AGV collision detection method based on motor current change

Abstract

The present application relates to a kind of AGV collision detection methods based on motor current variation, comprising the following steps: S1, the motor current signal in real time acquisition in AGV and the motion state parameter of AGV;S2, according to the current signal and the motion state parameter of AGV obtained by acquisition training to obtain dynamic current baseline model;S3, the motor current signal of real time acquisition is compared with dynamic current baseline model, calculates out current deviation and extracts collision feature;S4, collision feature is compared with pre-set threshold, combination determines whether there is collision event.The method of the present application does not depend on additional physical sensor, but reuses the motor current signal and motion state parameter inherent in AGV driving system, significantly reduces hardware cost and system complexity;And by establishing "dynamic current baseline model", fundamentally solves the core problem of high false alarm rate and high false negative rate of traditional fixed current threshold method, improves the comprehensiveness and reliability of detection.

Description

Technical Field

[0001] This invention relates to the field of AGV collision detection technology, specifically to an AGV collision detection method based on changes in motor current. Background Technology

[0002] Automated Guided Vehicles (AGVs), as highly efficient and flexible material handling equipment, have been widely used in warehousing and logistics, intelligent manufacturing, and port terminals. Ensuring the safety of AGVs during autonomous operation, particularly in relation to collisions with the environment, goods, and personnel, is a core prerequisite for the reliable and efficient operation of the system and a key technology for achieving a human-machine harmonious environment.

[0003] Currently, the AGV collision detection technologies used in the industry are mainly divided into two categories: contact and non-contact.

[0004] I. Contact-based Collision Detection Technology: This type of technology directly senses collisions by installing physical sensors on the AGV body. The most common method is to install mechanical anti-collision strips or edges on the front, rear, and sides of the AGV. When physical contact occurs, an internal switch is triggered, generating an emergency stop signal. Its advantages are simple principle, low cost, and high reliability. However, its disadvantages are also quite obvious: First, it is a passive detection method, requiring physical contact and deformation to occur before a response is generated, resulting in a reaction delay; second, frequent collisions can lead to mechanical fatigue and damage, increasing maintenance costs; finally, it may not effectively trigger the system for some minor or non-frontal collisions.

[0005] II. Non-contact Collision Detection Technology: To anticipate and avoid collisions earlier, non-contact technologies are widely researched and applied. These mainly include: 1. LiDAR / Laser Scanner: Constructs an environmental map by emitting laser beams and receiving reflected signals, achieving high-precision obstacle detection and avoidance. This is currently the mainstream solution for high-end AGVs, but it is expensive and has blind spots or errors in detecting reflective surfaces, glass, and other special objects. 2. Ultrasonic Sensor: Detects nearby obstacles using ultrasonic ranging principles. It is relatively inexpensive, but has a narrow detection angle, is easily affected by environmental noise (such as ultrasonic waves from other AGVs), and performs poorly on soft or sound-absorbing materials. 3. Vision Sensor (Camera): Determines obstacles ahead using image recognition technology. It can provide rich environmental information, but is greatly affected by lighting conditions and computational complexity, posing significant challenges in real-time performance and robustness. 4. Infrared Sensor: Similar in principle to ultrasonic sensors, but more susceptible to interference from ambient light (especially sunlight), typically used for short-range detection.

[0006] The common problems with non-contact sensors are: increased hardware costs and installation complexity; the sensors themselves have the risk of detection blind spots, false alarms (such as misjudging uneven ground or changes in light and shadow as obstacles) or missed alarms; and in complex and dynamic environments, the algorithmic complexity of multi-sensor information fusion and decision-making is high. Summary of the Invention

[0007] To address the aforementioned technical problems, the present invention aims to provide an AGV collision detection method based on motor current changes. This method integrates motion state parameters in real time, learns and predicts the theoretical motor current online, and extracts intelligent collision features for judgment. Thus, without the need for additional hardware, it achieves low-cost, high-reliability, and low-false-alarm collision detection, effectively improving the safety and environmental adaptability of AGVs.

[0008] To achieve the objectives of the invention described above, the present invention adopts the following technical solution:

[0009] An AGV collision detection method based on motor current variation includes the following steps:

[0010] S1. Real-time acquisition of current signals of each motor in the AGV and motion status parameters of the AGV;

[0011] S2. A dynamic current baseline model is trained based on the collected current signal and the motion state parameters of the AGV.

[0012] S3. Compare the real-time acquired motor current signal with the dynamic current baseline model, calculate the current deviation and extract the collision features.

[0013] S4. Compare the collision features with the preset threshold and combine them to determine whether a collision event exists.

[0014] As a preferred embodiment, in step S1, the current signals of each motor in the AGV are collected in real time. The sampling frequency is ≥ 10Hz; and the motion state parameters of the AGV are acquired synchronously, including speed v, acceleration a, steering angle θ and load w.

[0015] As a preferred embodiment, in step S2, the theoretical current baseline value is calculated based on the current vehicle motion state and load using a preset motor load model. Its form is: ,

[0016] By collecting speed, acceleration, steering angle, and load information at each moment. and the motor current at the corresponding moment Using regression analysis, the input was learned through the Scikit-learn library. and Train them to obtain the functional relationship between them. .

[0017] As a preferred embodiment, the current deviation is calculated in step S3 as follows: ,in For based on the input Using functional relationships The calculated reference current, This represents the current being collected in real time at the current moment.

[0018] As a preferred embodiment, in step S3, the extraction of collision features includes at least one of the following features:

[0019] A. Current sudden change amplitude threshold Under normal operating conditions Set it to be 3 times the variance;

[0020] B. Rate of change of current threshold Under normal operating conditions Set it to 3 times the variance.

[0021] As a preferred embodiment, in step S4, when one condition is met, or If a collision is detected, a safety mechanism such as emergency stop, alarm, or reporting will be triggered.

[0022] As a preferred embodiment, the motors in the AGV include a travel motor and a fork motor. Based on the action commands given by the AGV vehicle controller, the current of the corresponding motor is detected, and it is determined whether a collision has occurred.

[0023] As a preferred option, when the AGV uses dual drive motors, the current of each drive motor is detected separately to determine whether a collision has occurred.

[0024] The method of this invention does not rely on additional physical sensors (such as lidar or mechanical anti-collision strips), but reuses the inherent motor current signal and motion state parameters of the AGV drive system, which significantly reduces hardware costs and system complexity. Moreover, by establishing a "dynamic current baseline model", it fundamentally solves the core problems of high false alarm rate (misjudging normal acceleration and heavy load as collision) and high false alarm rate (setting a high threshold to reduce false alarms, resulting in the failure to detect minor collisions) of the traditional fixed current threshold method, thereby improving the comprehensiveness and reliability of detection.

[0025] The method of this invention also compares the real-time current with a baseline that dynamically changes with the AGV's motion state (speed, acceleration, steering, load), enabling the system to intelligently distinguish between "normal load changes" and "abnormal collision impacts," significantly improving detection accuracy and environmental adaptability. From data acquisition, model training, and feature extraction to threshold determination, the method of this invention defines a standardized and implementable automated process, forming a complete closed loop of detection, judgment, and response, providing a direct and reliable technical basis for the safety control of AGVs. Attached Figure Description

[0026] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute a limitation thereof.

[0027] Figure 1 This is a system architecture block diagram of the method implemented in this invention;

[0028] Figure 2 This is a diagram showing the normal operation of the AGV in the method of this invention and the changes in motor current during normal loading and unloading.

[0029] Figure 3 This is a graph showing the change in motor current after a collision or top-loading in the method of this invention. Detailed Implementation

[0030] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0031] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0032] Furthermore, in the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0033] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more, unless explicitly defined otherwise.

[0034] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0035] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "over," and "on top" of the second feature includes the first feature directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature includes the first feature directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0036] The present invention will be further described below with reference to the accompanying drawings and embodiments:

[0037] like Figures 1 to 3 As shown, an AGV collision detection method based on motor current variation includes the following steps:

[0038] Step S1: Real-time acquisition of current signals from each motor in the AGV and the motion status parameters of the AGV; in the above steps, the current signals of each motor in the AGV... The sampling frequency is ≥ 10Hz; the specified sampling frequency of ≥ 10Hz can effectively capture the rapid transient change characteristics of the motor current when a collision occurs, avoid the loss of key information due to slow sampling, and ensure the timeliness and sensitivity of the detection.

[0039] The motion parameters of the AGV include speed v, acceleration a, steering angle θ, and load w. Accurately and synchronously collecting speed, acceleration, steering angle, and load, these parameters collectively determine the theoretical load of the motor. This provides the necessary and sufficient data foundation for establishing a high-precision dynamic current baseline model in subsequent step S2, enabling the model to comprehensively reflect the normal current level under various operating conditions.

[0040] Step S2: Train a dynamic current baseline model based on the collected current signal and the AGV's motion state parameters; in the above steps, the theoretical current baseline value is calculated based on the current vehicle motion state and load using a preset motor load model. Its form is: By collecting speed, acceleration, steering angle, and load information at each moment and the motor current at the corresponding moment Using regression analysis, the input was learned through the Scikit-learn library. and Train them to obtain the functional relationship between them. .

[0041] The above process achieves accurate fitting and generalization capabilities of the model: by using regression analysis (such as using the Scikit-learn library) to automatically learn motion state parameters (v, a, θ, w) and actual motor current from historical operating data. Complex nonlinear functional relationships between them This method better reflects the characteristics of actual equipment than preset theoretical formulas, and can automatically compensate for unmodeled factors such as friction and efficiency, generating a reference current. More accurate

[0042] Meanwhile, the method of this invention enables the system to have self-learning and calibration capabilities: allowing the model to be continuously updated and optimized using daily AGV operation data. This allows the baseline to adapt to slow changes in system characteristics caused by vehicle aging and tire wear, improving the long-term stability and reliability of the detection system.

[0043] Step S3: Compare the real-time acquired motor current signal with the dynamic current baseline model, calculate the current deviation, and extract the collision features; the calculation of the current deviation is as follows: ,in For based on the input Using functional relationships The calculated reference current, This represents the current being collected in real time at the current moment.

[0044] The above process compares the current current with a dynamic baseline, yielding a clear deviation, quantifying the degree of anomaly, and transforming the collision detection problem into the identification of the deviation signal, making the subsequent analysis target clear and measurable. Furthermore, directly calculating the deviation effectively filters out large current changes caused by normal motion, allowing the analysis to focus on current components that cannot be explained by conventional motion, greatly improving the signal-to-noise ratio and making collision features easier to extract.

[0045] In step S3 above, the extraction of collision features includes at least one of the following features: A. Current sudden change amplitude threshold Under normal operating conditions Set it to 3 times the variance; B, current change rate. threshold Under normal operating conditions Set it to 3 times the variance.

[0046] The above process uses a multi-dimensional approach to capture collision features: the amplitude of the current change can directly reflect the intensity of the collision impact and is suitable for detecting instantaneous rigid collisions; while the rate of change of current can reflect the severity of load changes and can more sensitively detect collisions such as slight scratches and continuous compression that may cause the current to rise rapidly but whose amplitude may not be immediately large.

[0047] Step S4: Compare the collision features with a preset threshold to determine whether a collision event exists. In step S4, if one condition is met, or If a collision is detected, a safety mechanism such as emergency stop, alarm, or reporting will be triggered.

[0048] The above process uses "OR" logic (any feature exceeding the limit determines a collision), ensuring that both amplitude-type and rate-type anomalies can be quickly captured, reducing the risk of missed detection and improving the sensitivity and reliability of detection. At the same time, it clearly links the judgment result directly with safety mechanisms such as emergency stop, alarm, and reporting, realizing rapid and deterministic linkage from detection to safety control, and meeting the basic requirements of functional safety.

[0049] The AGV motors in the method of the present invention include a travel motor and a fork motor. According to the action instructions given by the AGV vehicle controller, the current of the corresponding motor is detected and it is determined whether a collision has occurred.

[0050] The walking motor is used to detect collisions between the AGV and its surrounding environment (people, objects, and other AGVs) during operation. The forklift motor is used to detect abnormal contact or obstruction with shelves, pallets, or goods during loading and unloading. By selecting and monitoring the appropriate motor according to controller instructions, this method can cover multiple collision risks throughout the entire AGV operation process, significantly improving the overall system safety level.

[0051] In the method of this invention, when the AGV uses dual drive motors, the current of each drive motor is detected separately to determine whether a collision has occurred. Independent detection and judgment of the current of the left and right drive motors can distinguish between a single-sided collision (e.g., brushing against an obstacle on the left, only the left motor current is abnormal) and a frontal collision (both sides have abnormal currents). This provides an information basis for more advanced safety strategies (such as fine-tuning obstacle avoidance in the opposite direction). Simultaneously, independent detection of the dual motors constitutes a degree of redundant judgment; when a sensor or signal chain on one side fails, the abnormal detection on the other side may still trigger safety protection, improving the system's fault tolerance and overall reliability.

[0052] The method of the present invention is implemented through a controller in an AGV. The controller specifically includes a signal acquisition layer, a core processing layer, and an execution output layer. The implementation of the method of the present invention also requires current model training, which can be run in the controller or on another server.

[0053] Signal Acquisition Layer: Defines the data source and interface, ensuring the real-time performance and reliability of the raw data input. Core Processing Layer: Concentrates core algorithms such as baseline calculation, deviation analysis, feature extraction, and threshold determination; it is the "brain" of the system, facilitating centralized algorithm maintenance and upgrades. Execution Output Layer: Ensures that safety commands are deterministically sent to the drivers and alarm units, achieving hardware-software linkage.

[0054] This layered architecture makes the system design clear and responsibilities well-defined, improving the maintainability, testability, and portability of the software. It also provides flexible model deployment options: running on the controller (local model deployment with extremely low response latency, no network dependency, and maximum real-time performance and reliability, suitable for collision-and-stop scenarios with stringent real-time requirements); or running on other servers (cloud or edge server deployment), facilitating centralized training, unified management, and continuous optimization of models from multiple AGVs, and allowing the use of more powerful servers to run more complex models, as well as facilitating data backhaul and big data analysis. This flexibility allows implementers to choose the optimal deployment strategy based on specific scenario requirements (real-time requirements or centralized management requirements, etc.) and hardware resources, greatly enhancing the applicability and scalability of the technical solution.

[0055] In summary, the principle of the method of this invention is that when an AGV collides, the drive wheel or steering mechanism experiences additional resistance, causing a sudden increase in motor load. This is reflected in the electrical characteristics as an abnormal increase or drastic fluctuation in motor current. Compared to installing external sensors, using the existing current feedback signal of the motor controller for collision detection has the potential advantages of no additional hardware cost, simple installation, and no detection blind spots (any collision affecting the drive load can be detected).

[0056] The method of the present invention is a highly reliable current-based collision detection method that can effectively distinguish between normal load changes and abnormal collision impacts. It establishes a benchmark model that can accurately predict the "normal" operating current of the motor under multiple variable conditions such as current speed, acceleration, steering angle and load, and sensitively and accurately captures abnormal features caused by collisions by comparing the deviation between the actual current and this dynamic baseline in real time.

[0057] The method of this invention has high real-time performance, with a detection delay of <10ms, which is better than mechanical switches; the method of this invention also has strong robustness: by training to obtain a dynamic reference current to eliminate operating condition interference, the false alarm rate is reduced by more than 90%; the method of this invention can also achieve full operating condition coverage: it can detect not only collisions of the vehicle body itself, but also collisions of goods on the forks.

[0058] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0059] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention without departing from the principles and spirit of the present invention. Any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. An AGV collision detection method based on motor current variation, characterized in that, Includes the following steps: S1. Real-time acquisition of current signals of each motor in the AGV and motion status parameters of the AGV; S2. A dynamic current baseline model is trained based on the collected current signal and the motion state parameters of the AGV. S3. Compare the real-time acquired motor current signal with the dynamic current baseline model, calculate the current deviation and extract the collision features. S4. Compare the collision features with the preset threshold and combine them to determine whether a collision event exists.

2. The AGV collision detection method based on motor current variation according to claim 1, characterized in that, In step S1, the current signals of each motor in the AGV are collected in real time. The sampling frequency is ≥ 10Hz; and the motion state parameters of the AGV are acquired synchronously, including speed v, acceleration a, steering angle θ and load w.

3. The AGV collision detection method based on motor current variation according to claim 1, characterized in that, In step S2, based on the current vehicle motion state and load, the theoretical current baseline value is calculated using a preset motor load model. Its form is: , By collecting speed, acceleration, steering angle, and load information at each moment. and the motor current at the corresponding moment Using regression analysis, the input was learned through the Scikit-learn library. and Train them to obtain the functional relationship between them. .

4. The AGV collision detection method based on motor current variation according to claim 3, characterized in that, In step S3, the current deviation is calculated as follows: ,in For based on the input Using functional relationships The calculated reference current, This represents the current being collected in real time at the current moment.

5. The AGV collision detection method based on motor current variation according to claim 4, characterized in that, In step S3, the extraction of collision features includes at least one of the following features: A. Current sudden change amplitude threshold Under normal operating conditions Set it to be 3 times the variance; B. Rate of change of current threshold Under normal operating conditions Set it to 3 times the variance.

6. The AGV collision detection method based on motor current variation according to claim 5, characterized in that, In step S4, when one condition is met... or If a collision is detected, a safety mechanism such as emergency stop, alarm, or reporting will be triggered.

7. The AGV collision detection method based on motor current variation according to claim 1, characterized in that, The motors in the AGV include a travel motor and a fork motor. Based on the action commands given by the AGV vehicle controller, the current of the corresponding motor is detected, and it is determined whether a collision has occurred.

8. The AGV collision detection method based on motor current variation according to claim 1, characterized in that, When an AGV uses dual drive motors, the current of each drive motor is detected separately to determine whether a collision has occurred.

Images