Charging pile automatic charging system based on adaptive infrared sensing

Abstract

The present invention relates to the field of automatic charging technology, and more specifically, to an automatic charging system for charging piles based on adaptive infrared sensing. It includes an infrared sensing detection unit that actively transmits infrared signals and receives reflected signals, and obtains the target's spatial orientation, distance, and charging port position based on the infrared signal reflected echo through time difference and signal strength; a signal judgment and control unit that determines whether the target has entered the charging area, whether it has accurately parked in place, and whether the charging port is aligned based on spatial position analysis and geometric feature matching of infrared sensing data, and generates control instructions based on the judgment results using manifold elastic drive and geometric estimation control technology. The present invention effectively solves the problem of traditional infrared sensing being susceptible to interference in complex environments such as rain and fog, high humidity, and metal reflection by introducing polarization phase entropy analysis, multipath suppression algorithm, and nonlinear environmental compensation mechanism.

Description

Technical Field

[0001] The present invention relates to the field of automatic charging technology, and in particular to an automatic charging system for a charging pile based on adaptive infrared sensing. Background Art

[0002] In complex or harsh environments, such as rain, fog, and high humidity, traditional recognition and positioning methods are susceptible to interference, leading to recognition failure or inaccuracy. This is especially true on highly reflective surfaces such as metal, where the multiple reflection paths generated by the infrared signal can cause delay estimation errors. Furthermore, the vehicle's paint and the metal charging port frame can produce similar highlights in the infrared image, making it difficult for traditional recognition methods to distinguish the true charging port location, especially in conditions of strong reflection, low light, or when covered in dirt. Furthermore, slight vehicle movement during charging, caused by factors such as airflow disturbances and subtle posture adjustments, requires the automatic docking system to adjust and compensate for these minor deviations in real time to ensure a stable and reliable docking process. To achieve efficient automatic docking, it is also necessary to avoid damage to the vehicle surface caused by contact impact and plan a safe path to avoid obstacles in complex mechanical environments. Therefore, an automatic charging system for charging piles based on adaptive infrared sensing is designed. Summary of the Invention

[0003] The purpose of the present invention is to provide an automatic charging system for charging piles based on adaptive infrared sensing, so as to solve the problems of low recognition accuracy, poor anti-interference ability and insufficient dynamic docking stability of the existing automatic charging system in complex environments proposed in the above background technology.

[0004] To achieve the above objectives, the present invention provides an automatic charging system for charging piles based on adaptive infrared sensing, comprising:

[0005] Infrared sensor detection unit, which actively transmits infrared signals and receives reflected signals, and obtains the target spatial orientation, distance, and charging port location based on the infrared signal reflected echo through time difference and signal strength;

[0006] A signal judgment and control unit, which determines whether the target has entered the charging area, is accurately parked, and is aligned with the charging port based on spatial position analysis and geometric feature matching of infrared sensor data, and generates control instructions based on the judgment results using manifold elastic drive and geometric estimation control technology;

[0007] A charging gun head control unit, which receives control instructions and drives the actuator to control the charging gun head;

[0008] A dynamic power supply control unit is configured to perform dynamic power supply according to a control instruction.

[0009] As a further improvement of the present technical solution, the infrared sensing detection unit includes an infrared signal receiving module and an infrared signal processing module;

[0010] The infrared signal receiving module actively transmits modulated infrared pulses at a preset frequency n, and synchronously receives the infrared signal reflected by the target through the infrared receiver array;

[0011] The infrared signal processing module includes a distance calculation module, a target orientation solution module and a charging port identification module;

[0012] The distance calculation module calculates the distance of the target object based on the signal propagation time difference and the speed of light;

[0013] The target direction calculation module calculates the target's direction based on the arrival time difference and signal strength difference, and introduces polarization phase entropy in the calculation process to distinguish and suppress signal distortion caused by multipath interference and ambient humidity;

[0014] The charging port identification module identifies the specific location of the charging port using target surface features.

[0015] As a further improvement of the present technical solution, the target position calculation module calculates the target position based on the arrival time difference and the signal strength difference, and introduces polarization phase entropy in the calculation process, including the following steps:

[0016] S1.1. Receive the original infrared signal, extract the polarization phase characteristics of each channel of the infrared signal, calculate the polarization phase entropy based on the polarization phase angle distribution, and determine whether the signal comes from a direct path;

[0017] S1.2. Calculate the time delay difference between a pair of receivers, perform temperature compensation on the inter-receiver baseline based on the ambient temperature, and calculate the initial azimuth based on the calibrated baseline and time delay difference;

[0018] S1.3. Calculate the compensation factor using a fusion model of humidity and polarization phase entropy, and perform nonlinear environmental compensation on the initial azimuth angle based on the compensation factor;

[0019] S1.4. Combine the normalized signal strength difference with the polarization phase entropy, dynamically adjust the fusion weight factor, and output the final fusion azimuth.

[0020] S1.5. Calculate the vertical TDOA difference, estimate the target height difference, calculate the target pitch angle using the elevation angle formula, and perform elevation angle correction based on polarization phase entropy to eliminate the elevation angle deviation caused by ground reflection.

[0021] As a further improvement of the present technical solution, the charging port identification module uses the target surface features to identify the specific charging port location, including the following steps:

[0022] S1.6. Collect target distance and orientation information and acquire thermal imaging raw data;

[0023] S1.7. Predict the search area for the charging port based on the target location, and select areas with a reflectivity exceeding d and made of metal as candidate areas.

[0024] S1.8. Collect thermal attenuation data of the target area within x seconds after being heated, construct the thermal attenuation probability distribution, and calculate the TIIE value;

[0025] S1.9. Create a metal contact mask with a TIIE value less than y, further screen it based on geometric features, and ultimately select the charging port border area that best matches the charging port characteristics.

[0026] S1.10. Perform three-dimensional fitting on the mask area where the TIIE value is less than the threshold y, extract the central coordinates of the area as the center position of the charging port, and estimate the average normal direction of the charging port border area based on the point cloud surface normal vector of the charging port border area.

[0027] As a further improvement of the present technical solution, in S1.8, constructing the thermal attenuation probability distribution and calculating the TIIE value includes the following steps:

[0028] S1.81. Use infrared thermal imaging equipment to record each frame of the image within the target area;

[0029] S1.82. For each pixel, calculate the rate of change of temperature between consecutive time frames;

[0030] S1.83. Combine the temperature change rate values ​​of all pixels into a one-dimensional array. Within a pre-set temperature change rate range, divide this range into e intervals of equal width. Count the number of occurrences of the temperature change rate value in each interval to form a histogram.

[0031] S1.84. Normalize the histogram to obtain the thermal attenuation probability distribution;

[0032] S1.85. Use the information entropy formula to calculate the TIIE value to quantify the uncertainty in the probability distribution of the temperature change rate.

[0033] As a further improvement of the present technical solution, the signal judgment and control unit includes a signal judgment module and a control instruction generation module;

[0034] The signal judgment module receives the infrared sensor data output by the infrared signal processing module, analyzes and determines whether the target has entered the charging area, whether it is parked accurately, and whether the charging port is within the identifiable range;

[0035] The signal judgment module includes a target entry and exit judgment module, a target parking judgment module, and a charging port judgment module;

[0036] The target entry and exit judgment module receives the target distance and direction data output by the infrared sensor detection unit and judges whether the target enters and leaves the charging area through a dynamic threshold algorithm;

[0037] The target parking judgment module analyzes the target position coordinates and motion status in real time, and determines whether the target has been parked through a dual verification mechanism of position and speed;

[0038] The charging port judgment module determines whether the charging port coordinates are within the charging gun working space through the accessibility mechanism;

[0039] The control instruction generation module generates a charging gun head control instruction that instructs the charging gun head control unit to act according to the judgment result of the signal judgment module by using manifold elastic drive and geometric estimation control technology, and sends a power supply start and stop control instruction to the dynamic power supply control unit after the charging gun head action is completed.

[0040] As a further improvement of the present technical solution, the method of generating a charging gun head control instruction for instructing the charging gun head control unit to operate based on the judgment result of the signal judgment module by using manifold elastic drive and geometric estimation control technology includes the following steps:

[0041] S2.1. Receive the judgment result of the signal judgment module and the center position and average normal direction of the charging port;

[0042] S2.2. Based on the current spatial position and posture information of the charging gun head, use the inverse kinematics method to calculate the candidate path from the current posture to the target charging port position and generate the initial target posture sequence;

[0043] S2.3. Call the impedance control module and the manifold constraint module to perform multi-factor correction, input the output results of the impedance control module and the manifold constraint module into the coupling matrix, perform fusion operation, and generate the final path command and joint torque command;

[0044] The impedance control module is used to detect the speed and acceleration of each joint of the charging gun robotic arm in real time. It uses FFT to extract low-frequency resonant components and constructs an adaptive impedance model to output torque correction terms to suppress vibration at the end of the charging gun tip.

[0045] The manifold constraint module is used to construct a signed distance field based on the environment point cloud, define the safe manifold metric tensor, solve the geodesic path on the Riemannian manifold, and output the corrected path;

[0046] S2.4. Analyze the micro-motion trajectory of the charging port center over a historical period. Establish a breathing disturbance model using bandpass filtering and ARIMA prediction methods. Apply advance compensation to the trajectory vector of the charging gun tip to generate the compensated final target trajectory.

[0047] S2.5. Quantize and encode the final trajectory vector and control torque to generate a control instruction packet that is resistant to electromagnetic interference;

[0048] S2.6. Send the control instruction packet to the charging gun head control unit.

[0049] As a further improvement to the present technical solution, in S2.4, the micro-motion trajectory of the charging port center in the historical time period is analyzed, and a breathing disturbance model is established by bandpass filtering and ARIMA prediction method, including the following steps:

[0050] S2.41. Collect the coordinates of the center of the charging port for each frame in the past 10 to 15 seconds to form a time series.

[0051] S2.42, applying a bandpass filter to retain only the signal components in the range z to u, thereby generating a filtered respiratory micromotion signal;

[0052] S2.43, performing differential operations on the respiratory micro-motion signal in the X, Y, and Z directions respectively;

[0053] S2.44, based on the differential respiratory micro-motion signal sequence, use the ARIMA model to predict the future Make advance predictions within seconds and generate future disturbance increments;

[0054] S2.45. Apply the predicted disturbance increment to all control point coordinates in the original inverse kinematics path planning result to obtain a control path after disturbance compensation.

[0055] As a further improvement of the present technical solution, the charging gun head control unit receives a control instruction and drives the actuator to control the charging gun head, including the following steps:

[0056] S3.1. Receive a control instruction packet from the signal judgment and control unit and perform an integrity check on the instruction packet using CRC-16.

[0057] S3.2. Decode the control instruction packet and restore it to the real physical quantity;

[0058] S3.3. Analyze the restored control instructions, extract the translation instructions and rotation instructions in the three-dimensional coordinate system, perform unit and precision conversion, and generate low-level motion control parameters;

[0059] S3.4. Send the low-level motion control parameters to the actuator control modules at all levels for mechanical docking. After the mechanical docking is completed, the key parameters are fed back to the signal judgment and control unit.

[0060] As a further improvement of the present technical solution, the dynamic power supply control unit includes a dynamic power supply module and a dynamic power outage module;

[0061] The dynamic power supply module receives the signal judgment and charging instructions from the control unit, controls the main relay to close, connects the power path in real time, and monitors various parameters during the charging process in real time;

[0062] The dynamic power-off module is used to quickly disconnect the power supply path when an abnormal state occurs during the charging process.

[0063] Compared with the prior art, the present invention has the following beneficial effects:

[0064] This adaptive infrared sensing-based automatic charging system for charging piles effectively addresses the interference vulnerability of traditional infrared sensors in complex environments such as rain, fog, high humidity, and metal reflections by introducing polarization phase entropy (PPE) analysis, a multipath mitigation algorithm, and a nonlinear environmental compensation mechanism. Dynamic baseline temperature compensation, combined with a fusion signal strength and PPE azimuth calculation method, significantly improves the accuracy of identifying the spatial position of targets and charging ports. Furthermore, the use of TIIE thermal inertia information entropy for material recognition overcomes visual interference such as reflections, low light, and dirt coverage, enabling stable and reliable charging port identification and positioning. The system maintains high-precision spatial perception capabilities in a variety of harsh environmental conditions, significantly enhancing the adaptability and stability of the automatic charging system.

[0065] 2. This adaptive infrared sensing-based automatic charging system for charging piles utilizes manifold elastic drive and geometric estimation control technology, combined with inverse kinematic path planning, impedance control, manifold constraint optimization, and disturbance prediction and compensation mechanisms to achieve high-precision trajectory control and force regulation of the charging gun tip. By constructing a breathing disturbance model (bandpass filtering + ARIMA prediction), position deviations caused by target micro-movements are pre-compensated, improving docking success rate and response speed. Furthermore, at the execution level, CRC command verification, temperature rise protection, contact status feedback, and dynamic power supply control modules are integrated to ensure the safety and reliability of the entire charging process. This series of intelligent control strategies enables the system to achieve autonomous decision-making, dynamic adjustment, and anti-interference capabilities, significantly enhancing the automation level and user experience of the charging pile system. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is the overall flow chart of the present invention;

[0067] The meaning of each number in the figure is:

[0068] 1. Infrared sensor detection unit; 11. Infrared signal receiving module; 12. Infrared signal processing module; 2. Signal judgment and control unit; 21. Signal judgment module; 22. Control instruction generation module; 3. Charging gun head control unit; 4. Dynamic power supply control unit; 41. Dynamic power supply module; 42. Dynamic power outage module. DETAILED DESCRIPTION

[0069] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0070] Example: See Figure 1 As shown, an automatic charging system for charging piles based on adaptive infrared sensing is provided, including:

[0071] The infrared sensing detection unit 1 actively transmits infrared signals and receives reflected signals. Based on the infrared signal reflected echo, the target spatial orientation, distance, and charging port position are obtained through time difference and signal strength. In this embodiment, the target is a vehicle;

[0072] In this embodiment, the infrared sensing detection unit 1 includes an infrared signal receiving module 11 and an infrared signal processing module 12;

[0073] The infrared signal receiving module 11 actively transmits a modulated infrared pulse or continuous wave signal according to a preset frequency n or a trigger condition, and synchronously receives the infrared signal reflected by the target through the infrared receiver array. In this embodiment, the trigger condition includes detecting the approach of the target, and the infrared receiver array is a multi-point receiving probe;

[0074] The infrared signal processing module 12 includes a distance calculation module, a target orientation solution module and a charging port identification module;

[0075] The distance calculation module calculates the distance of the target object based on the signal propagation time difference and the speed of light;

[0076] The target direction calculation module calculates the target's direction based on the time difference of arrival (TDOA) and the signal strength difference, and introduces polarization phase entropy in the calculation process to distinguish and suppress signal distortion caused by multipath interference and ambient humidity. In this embodiment, the target's direction refers to the azimuth angle and the pitch angle.

[0077] In practical applications, infrared signals are prone to multiple reflection paths on highly reflective surfaces such as metal, resulting in delay estimation errors. Furthermore, in high-humidity environments, infrared signals are affected by water molecule absorption, resulting in nonlinear attenuation, further weakening signal stability. Therefore, polarization phase entropy (PPE) is introduced as a discriminant and control factor to quantify the degree of distortion, converting the invisible molecular absorption effect into a measurable polarization entropy change signal, and realizing a physical-level distortion compensation closed loop. This can effectively identify and filter out indirect path signals and suppress multipath interference. Furthermore, PPE is integrated with environmental parameters to achieve adaptive compensation, enhancing the system's robustness to material and weather changes, thereby significantly improving the positioning accuracy and reliability of charging ports or targets.

[0078] The target bearing solution module calculates the target's bearing based on the time difference of arrival (TDOA) and signal strength difference, and introduces polarization phase entropy into the solution process. The module includes the following steps:

[0079] S1.1. Receive the original infrared signal, extract the polarization phase characteristics of each channel of the infrared signal, and calculate the polarization phase entropy based on the polarization phase angle distribution , determine whether the signal comes from a direct path. If the PPE value is less than a, select the earliest arriving signal as the valid input. If the PPE value is greater than b, use the NLOS (non-line-of-sight) signal suppression algorithm to filter multipath interference;

[0080] Given that humidity interference on infrared sensing is essentially due to the selective absorption of infrared wavelengths by water molecules, resulting in signal attenuation and nonlinear distortion, polarization phase entropy is calculated based on the polarization phase angle distribution. The NLOS (non-line-of-sight) signal suppression algorithm is used to filter multipath interference, leveraging the concentrated polarization distribution of metal targets to actively eliminate multipath paths and enhance positioning robustness.

[0081] In this embodiment, a is 0.3, which shows the reflection characteristics of a typical metal surface, and b is 0.6;

[0082] Calculation of polarization phase entropy based on polarization phase angle distribution The specific formula is:

[0083] ;

[0084] Where, Indicates the The central angle value of the polarization phase angle interval, in radians or degrees, represents the polarization state of light. Indicates the percentage of pixels falling into this polarization phase interval, that is, the phase angle is The ratio of the number of nearby pixels to the total number of pixels, represents the number of subintervals into which the entire polarization phase angle range is divided, represents the subinterval index;

[0085] S1.2. Calculate the time delay difference between a pair of receivers, perform temperature compensation on the baseline between receivers based on the ambient temperature, and calculate the initial azimuth angle based on the calibration baseline and time delay difference. The temperature compensation of the baseline between receivers based on the ambient temperature is because the ambient temperature change will cause the thermal expansion / contraction of the baseline length between sensors, which in turn affects the azimuth calculated based on TDOA. The baseline length is corrected by the temperature compensation function to ensure that the initial azimuth angle calculated is more accurate and reliable. It is especially suitable for outdoor and long-term operation equipment.

[0086] Calculate the initial azimuth The formula is:

[0087] ;

[0088] Where, represents the time delay difference between a pair of receivers. By performing cross-correlation operation on any pair of receivers, the time delay difference at the maximum cross-correlation peak is obtained. represents the signal propagation speed, represents the baseline length of the temperature-compensated receiver array, i.e., the physical distance between two receivers. , is the original baseline length, is the current ambient temperature, is the standard reference temperature, represents the temperature change, is the temperature expansion coefficient, which is determined by the material properties;

[0089] S1.3. Compensation factor calculation using the fusion model of humidity and polarization phase entropy , based on the compensation factor for the initial azimuth Perform nonlinear environmental compensation to significantly reduce the impact of air attenuation on angle measurement in high humidity environments such as rain and fog;

[0090] Calculate the compensation factor formula for: , is the weight coefficient of PPE, dimensionless, is the basic compensation constant, dimensionless. In this embodiment, there is no humidity sensor. , ;

[0091] Based on the compensation factor for the initial azimuth The nonlinear environmental compensation is:

[0092] ;

[0093] Where, is the corrected azimuth after compensation, is the nonlinear adjustment parameter that controls the exponential decay rate, The unit is the reciprocal of the square of radians;

[0094] S1.4. Normalized signal intensity difference Combined with polarization phase entropy, dynamic adjustment of fusion weight factor , so that the solution has the ability to adapt to different reflective surfaces and output the final fusion azimuth ;

[0095] Final fusion azimuth for:

[0096] ;

[0097] S1.5. Calculate the vertical TDOA difference , that is, the height difference in the vertical direction, estimate the target height difference, and calculate the target pitch angle through the elevation angle formula , and performs elevation angle correction based on polarization phase entropy to eliminate the elevation angle deviation caused by ground reflection. This correction uses the polarization entropy characteristics of mixed reflection to achieve quantitative removal and correction of ground false signals;

[0098] ;

[0099] Where, is the horizontal projection distance, which refers to the horizontal baseline distance or the projection distance on the horizontal plane. The pitch angle correction coefficient function based on polarization phase entropy is used to compensate for the deviation introduced by ground or surface reflection. , is the basic correction coefficient constant, reflecting the ideal correction value when there is no PPE influence. is the PPE weight coefficient, which quantifies the influence of polarization phase entropy on the correction amount.

[0100] The charging port identification module identifies the specific charging port location using target surface features;

[0101] The paint on the vehicle body and the metal charging port frame create bright areas in the infrared image, causing traditional recognition methods to mistakenly identify the reflection as the charging port. Appearance features (such as color and shape) become unreliable in rain, fog, stains, or when there is insufficient light at night, making missed detections / false detections more likely. Traditional methods ignore the thermal inertia differences of the materials themselves and lack the ability to utilize the inherent thermal response characteristics of the materials for identification. TIIE (thermal inertia information entropy) parameters are used to convert the thermal inertia differences of materials into quantifiable indicators to distinguish between metals and non-metals. By combining active thermal excitation with material thermal inertia feature extraction, a TIIE entropy value recognition mechanism is constructed, effectively solving the problem of misjudgment in traditional recognition in complex scenarios such as strong reflections, low light, rain, snow, and stains.

[0102] The charging port identification module uses the target surface features to identify the specific charging port location, including the following steps:

[0103] S1.6. Collect the distance and orientation information of the target and acquire raw thermal imaging data. This data reflects the temperature distribution of the target area. The main purpose of collecting raw thermal imaging data is to use the thermal inertia differences of the materials to identify the specific location of the charging port.

[0104] S1.7. Based on the target location, the search area for the charging port is predicted. Areas with a reflectivity exceeding d and made of metal are selected as candidate areas. A 980nm wavelength infrared laser pulse is applied to the selected target area for heating. The temperature rise is monitored in real time to ensure safety. If the temperature rise exceeds 45°C, the power is automatically reduced to 0.8W to avoid surface damage. A temperature rise threshold (45°C) and adaptive power adjustment (0.8W fuse) are introduced to ensure that laser heating does not damage the target surface. An active excitation safety protection mechanism is integrated into the charging port recognition.

[0105] S1.8. Collect thermal decay data for the target area within x seconds after heating, construct a thermal decay probability distribution, and calculate the TIIE value. The TIIE value reflects the difference in thermal inertia between different materials. Metal areas have a fast and uniform thermal decay characteristic, with a TIIE close to 0. Non-metal areas exhibit a large random slow decay characteristic, with a TIIE value greater than 0.7.

[0106] In this embodiment, the purpose of constructing the thermal attenuation probability distribution and calculating the TIIE value is to convert the thermodynamic properties of the material into a quantifiable entropy index to solve the problem of misjudgment of metal reflections and paint surfaces;

[0107] The construction of thermal attenuation probability distribution and calculation of TIIE value include the following steps:

[0108] S1.81. Use infrared thermal imaging equipment to record each frame of the target area, which contains the surface temperature information of each pixel. This data constitutes a temperature field sequence that changes over time.

[0109] S1.82. For each pixel, calculate the rate of change of temperature between successive time frames. This step generates a new matrix or array representing the temperature change of each pixel over time.

[0110] S1.83. Combine the temperature change rate values ​​of all pixels into a one-dimensional array. Within a pre-set temperature change rate range, divide this range into e intervals of equal width. Count the number of occurrences of the temperature change rate value in each interval to form a histogram. This histogram reflects the distribution of different temperature change rates across the entire target area. In this embodiment, the temperature change rate range is -0.5°C to 0°C.

[0111] S1.84. Normalize the histogram to obtain the thermal decay probability distribution. Normalization is performed so that the sum of the frequencies of all intervals is equal to 1.

[0112] S1.85. Use the information entropy formula to calculate the TIIE value to quantify the uncertainty in the probability distribution of the temperature change rate;

[0113] The TIIE value is calculated using the information entropy formula:

[0114] ;

[0115] Where, express The temperature change rate, Indicates that among all temperature change rates, The probability distribution value of occurrence, Indicates the length of the temperature time series.

[0116] S1.9. Create a metal contact mask with a TIIE value less than y. Further screening is performed based on geometric features. Ultimately, the charging port border area that best matches the charging port characteristics is selected. Generate a metal contact mask based on a threshold where the TIIE value is less than y. This directly targets the physical characteristics of the charging port metal border, rather than relying on appearance features. This overcomes the failure of traditional visual solutions in strong reflections, rain, fog, and stains. Geometric features include diameter and depth.

[0117] S1.10. Perform three-dimensional fitting on the mask area where the TIIE value is less than the threshold y, extract the central coordinates of the area as the center position of the charging port, and estimate the average normal direction of the charging port border area based on the point cloud surface normal vector of the charging port border area for the charging gun head alignment posture.

[0118] The signal judgment and control unit 2 determines whether the target has entered the charging area, whether it has parked accurately, and whether the charging port is aligned based on the spatial position analysis and geometric feature matching of the infrared sensor data. Based on the judgment results, it uses manifold elastic drive and geometric estimation control technology to generate control instructions;

[0119] In this embodiment, the signal judgment and control unit 2 includes a signal judgment module 21 and a control instruction generation module 22;

[0120] The signal judgment module 21 receives the infrared sensor data output by the infrared signal processing module 12, analyzes and determines whether the target has entered the charging area, whether it has been accurately parked in place, and whether the charging port is within the identifiable range;

[0121] The signal judgment module 21 includes a target entry and exit judgment module, a target parking judgment module, and a charging port judgment module;

[0122] The target entry and exit judgment module receives the target distance and direction data output by the infrared sensor detection unit 1, and judges whether the target enters and leaves the charging area through a dynamic threshold algorithm, wherein the dynamic threshold algorithm is as follows: when the target distance is less than a preset entry threshold a and meets the target size characteristics, a target entry state flag is triggered; when the target disappears for a period exceeding a time threshold b or the distance exceeds an exit threshold c, a target exit signal is generated;

[0123] The target parking judgment module analyzes the target position coordinates and motion status in real time, and determines whether the target has been parked through a dual verification mechanism of position and speed. Among them, position verification is: using infrared sensing, visual recognition or laser ranging to obtain the position relationship between the vehicle or target and the charging pile in real time, determine whether it has entered the preset parking area, and compare it with the standard docking position of the charging port to determine whether the position error is within an acceptable range (± a few centimeters). Speed ​​verification is: combining the displacement difference of the inertial sensor or continuous frame images to monitor the speed change of the target in a short period of time to determine whether it has approached a stationary state; the dual verification mechanism of position and speed is: only when the two conditions of "position deviation within the threshold" and "speed close to zero" are met at the same time, is it determined that the target has been correctly parked.

[0124] The charging port judgment module uses a reachability mechanism to determine whether the charging port coordinates are within the working space of the charging gun. In this embodiment, the working space of the charging gun, or the execution end of the automatic docking system, needs to be clearly defined. The working space refers to the set of all positions and postures that the charging gun can reach without considering any obstacles. The charging port center position coordinates and the average normal direction are converted to the local coordinate system of the charging gun. The inverse kinematics (IK) algorithm is used to calculate the specific joint angle combination required for the charging gun end to reach the target position. Physical reachability and obstacle-free path planning are performed. Among them, physical reachability refers to whether a given target position can be achieved by adjusting the various joints of the robotic arm. This is evaluated by comparing the existence and rationality of the IK solution. The obstacle-free path planning ensures that no obstacles are encountered during the movement from the current position to the target position.

[0125] The control instruction generation module 22 generates a charging gun head control instruction to instruct the charging gun head control unit 3 to act according to the judgment result of the signal judgment module 21, using the manifold elastic drive and geometric estimation control technology, and sends a power supply start and stop control instruction to the dynamic power supply control unit 4 after the charging gun head action is completed;

[0126] The manifold elastic drive and geometric estimation control technology integrates inverse kinematics, impedance control, manifold constraint optimization, and disturbance prediction and compensation to generate interference-resistant, high-precision, and dynamically responsive charging gun head control instructions, enabling safe and efficient automatic docking of charging ports in complex environments.

[0127] Based on the judgment result of the signal judgment module 21, a charging gun head control instruction for instructing the charging gun head control unit 3 to operate is generated by using manifold elastic drive and geometric estimation control technology, including the following steps:

[0128] S2.1. Receive the judgment result of the signal judgment module 21 and the center position and average normal direction of the charging port;

[0129] S2.2. Based on the current spatial position and posture information of the charging gun head, an inverse kinematics (IK) method is used to calculate candidate paths from the current posture to the target charging port position, generating an initial target posture sequence. The initial target posture sequence is generated by: reading the current position and posture of the charging gun head end effector in real time; setting the target position to the center of the charging port, and the target posture to align the charging gun axis with the normal direction; using a numerical iteration method to solve the joint angle sequence, generate multiple candidate paths, and perform preliminary evaluation; and output the initial target posture sequence. In this embodiment, the numerical iteration method includes the Jacobian transpose / pseudo-inverse method.

[0130] S2.3. Call the impedance control module and manifold constraint module to perform multi-factor correction, and input the output results of the impedance control module and manifold constraint module into the coupling matrix, perform fusion operation, and generate the final path instruction Joint torque command This system achieves high-precision, high-safety, and high-responsiveness in automatic docking of the charging gun tip to the target charging port in dynamic environments. The impedance control module enables real-time sensing and suppression of vibration interference during the robot arm's motion, dynamically adjusting stiffness and damping properties to ensure positioning accuracy while preventing damage caused by contact impact. The manifold constraint module constructs a safe manifold space based on the geometric information of the surrounding space, optimizing path planning and enabling the robot arm to avoid obstacles and follow the optimal geodesic path even in complex or confined environments. The data from these two systems is integrated through a cross-coupling IMC matrix. This not only enables coordinated control of physical behaviors (such as force control and posture), but also resolves the fundamental contradictions between "precision and compliance" and "efficiency and safety" in traditional methods. This generates high-quality path and torque commands that balance dynamic response, obstacle avoidance, and contact stability, significantly improving the intelligence and practical reliability of the charging pile system. The IMC matrix (coupling control matrix) is the data mapping matrix used to integrate impedance control and manifold constraint optimization in the automatic charging system. By coordinating torque control and trajectory constraints, it achieves high-precision and high-stability charging gun tip path planning and torque command output.

[0131] The impedance control module is used to detect the speed and acceleration of each joint of the charging gun robotic arm in real time. It uses FFT to extract low-frequency resonant components and constructs an adaptive impedance model to output torque correction terms to suppress vibration at the end of the charging gun tip.

[0132] Furthermore, the adaptive impedance model is:

[0133] ;

[0134] Where, Indicates the control torque command, which is the torque value that the motor or actuator needs to apply, calculated based on the difference between the current state and the target state. It is used to drive the actuators of the joints of the charging gun head. Represents the impedance function in the frequency domain, reflecting the stiffness and damping (vibration suppression) of the charging gun head at different frequencies. Indicates frequency, Indicates the expected acceleration, which is the linear acceleration of the end effector of the charging gun head. Indicates the actual acceleration;

[0135] The manifold constraint module is used to construct a signed distance field based on the environmental point cloud, define the safety manifold metric tensor, reflect the local space curvature and obstacle proximity, solve the geodesic path on the Riemannian manifold, and output the corrected path. , introducing Riemannian geometry into the control of the charging robot arm to improve the efficiency of the obstacle avoidance path;

[0136] The formula for solving the geodesic path on a Riemannian manifold is:

[0137] ;

[0138] Where, represents the time-integrated variable along the path, represents the joint velocity vector, represents the transpose operation, Represents the metric tensor, which defines the curvature of the space where the charging gun head is located;

[0139] Furthermore, the output results of the impedance control module and the manifold constraint module are input into the coupling matrix to perform a fusion operation. That is, the impedance control term suppresses low-frequency mechanical vibrations, and the manifold path term realizes the reasonable planning of obstacle avoidance and constraint paths. The coupling matrix coordinates the two and finally outputs the execution torque and trajectory instructions:

[0140] ;

[0141] Where, Represents the target position vector of the terminal trajectory control, that is, the position path point where the charging gun head should move to, including disturbance compensation. Represents the coupling control matrix, which integrates the information mapping matrix of impedance regulation and manifold planning. It is a multi-channel gain matrix used to weigh the coupling weights of torque control and trajectory constraints. represents the terminal acceleration vector, which is obtained from the second-order derivative of the trajectory, represents the target path solution under manifold constraints, that is, the ideal path state after considering obstacles and structural constraints, including geodesic solutions, Represents the current joint angle vector of the robotic arm, Represents external interference force / disturbance vector, including wind force, target vibration, and environmental disturbance;

[0142] S2.4. Analyze the micro-motion trajectory of the charging port center over a historical period. This captures subtle movement patterns of the charging port caused by external factors. Using bandpass filtering and ARIMA prediction, a breathing disturbance model is established. A 0.3-second lead compensation is applied to the trajectory vector of the charging gun tip to generate the compensated final target trajectory. External factors include adjustments to the target suspension system, uneven ground, and wind.

[0143] Furthermore, this solution addresses the problem of periodic micro-excursions of the charging port caused by airflow disturbances, attitude fine-tuning, or unstable hovering when the drone or target is in the waiting state. The goal is to extract the time series characteristics of these micro-perturbations and construct a predictable "breathing" disturbance model, thereby compensating for position fluctuations in path planning and contact control in advance, and improving the robustness and response speed of the docking process. The advantages of this method are: first, the low-frequency components that affect the position of the charging port can be effectively extracted through a bandpass filter, eliminating high-frequency noise interference; second, the ARIMA model is used to predict the disturbance trend in the short term in the future, and the motion path of the charging gun head can be adjusted in advance, thereby significantly improving the success rate and efficiency of docking, reducing docking failures or poor contact caused by small movements of the target, and ultimately improving the safety of the entire charging process and user experience. Integrating the above content, the purpose and advantage of this step is to enhance the adaptability of the charging device in complex environments by accurately modeling and predicting the breathing disturbance of the target charging port, achieving efficient and reliable automatic docking. It can effectively distinguish between stable drift and high-frequency disturbances, reduce the dependence on the real-time response capability of the end effector, and improve the docking success rate and robustness of the dynamic power supply system in non-rigid target environments.

[0144] Analyze the micro-motion trajectory of the charging port center over the historical time period and establish a breathing disturbance model using bandpass filtering and ARIMA prediction methods. The following steps are included:

[0145] S2.41. Collect the coordinates of the charging port center for each frame in the past 10 to 15 seconds to form a time series ;

[0146] S2.42. Apply a bandpass filter to retain only the signal components in the range z to u. In order to separate the low-frequency disturbance caused by the target suspension system from the historical position data of the charging port, a filtered breathing micro-motion signal is generated. ;

[0147] In this embodiment, z is 0.05 Hz and u is 0.15 Hz;

[0148] S2.43, respiratory micro-movement signal Perform differential operations in the X, Y, and Z directions to achieve stabilization;

[0149] S2.44, based on the differential respiratory micro-motion signal sequence, use the ARIMA model to predict the future Make advance predictions within seconds and generate future disturbance increments ;

[0150] ;

[0151] ;

[0152] Where, Indicates the current time The respiratory micro-motion position vector, Indicates time;

[0153] S2.45. Apply the predicted disturbance increment to all control point coordinates in the original inverse kinematics path planning result to obtain the control path after disturbance compensation. ;

[0154] ;

[0155] Where, Control cycle for path planning;

[0156] S2.5. Quantize and encode the final trajectory vector and control torque to generate a control instruction packet that is resistant to electromagnetic interference;

[0157] S2.6. Send the control instruction packet to the charging gun head control unit 3.

[0158] The charging gun head control unit 3 receives the control instruction and drives the actuator to control the charging gun head;

[0159] In this embodiment, the charging gun head control unit 3 receives a control instruction and drives the actuator to control the charging gun head, including the following steps:

[0160] S3.1. Receive a control instruction packet from Signal Judgment and Control Unit 2. This control instruction is generated based on the target charging port posture prediction result and the real-time positioning error correction result, and contains the spatial position and posture parameters required by the charging gun head. Perform a CRC-16 integrity check on the instruction packet. CRC-16 is a 16-bit cyclic redundancy check algorithm used to detect errors during data transmission or storage to ensure that there are no data errors caused by electromagnetic interference.

[0161] S3.2. Decode the control instruction packet and restore it to the real physical quantity;

[0162] S3.3. Analyze the restored control instructions, extract the translation instructions and rotation instructions in the three-dimensional coordinate system, perform unit and precision conversion, and generate low-level motion control parameters suitable for the actuator;

[0163] S3.4. Send low-level motion control parameters to the actuator control modules at all levels for mechanical docking, including but not limited to: linear motors for coarse positioning, servo motors for posture fine-tuning, and end-to-end micro-motion compensation mechanisms for approaching docking. After the mechanical docking is completed, key parameters such as docking status, electrical contact resistance, and contact pressure are fed back to the signal judgment and control unit 2.

[0164] The dynamic power supply control unit 4 performs dynamic power supply according to the control instruction;

[0165] In this embodiment, the dynamic power supply control unit 4 includes a dynamic power supply module 41 and a dynamic power outage module 42;

[0166] The dynamic power supply module 41 receives the signal judgment and charging instruction of the control unit 2, controls the main relay or solid-state switch to close, connects the power path in real time, and monitors various parameters of the charging process in real time, including voltage, current, temperature, etc.;

[0167] The dynamic power-off module 42 is used to quickly disconnect the power supply path when an abnormal state occurs during the charging process. The abnormal state includes gun head posture drift, arc abnormality, insufficient contact pressure, overtemperature, accidental contact, user termination command, etc.

[0168] The basic principles, main features, and advantages of the present invention are shown and described above. It should be understood by those skilled in the art that the present invention is not limited to the above-described embodiments. The above-described embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention, and such changes and modifications fall within the scope of the invention claimed.

Claims

1. The automatic charging system of charging pile based on adaptive infrared sensing is characterized by: include: An infrared sensing detection unit (1), the infrared sensing detection unit (1) actively transmits an infrared signal and receives a reflected signal, and obtains a target spatial orientation, distance, and charging port position based on a reflected echo of the infrared signal by using a time difference and signal strength; The infrared sensing detection unit (1) includes an infrared signal receiving module (11) and an infrared signal processing module (12); The infrared signal receiving module (11) actively transmits modulated infrared pulses at a preset frequency n, and synchronously receives infrared signals reflected by the target through an infrared receiver array; The infrared signal processing module (12) includes a distance calculation module, a target orientation solution module and a charging port identification module; The distance calculation module calculates the distance of the target object based on the signal propagation time difference and the speed of light; The target direction calculation module calculates the target's direction based on the arrival time difference and signal strength difference, and introduces polarization phase entropy in the calculation process to distinguish and suppress signal distortion caused by multipath interference and ambient humidity; The charging port identification module identifies the specific charging port location using target surface features; A signal judgment and control unit (2), wherein the signal judgment and control unit (2) judges whether the target has entered the charging area, whether it has been accurately parked in place, and whether the charging port is aligned based on spatial position analysis and geometric feature matching of infrared sensor data, and generates control instructions based on the judgment results using manifold elastic drive and geometric estimation control technology; The signal judgment and control unit (2) includes a signal judgment module (21) and a control instruction generation module (22); The signal judgment module (21) receives the infrared sensing data output by the infrared signal processing module (12), analyzes and judges whether the target has entered the charging area, whether it has been accurately parked in place, and whether the charging port is within the identifiable range; The signal judgment module (21) includes a target entry and exit judgment module, a target parking judgment module, and a charging port judgment module; The target entry and exit judgment module receives the target distance and orientation data output by the infrared sensor detection unit (1), and judges whether the target enters and leaves the charging area through a dynamic threshold algorithm; The target parking judgment module analyzes the target position coordinates and motion status in real time, and determines whether the target has been parked through a dual verification mechanism of position and speed; The charging port judgment module determines whether the charging port coordinates are within the charging gun working space through the accessibility mechanism; The control instruction generation module (22) generates a charging gun head control instruction for instructing the charging gun head control unit (3) to act according to the judgment result of the signal judgment module (21), and sends a power supply start-stop control instruction to the dynamic power supply control unit (4) after the charging gun head action is completed; According to the judgment result of the signal judgment module (21), a charging gun head control instruction for instructing the charging gun head control unit (3) to act is generated by using manifold elastic drive and geometric estimation control technology, including the following steps: S2.1, receiving the judgment result of the signal judgment module (21) and the center position and average normal direction of the charging port; S2.

2. Based on the current spatial position and posture information of the charging gun head, use the inverse kinematics method to calculate the candidate path from the current posture to the target charging port position and generate the initial target posture sequence; S2.

3. Call the impedance control module and the manifold constraint module to perform multi-factor correction, input the output results of the impedance control module and the manifold constraint module into the coupling matrix, perform fusion operation, and generate the final path command and joint torque command; The impedance control module is used to detect the speed and acceleration of each joint of the charging gun robotic arm in real time. It uses FFT to extract low-frequency resonant components and constructs an adaptive impedance model to output torque correction terms to suppress vibration at the end of the charging gun tip. The manifold constraint module is used to construct a signed distance field based on the environment point cloud, define the safe manifold metric tensor, solve the geodesic path on the Riemannian manifold, and output the corrected path; S2.

4. Analyze the micro-motion trajectory of the charging port center over a historical period. Establish a breathing disturbance model using bandpass filtering and ARIMA prediction methods. Apply advance compensation to the trajectory vector of the charging gun tip to generate the compensated final target trajectory. S2.

5. Quantize and encode the final trajectory vector and control torque to generate a control instruction packet that is resistant to electromagnetic interference; S2.6, sending the control instruction packet to the charging gun head control unit (3); A charging gun head control unit (3), the charging gun head control unit (3) receiving a control instruction and driving an actuator to control the charging gun head; A dynamic power supply control unit (4), wherein the dynamic power supply control unit (4) performs dynamic power supply according to a control instruction.

2. The automatic charging system for charging piles based on adaptive infrared sensing according to claim 1 is characterized in that: The target direction calculation module calculates the target direction based on the arrival time difference and the signal strength difference, and introduces polarization phase entropy in the calculation process, including the following steps: S1.

1. Receive the original infrared signal, extract the polarization phase characteristics of each channel of the infrared signal, calculate the polarization phase entropy based on the polarization phase angle distribution, and determine whether the signal comes from a direct path; S1.

2. Calculate the time delay difference between a pair of receivers, perform temperature compensation on the inter-receiver baseline based on the ambient temperature, and calculate the initial azimuth based on the calibrated baseline and time delay difference; S1.

3. Calculate a compensation factor using polarization phase entropy, and perform nonlinear environmental compensation on the initial azimuth angle based on the compensation factor; S1.

4. Combine the normalized signal strength difference with the polarization phase entropy, dynamically adjust the fusion weight factor, and output the final fusion azimuth. S1.

5. Calculate the vertical TDOA difference, estimate the target height difference, calculate the target pitch angle using the elevation angle formula, and perform elevation angle correction based on polarization phase entropy to eliminate the elevation angle deviation caused by ground reflection.

3. The automatic charging system for charging piles based on adaptive infrared sensing according to claim 1 is characterized in that: The charging port identification module identifies the specific location of the charging port using the target surface features, including the following steps: S1.

6. Collect target distance and orientation information and acquire thermal imaging raw data; S1.

7. Predict the search area for the charging port based on the target location, and select areas with a reflectivity exceeding d and made of metal as candidate areas. S1.

8. Collect thermal attenuation data of the target area within x seconds after being heated, construct the thermal attenuation probability distribution, and calculate the TIIE value; S1.

9. Create a metal contact mask with a TIIE value less than y, further screen it based on geometric features, and ultimately select the charging port border area that best matches the charging port characteristics. S1.

10. Perform three-dimensional fitting on the mask area where the TIIE value is less than the threshold y, extract the central coordinates of the area as the center position of the charging port, and estimate the average normal direction of the charging port border area based on the point cloud surface normal vector of the charging port border area.

4. The automatic charging system for charging piles based on adaptive infrared sensing according to claim 3 is characterized in that: In S1.8, constructing the thermal attenuation probability distribution and calculating the TIIE value includes the following steps: S1.

81. Use infrared thermal imaging equipment to record each frame of the image within the target area; S1.

82. For each pixel, calculate the rate of change of temperature between consecutive time frames; S1.

83. Combine the temperature change rate values ​​of all pixels into a one-dimensional array. Within a pre-set temperature change rate range, divide this range into e intervals of equal width. Count the number of occurrences of the temperature change rate value in each interval to form a histogram. S1.

84. Normalize the histogram to obtain the thermal attenuation probability distribution; S1.

85. Use the information entropy formula to calculate the TIIE value to quantify the uncertainty in the probability distribution of the temperature change rate.

5. The automatic charging system for charging piles based on adaptive infrared sensing according to claim 1 is characterized in that: In S2.4, the micro-motion trajectory of the charging port center in the historical time period is analyzed, and a breathing disturbance model is established through bandpass filtering and ARIMA prediction method, including the following steps: S2.

41. Collect the coordinates of the center of the charging port for each frame in the past 10 to 15 seconds to form a time series. S2.42, applying a bandpass filter to retain only the signal components in the range z to u, thereby generating a filtered respiratory micromotion signal; S2.43, performing differential operations on the respiratory micro-motion signal in the X, Y, and Z directions respectively; S2.44, based on the differential respiratory micro-motion signal sequence, use the ARIMA model to predict the future Make advance predictions within seconds and generate future disturbance increments; S2.

45. Apply the predicted disturbance increment to all control point coordinates in the original inverse kinematics path planning result to obtain a control path after disturbance compensation.

6. The automatic charging system for charging piles based on adaptive infrared sensing according to claim 1 is characterized in that: The charging gun head control unit (3) receives a control instruction and drives the actuator to control the charging gun head, comprising the following steps: S3.1, receiving the control instruction packet from the signal judgment and control unit (2), and performing integrity check on the instruction packet using CRC-16; S3.

2. Decode the control instruction packet and restore it to the real physical quantity; S3.

3. Analyze the restored control instructions, extract the translation instructions and rotation instructions in the three-dimensional coordinate system, perform unit and precision conversion, and generate low-level motion control parameters; S3.

4. Send the low-level motion control parameters to the actuator control modules at all levels for mechanical docking. After the mechanical docking is completed, the key parameters are fed back to the signal judgment and control unit (2).

7. The automatic charging system for charging piles based on adaptive infrared sensing according to claim 1 is characterized in that: The dynamic power supply control unit (4) comprises a dynamic power supply module (41) and a dynamic power outage module (42); The dynamic power supply module (41) receives the charging instruction from the signal judgment and control unit (2), controls the main relay to close, connects the power path in real time, and monitors various parameters during the charging process in real time; The dynamic power-off module (42) is used to quickly disconnect the power supply path when an abnormal state occurs during the charging process.

Images