a charging connector and is a part of the ACD where higher posture accuracy is desired to achieve a successful connection process. The movable part may also be a component of the ACD, such as a robot arm end effector or a connector holder attached to an end effector. In some cases, the connector may be incorporated into the robot arm or any part thereof, and therefore the movable part may be the connector itself.

[0040] Motor commands In the context of this invention, the term "motion command" refers to at least one instruction or set of instructions for moving a movable part of a manipulator from its current position to a desired or target terminal posture or position, also referred to as a target posture. A motion command is used to instruct the manipulator to move using at least one motion command and to acquire an image of a reference marker. A motion command may include instructions such as coordinates, velocity, acceleration, waypoints, and other parameters that may be used to control the movement of the movable part of the manipulator. The target posture may be specified in relation to Cartesian coordinates, joint angles, or other reference parameters. Furthermore, a motion command may also encompass a set of target postures, in which the robot may stop at each posture and, optionally, take an image or perform a specific task. In some cases, waypoints involve continuous motion that does not require the robot to come to a complete stop at each point. In some cases, waypoints involve discontinuous motion in which the manipulator comes to a complete stop at each point, one or more times. In the context of the present invention, a motion command may encompass both individual target postures and posture sets, as well as waypoints with and without complete stops.

[0041] In some cases, motion commands may also include other ACD parameters, such as the positioning of specific actuators and the angles formed by those actuators. In some cases, such as ACDs with multiple robot arms, motion commands may include instructions on which arm to use and instructions on the desired orientation of the arm. Similarly, in a robot with multiple joints, motion commands may include instructions on the desired joint angles and the positioning of the actuators controlling those joints. Similarly, in a robot with six-legged or Stewart platform-equipped robot arms, motion commands may include instructions on the desired angles and the positioning of actuators controlling the six-legged motion.

[0042] In some cases, the motion command may also include one or more waypoints. These waypoints can serve as intermediate positions that the robot arm should pass through while moving toward a target posture. By incorporating waypoints, the robot arm can follow a planned path, enabling controlled movement during the charging process. In some cases, the waypoints may act as checkpoints where an image acquisition device, preferably mechanically mounted on the robot arm, can record visual data of the surroundings, charging port, or any relevant object. By including these additional parameters in the motion command, the motion control unit can generate a more precise and accurate trajectory for the manipulator to follow.

[0043] In some cases, when a motion command is given to or generated by the ACD, the ACD motion control unit (20) interprets the command and generates an appropriate trajectory for the robot arm to follow. During operation, the motion control unit operates the manipulator based on a compensation function such that the deviation of the resulting actual posture from the target posture is minimized, for example, by generating an appropriate trajectory for the robot arm to follow based on the motion command. The trajectory may be calculated using equations of motion that take into account the ACD physical properties and the desired motion parameters specified in the motion command. In some cases, the motion command may include a specific set of waypoints or trajectory that the movable part of the manipulator should follow to reach a desired terminal posture.

[0044] Once a trajectory is generated, the motion control unit sends a command to at least one ACD actuator to execute the motion. During the motion, the ACD's feedback sensors may also provide information about the robot's actual position and velocity, which may be used by the motion control unit to adjust the trajectory and ensure that the robot reaches the desired endpoint posture accurately and efficiently.

[0045] In the context of the present invention, the generation of motion commands can be achieved through various methods. Motion commands in the context of the present invention may be generated internally or externally by an autonomous charging device (ACD). Preferably, motion commands may be generated by an ACD control system and executed by a motion control unit.

[0046] In some cases, the ACD itself may be configured to generate motion commands based on its internal control system or through algorithms. In some cases, motion commands may be provided to the ACD manually, and the operator may provide the necessary commands, such as coordinates, velocity, acceleration, waypoints, and other parameters, using an interface or control panel to provide such commands. In some cases, external generation of motion commands may involve a central command center or remote control system providing remote commands to the ACD, which may be further processed by the ACD control system and / or executed by the motion control unit. In some cases, motion commands are generated using a scripting language or programming language. In this scenario, a set of commands is written or generated to specify the desired movement of the manipulator. Motion commands may be based on mathematical algorithms, predefined patterns, or a complex series of motions. In some cases, motion commands may be generated by an AI-based algorithm or planning module that takes into account various factors such as the environment, charging requirements, obstacle avoidance, and optimization goals.

[0047] The method according to the present invention involves obtaining a compensation function that covers at least a portion of the working space of a manipulator, the compensation function utilizing a relationship between at least one motion command and a corresponding actual posture, and operating the manipulator based on the obtained compensation function such that the resulting actual posture deviates from a target posture to the minimum.

[0048] In the context of this invention, a compensation function may be a mathematical relationship, algorithm, or model used to adjust the position or trajectory of an ACD manipulator to compensate for various sources of errors or deviations from a desired path or position. While compensation functions can take many forms, they are typically designed to account for factors such as joint and link deflection, thermal expansion, and fluctuations in the ACD environment. By modeling these factors and operating the manipulator based on the resulting compensation function, the accuracy and precision of ACD motion can be improved. More specifically, a compensation function refers to an adjustment or correction applied, for example, by a motion control unit to a manipulator's motion command to minimize the deviation between the target and actual attitudes of the movable parts of the Autonomous Charger (ACD). The compensation function incorporates a relationship between a motion command and the corresponding actual attitude. This relationship is established by capturing the manipulator's behavior and characteristics throughout at least a portion of its workspace. By utilizing the compensation function, the manipulator operates during the ACD process to minimize the deviation between the actually achieved attitude (actual attitude) and the target attitude.

[0049] The method of the present invention includes obtaining, calculating, deriving, establishing, or otherwise determining such a relationship between at least one motion command and the calculated actual posture of a movable part of an autonomous charging device (ACD) manipulator. The relationship is preferably expressed as a function, such as a template function that associates at least one motion command with the calculated actual posture.

[0050] The relationship between at least one motor command and the corresponding actual posture is, The image acquisition device is attached to the movable part of the manipulator, and the device is positioned to acquire an image including a visible reference marker placed in a reference position. Command the manipulator to move using at least one motion command and acquire an image of the reference marker, The computer vision unit and the above image are used to determine the relative orientation of the reference marker to the image acquisition device, Using the reference orientation of the reference marker and the determined relative orientation of the reference marker with respect to the image acquisition device, the actual orientation of the manipulator's movable parts is calculated. Fitting a template function that associates at least one motor command with the calculated actual posture. It is obtained by [method].

[0051] In some cases, the image acquisition device may be a 2D camera configured to capture images with a wide field of view, enabling comprehensive scene analysis and accurate reference marker detection. In some cases, the image acquisition device may also include a depth sensor, enabling 3D pose estimation of the reference marker.

[0052] The term "reference orientation" in which the reference marker is positioned may also refer to a reference orientation defined in an absolute coordinate system, thereby representing the position and orientation of the reference marker relative to an absolute coordinate system fixed to the base of the movable part, and the orientation of the reference marker is expressed using fixed XYZ coordinates and Euler angles or quaternion representation. In some embodiments, the reference orientation is defined relative to the base frame of the manipulator or ACD. In some embodiments, the reference orientation represents the position and orientation of the reference marker relative to a tool frame attached to the manipulator. In some embodiments, the reference orientation is defined and specified by the user or operator based on their desired reference point.

[0053] In some cases, the reference posture may be a substantially stationary reference posture, and the posture of the reference marker remains substantially fixed or substantially constant throughout the entire process of establishing the relationship between the motion command and the actual posture. Provided that the posture of the reference marker remains substantially stationary, such a reference posture may be a posture that is not known or unclear beforehand. Preferably, the reference marker does not move relative to the immobile part of the manipulator while the image is being recorded. In some cases, the reference posture is not substantially fixed or substantially constant throughout the entire method of establishing the relationship. Instead, the reference posture may change or fluctuate over time or in response to specific conditions. In this case, the reference posture is preferably a known posture, in particular, the variation in posture throughout the method is known and taken into account when calculating the actual posture of the movable part of the manipulator. Such prior knowledge may be obtained through calibration, mapping, or other techniques for defining an initial reference coordinate system for posture estimation.

[0054] The reference orientation may be arbitrary or selected based on the convenience of the ACD or specific requirements. This means that the position and orientation of the reference marker, as well as its relationship to the image acquisition device, may be set to any desired value without strictly adhering to a predefined coordinate system or reference coordinate system. The reference marker is preferably positioned so that it is visible and / or remains visible throughout the entire portion of the workspace to which the compensation function is applied, and so that the computer vision unit can estimate its orientation with sufficient accuracy.

[0055] This method includes commanding a manipulator to move using at least one motion command and acquiring an image of a reference marker. Commanding a manipulator to move using at least one motion command and acquiring an image of a reference marker may include any or more of the following steps, in any order: commanding the manipulator to move, commanding an image acquisition device to acquire an image, enabling the manipulator to move, and acquiring an image of the reference marker after enabling the manipulator to move. Commanding a manipulator to move using at least one motion command and acquiring an image of a reference marker may be a command generated by an ACD control system, an ACD motion control unit, or a combination thereof. In some cases, an image including the reference marker is acquired when the manipulator moves based on the motion command. In some cases, an image including the reference marker is acquired while the manipulator is moving. The image acquisition process may be performed concurrently with the manipulator's movement, allowing for real-time feedback of the reference marker's posture. Alternatively, the image of the reference marker is acquired immediately after the manipulator completes its movement. The manipulator may momentarily pause to acquire an image of a reference marker once it reaches a desired position or orientation. This technique enables stable, still images and facilitates accurate orientation estimation. In another embodiment, image acquisition including the reference marker is performed throughout the entire movement of the manipulator. The image acquisition device may capture the position and orientation of the reference marker at regular intervals while the manipulator is moving. Continuous image acquisition can provide a comprehensive dataset for orientation estimation and can enhance the accuracy of the relationship established between the motion command and the calculated actual orientation.

[0056] This method involves using a computer vision unit and the aforementioned images to determine the relative pose of a reference marker with respect to an image acquisition device. In some cases, the computer vision unit may utilize a deep learning-based method for pose estimation, thereby training a convolutional neural network (CNN) on a large dataset of reference marker images and corresponding ground truth poses. In this case, the CNN takes the acquired images as input and directly predicts the relative pose of the reference marker with respect to the image acquisition device.

[0057] This method includes calculating the actual orientation of the manipulator's movable parts using the reference orientation of a reference marker and the determined relative orientation of the reference marker with respect to the image acquisition device. The manipulator's kinematic model and calibration parameters may be used to calculate the transformation between the image acquisition device and the manipulator's end effector. By combining this transformation with the relative orientation of the reference marker, the actual orientation of the manipulator's end effector can be calculated.

[0058] The method further includes fitting a template function that associates at least one motion command with a calculated actual pose. The template function may be selected from at least one of the parameterized functions, including a linear function, a polynomial function, a matrix function, a piecewise function, a lookup table, or a heuristic algorithm. In some cases, the template function is a parameterized function, thereby allowing the function to be expressed as an equation or algorithm that takes motion commands as input and produces the predicted actual pose as output. The parameters of the function may be tuned and optimized to accurately capture the relationship between the input (motion command) and the output (actual pose). A regression function or optimization function, such as linear least squares, may be used to fit the function parameters to the data.

[0059] In some cases, this method involves establishing a relationship between motion commands and calculated actual poses through analytical modeling to mathematically model the behavior of the ACD in various scenarios. The analytical model can be used to describe the behavior of the end effector within a specific pose in the workspace.

[0060] In some cases, the method further includes obtaining a model that represents the relationship between at least one motor command and the calculated actual pose, the model being selected from a group that includes feedforward neural networks (FNNs), convolutional neural networks (CNNs), recurrent neural networks (RNNs), and graph neural networks (GNNs).

[0061] In some cases, the method involves establishing a relationship between motion commands and calculated actual postures through experimental testing to measure the robot's behavior in various scenarios, such as building relationships under various environmental conditions, and using the resulting data in a compensation function to explain these variations. For example, an experimental setup may be constructed in which the ACD is tasked with performing a plug-in process under low-temperature conditions that are expected to affect posture estimation while measuring the manipulator's actual posture in relation to a target posture. By varying the temperature in which the ACD operates and measuring the resulting deviation from the desired position, a dataset relating motion commands to actual postures can be obtained and subsequently used to construct a compensation function. A regression model may be constructed to fit the dataset and identify the relationship between temperature conditions and the resulting deviation from the target position. Preferably, linear regression may be used to model the relationship between temperature conditions and posture deviations, or a more complex model such as polynomial regression may be used to capture more complex relationships.

[0062] Preferably, the method includes establishing relationships between a plurality of motion commands and their calculated actual postures. Preferably, the plurality of motion commands encompass a substantial portion of the manipulator's workspace, covering a range of positions and orientations sufficient to enable effective operation of the ACD.

[0063] In some cases, the method may further include using existing functions in the workspace, which include representations of key features and existing data relationships between motion commands and actual poses for multiple motion commands. These existing functions can serve as initial references for establishing relationships between at least one motion command and a calculated actual pose in other poses or positions in the workspace.

[0064] The method according to the present invention involves operating a manipulator based on a obtained compensation function such that the deviation of the resulting actual posture from the target posture is minimized, and the compensation function utilizes the obtained relationship between at least one motor command and the corresponding actual posture.

[0065] The compensation function serves as a mathematical model or representation of the relationship between motion commands and posture, enabling the manipulator to accurately estimate and adjust its posture. The compensation function can determine the motion commands given for a target posture by utilizing a lookup table representing the above relationship. When the ACD is operating, the lookup table can be used to adjust the robot's trajectory in real time based on the observed deviation between the desired posture and the actual posture of the ACD manipulator.

[0066] A compensation function can determine appropriate adjustments to a motion command to compensate for this discrepancy based on a representation of the relationship between the motion command and the posture, using a lookup table. For example, if a particular motion command with a target posture results in a calculated actual posture, this relationship is recorded in the lookup table. During operation, the lookup table can be used to adjust the trajectory of the robot motion command in real time based on the observed discrepancy between the desired posture and the actual posture of the ACD manipulator. Thus, a compensation function can utilize the lookup table to determine appropriate adjustments to the motion command to compensate for this discrepancy. By applying the compensation function, the motion command can be dynamically adjusted, ensuring that the resulting actual posture aligns with the desired or target posture, thereby improving the accuracy and performance of the ACD manipulator. In general, compensation functions enable bridging the gap between theoretical systems and their actual implementations, allowing for accurate estimation and control of the manipulator's posture. Preferably, the method includes updating the compensation function based on data acquired during the ACD's operation.

[0067] The compensation function can be modeled to accurately represent the behavior of the ACD under various conditions and to allow for real-time adjustment of the robot's trajectory or position to compensate for any source of error or deviation from motion commands and / or target posture.

[0068] An optimization algorithm may be incorporated to determine and refine the compensation function. This algorithm can facilitate the discovery of appropriate parameter values ​​that align the compensation function with a specified relationship. The optimization algorithm can iteratively adjust parameters based on optimization criteria that typically minimize the deviation between the actual pose and the target pose. The optimization process involves iteratively refining parameter values ​​to improve the fit of the function to data points, including motion commands and corresponding actual poses. The algorithm can adjust parameters based on the gradient of the error function, guiding it in a direction that minimizes the error. The optimization algorithm can continue iterating until a convergence criterion is met, ensuring a desired level of accuracy or stability in the parameter estimation. By utilizing the optimization algorithm, the compensation function can fit and improve its accuracy over time. The optimization algorithm can fine-tune parameters to align the predicted pose with the desired target pose, ensuring that the manipulator is operated in such a way that the minimum deviation between the actual pose and the target pose is obtained.

[0069] The compensation function can be learned from the behavior of the ACD in various scenarios and further constructed through machine learning algorithms to build the compensation function based on this data. In some embodiments, the machine learning algorithm is trained on data from the ACD sensor to predict the influence of various environmental conditions on the robot's behavior and to use this information to adjust the robot's movement in real time. In such cases, a neural network may be used to learn the relationship between sensor data, such as a camera mounted on the ACD manipulator, and the desired position of the manipulator. The neural network can be trained on a large dataset of sensor readings and corresponding end-effector poses using techniques such as backpropagation to adjust the network weights and biases to minimize the error between the predicted end-effector position and the actual end-effector position. Once trained, the neural network can be used in real time to predict appropriate compensation to apply to the ACD trajectory based on the current sensor data.

[0070] In the context of the present invention, an interpolation procedure may be used to estimate the values ​​between data points in a lookup table, where the data points correlate the motion command with the corresponding calculated actual posture.

[0071] The calculated actual posture preferably includes translational and / or rotational values ​​for each motion command. One common interpolation method is linear interpolation, which assumes a linear relationship between data points. To apply linear interpolation to a lookup table for translation and / or rotation, the following steps can be followed. Sort the lookup table in ascending order of translation and rotation values. Given input translation and rotation values, find the two closest data points in the lookup table whose translation and rotation values ​​enclose the input value. Weights are calculated for two data points based on their distances from the input values. A weight is assigned to a closer data point than to a weight assigned to a farther data point. The translational and rotational values ​​are interpolated by combining two data points using their respective weights. For example, a weighted average of the translational and / or rotational values ​​may be used.

[0072] In the context of this invention, the term "workspace of a robotic manipulator" refers to the region within the space that a robotic arm can reach and operate in. This can be defined as the volume of space that can be accessed by the movable part of the robotic manipulator while the robot remains fixed to its base.

[0073] In some cases, the method includes dynamically evaluating the deviation between the resulting actual posture and the target posture during the operation of the manipulator, and adjusting at least one boundary of the workspace if the deviation exceeds a predetermined threshold. Preferably, the adjustment of at least one boundary of the workspace includes adjusting the range of motion of the manipulator to a narrower range so that the deviation of the resulting actual posture from the target posture is minimized. If the deviation between the resulting actual posture and the target posture during the operation of the manipulator exceeds a predetermined threshold, the method further includes dynamically reducing the boundary of the workspace. This reduction may be carried out to ensure that the deviation falls within an acceptable threshold, thereby maintaining the desired performance and motion limits of the robot arm. The reduction of the workspace boundary is performed by modifying the limits or constraints on the range of motion of the arm. This may include limiting the maximum extension, rotation, or other parameters that define the acceptable movement of the arm. By doing so, the system effectively narrows the workspace in which the arm operates, ensuring that the evaluated deviation remains within a predetermined threshold.

[0074] In some cases, the method according to the present invention includes defining boundaries within the manipulator's workspace from which a relationship should be obtained and / or a compensation function will be obtained. The compensation function may cover the entire workspace of the manipulator, and in some cases, the function may cover at least a portion or part of it, depending preferably on the relationship obtained. When the function covers a portion or part of the workspace, the method may include the step of defining boundaries within the workspace. Preferably, the portion of the workspace within the boundary overlaps with the spatial volume from which the ACD system should perform its task. For example, if a connector attached to the manipulator is considered a moving part, the boundary within the workspace of the connector from which the compensation function is obtained ideally overlaps with the location where the vehicle's entry point may be found during operation.

[0075] Defining the workspace boundaries includes defining the range of motion of the manipulator, defining the maximum reach of the manipulator, and defining the directional adjustment capability of the manipulator. In some cases, defining the workspace boundaries includes defining the maximum and minimum joint angles of each joint in the manipulator. In some cases, defining the workspace boundaries includes creating a grid of sample points in the workspace, the grid having a sample density that allows all areas of the workspace to be covered but is sparse enough to keep data collection manageable. In some cases, defining the workspace boundaries includes generating a map of the workspace, thereby depicting a representation of the range of positions and orientations that the manipulator can reach within its physical constraints. In some cases, the method further includes optimizing the workspace to map and identify areas of potential improvement by adjusting the design or configuration of the manipulator to improve its range of motion and optimize its performance within the workspace. The size and shape of the ACD workspace may depend on various factors, such as the physical dimensions of the ACD, the range of motion of its joints, and any constraints or obstacles in the environment. The workspace can be defined by its boundaries, which are typically determined by the maximum and minimum joint angles or positions of the robotic arm.

[0076] In some cases, the workspace is defined in relation to motion limitations, and the workspace boundary is determined by the robot arm's maximum reach, extension, or movement capabilities, taking into account its mechanical design and constraints. Alternatively, the workspace boundary may be defined based on physical constraints, taking into account the dimensions and limitations of the environment in which the robot arm operates. These physical constraints may include the dimensions of the ACD or the presence of surrounding obstacles.

[0077] In certain embodiments, the workspace boundary may be defined based on task-specific requirements. This includes considering factors such as vehicle type, ACD location, or motion constraints specific to the charging process to be performed by the robotic arm. In some embodiments, the workspace boundary may be established using real-time sensor feedback. Proximity sensors, cameras, or other sensing mechanisms may provide feedback regarding the presence of objects or obstacles, thereby influencing the acceptable range of arm movement within the workspace. Software constraints may also contribute to the demarcation of the workspace. In some cases, algorithms and software modules may optimize the movement of the manipulator's movable parts by considering factors such as joint limitations, collision avoidance, or motion smoothness, thereby determining the workspace boundary. Furthermore, the workspace may be defined using a specific coordinate system or reference coordinate system relevant to the task at hand. The coordinate system may include a Cartesian coordinate system, joint angles, or position vectors, enabling precise definition and control of the workspace boundary.

[0078] In some cases, when the relationship between at least one motion command and the calculated actual posture remains within a predetermined acceptable threshold, the method may include dynamically extending the boundaries of the workspace. This extension is performed in a manner that benefits from the available range of motion, thereby expanding the operational capabilities of the robotic arm.

[0079] Adjusting the workspace boundary based on the relationship between at least one motion command and the calculated actual posture can ensure that the robot arm's motion remains within predefined limits, thereby optimizing its performance, safety, and efficiency. By analyzing the relationship between the motion command and the calculated actual posture, the ACD can identify any deviations or discrepancies that may occur during the movement of the ACD manipulator's movable parts.

[0080] In the context of the present invention, a reference marker may be a physical object or feature used as a reference point for aligning the robot's coordinate system with another coordinate system or a reference coordinate system. Reference markers can take many forms, but they are typically designed to be easily detected and recognized by sensors or cameras mounted on the robot or elsewhere in the workspace. Examples of reference markers include reflective spheres, grid patterns, and QR codes. In some cases, the method includes using a vehicle socket as a reference marker.

[0081] The robot's image acquisition device is positioned to detect reference markers and use them as reference points for calculating the position and orientation of the robot's end effectors and other components, preferably the movable parts of the manipulator. This information is then used to generate a transformation matrix that relates the robot's coordinate system to another coordinate system or a reference coordinate system.

[0082] In some embodiments, a reference marker is designed and positioned to enable attitude estimation with a certain level of variability or uncertainty. The attitude estimation provided by the reference marker may range from ground truth attitude to attitudes within a defined confidence interval. The confidence interval represents the range of attitudes within which the estimated attitude of the reference marker is expected to fall with a certain level of confidence. The confidence interval takes into account various factors inherent in the attitude estimation process, such as measurement errors, noise, and uncertainties.

[0083] In some embodiments, the method involves obtaining a comparison matrix generated by comparing obtained functions representing the relationship between at least one motion command and the corresponding actual posture of various ACDs. In this example, multiple ACDs having similar functions and manipulators are considered, each ACD having its own function mapping the relationship between a motion command and the corresponding actual posture. The comparison matrix may provide a template that captures common patterns and trends in the identified relationships across multiple ACDs, thereby reducing the number of data points required to establish the relationship between the motion command and the actual posture. Thus, instead of individually collecting a large amount of data for each ACD, the comparison matrix allows a smaller subset of representative data points to be fed into the method for obtaining a function representing the relationship between at least one motion command and the corresponding actual posture of the movable part of the ACD manipulator.

[0084] In some embodiments, the method includes obtaining a compensation function, thereby allowing the compensation function to be appropriately transferred to multiple ACDs. In some cases, obtaining the compensation function includes a regularization step selected from weight decay and dropout, thereby preventing overfitting of the mapping function to sample data and enabling the function to perform better on new inputs.

[0085] In some cases, obtaining a compensation function involves using a dataset of motor commands that preferably represent a range of inputs and outputs expected to be encountered during operation, in order to allow the function to generalize better and improve its ability to predict new, unknown inputs.

[0086] In some cases, obtaining a compensation function involves a transfer learning step, where the function becomes a model that can be trained on one set of motor commands, which is then used as a starting point for motor commands, thereby allowing the model to improve its performance on target commands by leveraging knowledge gained from source commands.

[0087] In some cases, obtaining a compensation function involves models that use domain adaptation techniques, thereby allowing the model to operate in environments different from the one in which it was trained.

[0088] In some cases, obtaining a compensation function involves combining multiple models to construct a new one, thereby increasing the robustness and generalizability of the compensation function by reducing the risk of relying on a single model whose transferability may be limited.

[0089] The method according to the present invention enables the time-efficient acquisition of a relationship between at least one motion command and a corresponding actual posture. In some cases, the method includes a data preprocessing step that enables a reduction in the amount of data that the function needs to process. In some cases, the method includes parallel processing using multiple processors to reduce overall computation time by performing calculations simultaneously so that the workload is distributed across multiple processors. In some cases, the method includes a linearization step that enables a reduction in the number of calculations required to generate the calculation. In some cases, the method includes a model compression step that enables a reduction in the size of data points without significantly affecting its performance, and the model compression step includes one or more techniques selected from pruning, quantization, and distillation, thereby improving the method by reducing the number of calculations required to generate the relationship with a smaller model size.

[0090] In one embodiment of the present invention, the Gauss-Newton method is used to iteratively estimate the relationship between a motion command and a calculated actual posture. The Gauss-Newton method can be used to iteratively determine the relationship between a motion command and a calculated actual posture with a high level of confidence while minimizing the number of iterations required for each posture or workspace. In some cases, the method may begin with an initial prediction of the relationship parameters, which is iteratively refined based on the Jacobian matrix and weighted least-squares optimization. The initial prediction can be based on prior knowledge, historical data, or approximations derived from ACD specifications.

[0091] By minimizing the error between the motor command and the calculated actual posture, this method efficiently converges towards an accurate estimation of the relationship within a reduced number of iterations.

[0092] In some cases, one or more components may be referred to herein as “configured to do so,” “composed of by,” “configurable to do so,” “operable to do so,” “adaptable to do so,” etc. It will be recognized by those skilled in the art that such terms (e.g., “configured to do so”) generally encompass active state components and / or inactive state components and / or standby state components unless otherwise required by the context.

[0093] In particular, conditional language used herein, such as “can,” “could,” “might,” “may,” and “e.g.,” is generally intended to convey that a particular embodiment includes certain features, elements, and / or steps, while other embodiments do not, unless otherwise specifically stated or understood in the context in which they are used. Accordingly, such conditional language is generally not intended to imply that features, elements, and / or steps are required in some way in one or more embodiments, or that one or more embodiments necessarily include logic for determining, with or without author input or instruction, whether the features, elements, and / or steps are included in or implemented in any particular embodiment.

[0094] Terms such as “equip,” “include,” and “possess” are synonyms and are used inclusively, in an open-ended manner, without excluding any additional elements, features, actions, or behaviors. The term “or” is used in its inclusive sense (rather than its exclusive sense), for example, when used to connect a list of elements, so that the term “or” means one, some, or all of the elements in the list. In addition, the articles “a,” “an,” and “the,” as used in this application and the attached claims, shall be interpreted as meaning “one or more” or “at least one,” unless otherwise specified. Where used herein, the phrases “at least one of” or “and / or” of the enumerated items refer to any combination of those items, including a single member.

[0095] Similarly, while method steps are depicted in a specific order in the drawings, it should be recognized that such method steps do not need to be performed in the specific or sequential order shown, or that not all illustrated actions need to be performed to achieve the desired result. Furthermore, the drawings may schematically depict another exemplary process in the form of a flowchart. However, other actions not depicted can be incorporated into the schematicly shown exemplary methods and processes. For example, one or more additional actions can be performed before, after, simultaneously with, or in between any of the illustrated actions. Additionally, actions may be rearranged or reordered in other embodiments. In certain environmental conditions, multitasking and parallel processing may be advantageous.

[0096] The detailed descriptions provided above are essentially illustrative, and it should be understood that any variations that do not deviate from the spirit and / or idea of ​​the claimed subject matter are intended to fall within the scope of the claims. Such variations should not be considered deviations from the spirit and scope of the claimed subject matter.

Claims

1. A method for minimizing the deviation between the target and actual orientation of the movable part of an autonomous charging device (ACD) manipulator, wherein the manipulator is adapted to connect a charging connector to an electric vehicle socket, and the method is: a. Obtaining a compensation function that covers at least a portion of the working space of the manipulator, wherein the compensation function utilizes the relationship between at least one motion command and the corresponding actual posture. b. Operating the manipulator based on the obtained compensation function so that the resulting deviation of the actual posture from the target posture is minimized. Includes, The relationship between at least one motor command and the corresponding actual posture is established by at least the following steps, namely: - Attaching an image acquisition device to the movable part of the manipulator, wherein the device is positioned to acquire an image including a visible reference marker placed in a reference position, - Commanding the manipulator to move using at least one motion command, and acquiring an image of the reference marker, - To determine the relative orientation of the reference marker with respect to the image acquisition device, the computer vision unit and the image are used. - Using the reference orientation of the reference marker and the determined relative orientation of the reference marker with respect to the image acquisition device, the actual orientation of the movable part of the manipulator is calculated. - Fitting a template function that associates the at least one motor command with the calculated actual posture. A method obtained by a method including the following.

2. The method according to claim 1, wherein the template function is selected from at least one of parameterized functions, including linear functions, polynomial functions, matrix functions, piecewise functions, lookup tables, or heuristic algorithms.

3. The method according to any one of claims 1 or 2, comprising obtaining a model representing the relationship between the at least one motor command and the calculated actual posture, wherein the model is selected from the group including feedforward neural networks (FNNs), convolutional neural networks (CNNs), recurrent neural networks (RNNs), and graph neural networks (GNNs).

4. The method according to any one of the preceding claims, wherein the vehicle socket is used as the reference marker.

5. The method according to any one of the preceding claims, comprising establishing relationships between a number of motor commands and their corresponding calculated actual postures.

6. The method according to any one of the preceding claims, wherein the plurality of motion commands encompass a substantial portion of the working space of the manipulator, covering a range of positions and orientations sufficient to enable effective operation of the ACD.

7. The method according to any one of the preceding claims, comprising using an existing function of the workspace which includes representations of key features and existing data relationships between motion commands and actual postures for a plurality of motion commands.

8. The method according to any one of the preceding claims, wherein the compensation function includes a lookup table representing the relationship between at least one motor command and the corresponding actual posture for determining a motor command to be given to a target posture.

9. The method according to any one of the preceding claims, comprising updating the compensation function based on data acquired during the operation of the ACD.

10. The method according to claim 8, wherein the compensation function includes an optimization algorithm for improving and optimizing the parameter values ​​of the lookup table to minimize the discrepancy between the actual posture and the target posture.

11. The method according to claim 10, wherein the optimization algorithm includes an interpolation step for estimating values ​​between data points in the lookup table, the data points include at least one of the translational and rotational values ​​of each motion command.

12. The method according to any one of the preceding claims, further comprising: dynamically evaluating the resulting deviation between the actual posture and the target posture during the operation of the manipulator; and adjusting at least one boundary of the workspace if the deviation exceeds a predetermined threshold.

13. The method according to claim 12, wherein the adjustment of the at least one boundary of the workspace includes adjusting the range of motion of the manipulator to a narrower range such that the resulting deviation of the actual posture from the target posture is minimized.

14. The method according to any one of the preceding claims, wherein the motion command includes at least one of a predefined set of waypoints and coordinates, and the predefined waypoints for multiple motion commands are substantially the same.

15. The method according to any one of the preceding claims, wherein the motion command includes a series of waypoints that serve as intermediate positions for the movable part of the manipulator and for the image acquisition device to record subsequent images of the reference marker.

16. The method according to any one of the preceding claims, wherein the reference posture is determined by acquiring an image of the reference marker when the manipulator is in a state where the difference between the actual posture and the target posture is minimized, determining the posture of the reference marker relative to the camera, and calculating the absolute posture of the reference marker.

17. The method according to any one of the preceding claims, wherein the reference position is unknown when the reference position is substantially stationary.

18. The method according to any one of the preceding claims, wherein the reference position is a known position when the reference position is not substantially stationary.

19. The method according to any one of the prior claims, wherein the aforementioned reference position is known with a confidence interval.

20. The method according to any one of the preceding claims, comprising using a plurality of reference markers, each positioned in a different orientation.

21. The method according to any one of the preceding claims, wherein the manipulator is operated based on the compensation function obtained by combining the compensation function and the predefined waypoints to compensate for the deviation in each of the plurality of working positions.

22. The method according to any one of the preceding claims, comprising using a socket having a reference marker to verify the compensation function, wherein the manipulator is commanded to operate the socket to a target posture, and the actual posture of the socket is compared to a posture predicted based on the compensation function.

23. - A manipulator adapted to connect the charging connector to an electric vehicle socket, - An image acquisition device attached to the movable part of the manipulator, - Computer vision module, - Motion control unit and Equipped with, The aforementioned motion control unit device performs the following actions, namely: - To obtain a compensation function that covers at least a portion of the working space of the manipulator, wherein the compensation function utilizes the relationship between at least one motion command and the corresponding actual posture. - The manipulator is operated based on the obtained compensation function so that the resulting deviation of the actual posture from the target posture is minimized. It is configured to perform the following actions: The relationship between at least one motor command and the corresponding actual posture is established by at least the following steps, namely: - Attaching an image acquisition device to the movable part of the manipulator, wherein the device is positioned to acquire an image including a visible reference marker placed in a reference position, - Commanding the manipulator to move using at least one motion command, and acquiring an image of the reference marker, - To determine the relative orientation of the reference marker with respect to the image acquisition device, the computer vision unit and the image are used. - Using the reference orientation of the reference marker and the determined relative orientation of the reference marker with respect to the image acquisition device, the actual orientation of the movable part of the manipulator is calculated. - Fitting a template function that associates the at least one motor command with the calculated actual posture. An autonomous charging device (ACD) system obtained by a method including the following.

24. A method for obtaining a function that represents the relationship between at least one motion command and the corresponding actual posture of a movable part of an ACD manipulator, a) Attaching an image acquisition device to the movable part of the manipulator, wherein the device is positioned to acquire an image including a visible reference marker positioned in a reference posture, b) Commanding the manipulator to move using at least one motion command and acquiring an image of the reference marker, c) Using a computer vision unit and the image to determine the relative orientation of the reference marker with respect to the image acquisition device, d) Using the reference orientation of the reference marker and the determined relative orientation of the reference marker with respect to the image acquisition device, calculate the actual orientation of the movable part of the manipulator, e) fitting a template function that associates the at least one motion command with the calculated actual posture. Methods that include...

Images