Four-way shuttle vehicle safety control system and method with autonomous recovery decision function

Abstract

The application provides a four-way shuttle vehicle safety control system and method with autonomous recovery decision function, and relates to the fields of industrial automation and logistics storage. Active sensing is used to realize safe execution of upper command, and a decision mechanism of derailment real-time detection and autonomous addressing recovery after failure is introduced, so as to improve the operation safety, reliability and availability of the system as a whole. The system comprises a controller, a four-way motion chassis and an execution mechanism, a multi-modal environment sensing module, a derailment detection special sensor module, an RFID positioning module and an autonomous recovery decision module for judging whether the shuttle vehicle can move to a specified RFID position for recovery according to derailment, environmental obstacles, traffic conditions and / or battery capacity. An accurate motion control and execution module is based on model prediction and controls the shuttle vehicle to realize trajectory tracking and state switching. An abnormality monitoring and autonomous response module is provided. The modules are connected through a vehicle-mounted bus to realize data interaction and collaborative control.

Description

Technical Field

[0001] This application relates to the fields of industrial automation and logistics warehousing, and specifically proposes a safety control system and method for four-way shuttle vehicles to achieve safe execution of upper-level commands, real-time derailment detection, and autonomous addressing and recovery after a fault by utilizing active perception. Background Technology

[0002] Currently, in automated production and automated warehouses in the e-commerce and express delivery industries, four-way shuttles are typically used to move rapidly along longitudinal and transverse aisles to accurately store or retrieve containers from designated warehouses. Therefore, ensuring accurate status monitoring and behavior verification of the four-way shuttles is crucial for guaranteeing system reliability and operational efficiency.

[0003] Existing technologies for four-way shuttles generally employ a combination of track guidance and fixed-program control. Because they primarily rely on physical contacts or QR codes on the track for position correction, they lack the ability to actively perceive the surrounding environment. In actual operation, unexpected situations such as foreign objects on the track, abnormal storage conditions, rack deformation, or pallet misalignment often lead to mission failures or even equipment damage. Furthermore, when the higher-level scheduling system issues route instructions, existing four-way shuttles merely execute them passively, unable to verify the feasibility of the route before execution, and struggling to provide autonomous and safe responses in the event of anomalies. In particular, over long-term operation, the track may wear out, deform, or the wheels may derail due to external forces; if these issues are not detected and addressed promptly, they can cause serious accidents.

[0004] Existing technologies address this by installing RFID chips in each grid cell of the track for position calibration. However, how to autonomously determine whether a shuttle can move to the nearest RFID location for initialization and recovery after a malfunction remains a critical technical problem. For example, if a shuttle has severely derailed or there are obstacles on the track ahead, forcibly moving it will lead to secondary accidents; conversely, if a vehicle is in good condition but lacks the ability to make autonomous judgments and awaits manual rescue, it will cause a prolonged system downtime.

[0005] Therefore, the applicant believes that the urgent task is to enable the four-way shuttle to autonomously decide whether to go to the RFID location for recovery based on the current operating status, which is crucial to improving the intelligence level and availability of the warehousing system.

[0006] In view of the above, this patent application is hereby filed. Summary of the Invention

[0007] This application proposes a four-way shuttle safety control system and method with autonomous recovery decision-making function. It aims to solve the problems existing in the prior art by using active perception to realize the safe execution of the upper-level command, and at the same time introduces a decision-making mechanism for real-time derailment detection and autonomous addressing and recovery after the fault, so as to improve the overall operational safety, reliability and availability of the system.

[0008] To achieve the aforementioned objectives, the four-way shuttle safety control system with autonomous recovery decision-making function, which is installed on the four-way shuttle to implement a safety control method, includes: The controller is connected and interacts with the multi-sensor fusion track recognition and positioning device and the walking motor that drives the four-way shuttle through signals; The four-way motion chassis and actuator include two sets of mutually perpendicular longitudinal wheels and transverse wheels. The lifting mechanism enables the longitudinal wheels and transverse wheels to switch their running states along tracks in different directions. A position sensor is installed on one side of the lifting mechanism. The multimodal environmental perception module includes a forward-facing LiDAR installed at the front of the vehicle body to scan the track area ahead and detect obstacles and track features; a rearward-facing LiDAR installed at the rear of the vehicle body for rearward environmental perception when reversing; front and rear binocular depth cameras installed at both ends of the vehicle body to identify cargo location status, pallet position, and label information; wheel encoders installed on the axles of the longitudinal and lateral wheels to measure travel distance and speed; and a six-axis inertial measurement unit (IMU) installed at the center of the vehicle body to measure vehicle attitude and acceleration. The sensor data from the above units are aligned through a time synchronization mechanism and input to the controller. The dedicated derailment detection sensor module uses multiple sensors combined with vehicle height status to detect and distinguish between normal reversal and abnormal derailment in real time, triggering a timely safety response. This module includes two sets of laser displacement sensors symmetrically mounted on both sides of each longitudinal and transverse wheel to measure the distance between the wheel flange and the side of the track, and a sensor mounted on the bottom of the vehicle body to measure the height of the vehicle body relative to the top surface of the track. h Point laser sensor; The RFID positioning module is an RFID reader / writer installed at the bottom of the vehicle body, with an RFID chip pre-embedded in each cargo location to record the unique code information of that cargo location. The autonomous recovery decision module is used to determine whether the vehicle can be moved to a designated RFID location for recovery based on whether the vehicle has derailed, environmental obstacles, traffic conditions, and / or remaining battery power. The precision motion control and execution module predicts and controls the shuttle to achieve trajectory tracking and state switching based on model prediction. The anomaly monitoring and autonomous response module monitors potential risks in real time and decides what safety measures the shuttle should take. The modules mentioned above are connected via an onboard bus to enable data interaction and collaborative control.

[0009] Based on the aforementioned four-way shuttle safety control system with autonomous recovery decision-making function, this application also proposes a four-way shuttle safety control method with autonomous recovery decision-making function, including the following stages: Stage (1), receiving and parsing instructions from the host computer; Receive order task instructions issued by the upper-level scheduling system; Phase (2), Pre-operation safety verification; Based on the multimodal environment perception module, the characteristics of the perception track and the cargo location are actively combined to verify the safety and feasibility of the shuttle vehicle's travel path. Stage (3), Path tracking and motion control; The system monitors the vehicle's status in real time during operation and autonomously selects and responds to malfunctions; this includes... (3.1) Conditions for triggering fault detection and recovery; The autonomous recovery process is triggered when the shuttle vehicle experiences one of the following conditions: Communication interruption with the upper control system exceeds the threshold time T comm = 5s; During task execution, recoverable anomalies were detected, including walking distance deviation exceeding the threshold Δp and motor overload; The upper control system periodically requires position calibration self-testing, such as triggering it every 8 hours; Received an initialization command issued remotely by a human; (3.2) Self-inspection and status assessment; The vehicle immediately stops and performs a self-inspection procedure, including derailment status detection, vehicle body integrity detection, and power status detection; if an abnormality is found, proceed to stage (3.7). (3.3) Environmental safety verification; This includes detecting and scanning the path taken to the nearest RFID tag location; Traffic condition detection: real-time acquisition of the operating status of other shuttles in the surrounding area. If there is an oncoming vehicle approaching in the direction of the route and there is no way to pass, a traffic conflict is determined and the execution phase (3.7) is initiated. (3.4) Autonomous movement decision-making; Based on the evaluation results of stage (3.3), a multi-dimensional comprehensive decision function is performed using the following judgment criteria. C Execution of move: C move=( δ derail≤1) and (normal posture) and ( Ecurr− E req> E min) and ( d obs≥ Ds (and (no traffic conflict)); like C If the move condition is True, then execute phase (3.5). If a temporary traffic conflict is caused by other shuttle vehicles or dynamic obstacles, then proceed to phase (3.6). Otherwise, during the execution phase (3.7), the shuttle will stop operating; (3.5) Autonomous path planning and movement; Determine the location of the nearest RFID tag; Based on the current vehicle location coordinates ( X curr ,Y curr), calculate the location distance to all RFID tags, and select the target point with the smallest distance ( X RFID Y RFID): Autonomous route planning; The A* algorithm is used to search for the shortest path in the track grid map and generate a sequence of path points, including reversal points and vehicle status switching instructions. Execute the move command; The shuttle operates at a low speed, such as when the walking speed decreases to... v recover=0.5m / s, travel along the planned path; during the journey, execute phase (3.2) in real time, and stop immediately if a new safety risk occurs; (3.6) Wait for retry; Entering a waiting state, continuously monitoring environmental changes, and setting a maximum waiting time. T wait,max If environmental conditions improve during the waiting period, proceed to stage (3.5); if they do not improve after the timeout, proceed to stage (3.7). (3.7) Moving or reporting is prohibited; The shuttle stops in place, reports the abnormal status and cause to the upper control system, and waits for manual intervention. (3.8) RFID reading and status initialization; Once the shuttle arrives at the nearest target RFID tag location, the following steps will be executed sequentially: The RFID reader reads the chip information and obtains the absolute coordinate code. zone , level , row , col), which is converted to global physical coordinates through a mapping function. X , Y , Z ); Reset the current row and column coordinates of the shuttle to RFID coordinate values; If recovery is triggered due to communication interruption, attempt to re-establish the connection with the upper control system; after successful connection, report the current location and status; if the connection still fails or the order task has been lost, wait for the upper control system to reschedule. Phase (4), picking up and placing goods and status feedback; Once the shuttle arrives at the target location, it will perform either a pickup or delivery operation depending on the task type.

[0010] The aforementioned stage (2) includes: using lidar to scan the track area covered by the path points ahead, detecting obstacles, and verifying the feasibility of the path; if the task type is picking up goods, using a depth camera to identify whether the target pallet and / or bin is in place and whether the pallet is neatly arranged, and verifying that the size of the storage location meets the vehicle's entry requirements; if the task type is delivery, using a depth camera to detect whether the target storage location is vacant; if all the above detection results are passed, proceed to the next stage; otherwise, the shuttle stops running and reports the abnormal situation.

[0011] In stage (3), the controller drives the four-way shuttle to travel according to the path point sequence, and solves the optimization problem at a preset time interval (e.g., 20ms as a cycle) using the following formula: Where X is the predicted state vector, X ref Here, M is the desired state vector, N is the control time domain, U is the prediction time domain, Q is the control input vector, and R is the state weight matrix (such as the optimization objective of position and velocity). Specifically, R is used to adjust the optimization objective of each control input. By solving and optimizing the following formula, a set of optimal control input sequences can be obtained. Under the premise of satisfying dynamic constraints and state constraints, the reference trajectory can be tracked and its energy consumption can be predicted to be within a reasonable range. Subsequently, by controlling the input sequence, starting from the current time k, the optimal control sequence is obtained by rolling the solution according to each consecutive control cycle, thus realizing rolling time-domain calculation.

[0012] The distance deviation calculation formula for stage (3) is as follows: Where Δp is the distance deviation value; X meas、 Y meas are the measured X-axis and Y-axis coordinate values, respectively. X ref、 Yref represents the expected reference coordinates for the X and Y axes, respectively. If Δp is greater than the preset threshold, such as Δp > 10mm, the shuttle is triggered to reposition itself.

[0013] In stage (3.2), the derailment detection involves real-time detection and calculation of the lateral offset and vertical height of the wheels in contact with the track. h The reference height is determined based on the current condition of the vehicle body. h The base is used to calculate the derailment coefficient; if the calculation result exceeds the preset threshold, it is determined that the vehicle is in a derailment state and an emergency stop is initiated. Calculate the lateral offset Δy: Δ y = K ⋅( dL - dR ) Where K is the calibration coefficient; Set the horizontal offset threshold Δy max as follows: in, W r For track width, Ww For wheel width, δ s For safety margin; Real-time measurement of the height of the vehicle body relative to the top surface of the track. h Based on the aforementioned reference height of the vehicle body, the reference height of each traveling wheel under different operating conditions is set. h base ; Calculate the derailment coefficient using the following formula. δ derail : in, T h This is the vertical threshold, which can be 5mm. like δ derail If the value is greater than 1, it is determined that the shuttle car has derailed, and the vehicle is stopped and the incident is reported.

[0014] In stage (3.3), the area in front of the track is scanned to identify other shuttles or objects encroaching on the track and predict their movement trends; if an obstacle is detected and the obstacle is within a certain distance... d obs < minimum safe distance D If s, then execute phase (3.7); Minimum safe distance Ds The calculation formula is as follows: in, v recover refers to the recovery speed. t react refers to reaction time. a max For maximum braking deceleration, d margin is the safety distance allowance.

[0015] In summary, this application has the following advantages and beneficial effects compared with the prior art: 1. By introducing an active sensing, derailment detection, and autonomous recovery decision-making mechanism, this application enables the four-way shuttle to autonomously verify path safety and monitor wheel status in real time during mission execution. After a fault, it can autonomously decide whether to move to the RFID location to resume operation based on factors such as the degree of derailment, environmental obstacles, and battery power, thereby significantly improving the system's operational safety, reliability, and availability.

[0016] 2. Existing shuttle vehicles passively receive and execute path instructions from the host system, lacking the ability to autonomously verify path feasibility before operation and failing to provide real-time safety responses to emergencies such as track obstructions, abnormal cargo location status, and pallet misalignment. This application addresses these shortcomings by integrating a multimodal sensing module. This module enables autonomous verification of path feasibility before task execution and real-time detection of risks such as track obstructions, abnormal cargo location status, and pallet misalignment during execution. This effectively mitigates task failures or equipment damage caused by environmental changes, significantly improving system reliability.

[0017] 3. Existing technologies lack real-time monitoring methods for the contact state between wheels and rails. When wheel derailment occurs due to rail wear, deformation, or external forces, it is impossible to detect and take timely braking measures, which can easily lead to serious accidents. This application employs a derailment detection method that integrates laser displacement sensors with vehicle height status. This method can accurately distinguish between normal reversing and abnormal derailment, triggering emergency braking at the moment of derailment or even before it occurs, thus preventing serious accidents caused by derailment and ensuring the safety of equipment and personnel.

[0018] 4. Existing technologies, when shuttle vehicles experience communication interruptions or exceed positioning deviation limits, cannot autonomously determine whether they can move to the nearest RFID location for initialization and recovery based on their own status and environmental information. This results in the need for manual intervention even for minor faults, leading to prolonged system downtime. This application designs a vehicle autonomous recovery decision-making function. When a communication interruption, a recoverability anomaly, or an initialization command is received, the vehicle can autonomously determine whether it can move to the nearest RFID location for status initialization based on multi-dimensional information such as the severity of derailment, vehicle integrity, remaining battery power, path obstacles, and traffic conditions. This mechanism effectively reduces the need for manual intervention due to minor faults, avoids prolonged system downtime, and significantly improves the availability and intelligence level of the warehousing system.

[0019] 5. This application utilizes RFID chips pre-installed within the track grid to provide an absolute position reference. Combined with multi-sensor fusion positioning, it can effectively eliminate the cumulative errors generated during long-term operation, ensuring high-precision stopping and reliable operation of the shuttle in dense storage environments.

[0020] 6. This application mainly relies on the on-board perception and decision-making module to achieve the above functions, without the need for large-scale modification of the existing rail system or the addition of a large number of ground beacons, which facilitates the intelligent upgrading and promotion of existing warehouses. Attached Figure Description

[0021] Figure 1 This is a flowchart of the four-way shuttle safety control method with autonomous recovery decision-making function described in this application; Figure 2 Flowchart for autonomous recovery decision-making; Detailed Implementation

[0022] Example 1: This application proposes a four-way shuttle safety control system and method with autonomous recovery decision-making function to control the four-way shuttle to travel in a grid-like aisle composed of an array of straight and horizontal tracks. Shelves are set between the matrix aisles, and each shelf has several storage locations arranged in a matrix. The four-way shuttle receives and executes the order tasks issued by the upper system, and travels to the designated storage location according to the path instructions to store or retrieve the material box there.

[0023] A control system for implementing safety control methods is installed on the four-way shuttle, and the system includes: The controller is connected and interacts with the multi-sensor fusion track recognition and positioning device and the walking motor that drives the four-way shuttle through signals. In order to meet the dynamics and state constraints, the controller outputs state commands including the walking motor speed and the lifting motor position, so as to realize the precise switching of the vehicle height state at the reversing point according to the command. The four-way motion chassis and actuator include two sets of mutually perpendicular longitudinal and lateral wheels. A lifting mechanism switches the longitudinal and lateral wheels between different directional tracks. A position sensor (preferably a magnetostrictive displacement sensor in this embodiment) is installed on one side of the lifting mechanism to provide real-time feedback on the current lifting height (Hlift) to determine the vehicle's status. For example, when the shuttle is in a low position, the lateral wheels are in contact with the lateral track, while the longitudinal wheels are suspended. When the shuttle is in a mid-position, the longitudinal wheels are in contact with the longitudinal track, while the lateral wheels are suspended. The pallet is in a low position and not lifted, allowing the shuttle to travel longitudinally and access the area below the cargo space. When the shuttle is in a high position, the longitudinal wheels are in contact with the longitudinal track, while the lateral wheels are suspended. The pallet is in a high-lifted state for carrying cargo, but the vehicle cannot access the area below the cargo space.

[0024] The multimodal environmental perception module includes a forward-facing LiDAR installed at the front of the vehicle body to scan the track area ahead and detect obstacles and track features; a rearward-facing LiDAR installed at the rear of the vehicle body for rearward environmental perception when reversing; front and rear binocular depth cameras installed at both ends of the vehicle body to identify cargo location status, pallet position, and label information; wheel encoders installed on the axles of the longitudinal and lateral wheels to measure travel distance and speed; and a six-axis inertial measurement unit (IMU) installed at the center of the vehicle body to measure vehicle attitude and acceleration. The sensor data from the above units are aligned through a time synchronization mechanism and input to the controller. The dedicated derailment detection sensor module uses multiple sensors combined with vehicle height status to detect and distinguish between normal reversal and abnormal derailment in real time, triggering a timely safety response. This module includes two sets of laser displacement sensors symmetrically mounted on both sides of each longitudinal and transverse wheel to measure the distance between the wheel flange and the side of the track, and a sensor mounted on the bottom of the vehicle body to measure the height of the vehicle body relative to the top surface of the track. h Point laser sensor; An RFID positioning module, consisting of an RFID reader / writer installed at the bottom of the vehicle body, has an RFID chip pre-embedded at each cargo location, recording the unique code information for that location; the preferred format of the code information is (…). zone , level , row , col These correspond to the area, floor, row number, and column number, respectively. When the shuttle vehicle travels to a certain storage location area, the RFID reader reads the encoded information of the RFID chip and converts it into global physical coordinates using the following coordinate mapping function: ( X , Y , Z )= Map ( zone , level , row ,col ) The aforementioned mapping function is pre-calibrated based on warehouse design parameters, such as the origin coordinates of the shelving area and the track spacing, and stored in the vehicle-mounted map database. When an RFID reader event is triggered, the system resets the current positioning coordinates to the RFID coordinate values ​​to eliminate accumulated errors.

[0025] The autonomous recovery decision module is used to determine whether the vehicle can move to a designated RFID location to recover based on whether it has derailed, environmental obstacles, traffic conditions and / or battery remaining capacity. For example, when a communication interruption occurs, a recoverability error occurs or an initialization command is received, the controller uses this module to self-check the derailment status, vehicle integrity and battery remaining capacity, and scans for path obstacles and traffic conflicts to comprehensively determine whether it can autonomously move to the nearest RFID location to reset coordinates and thus resume operation. The precision motion control and execution module predicts and controls the shuttle to achieve trajectory tracking and state switching based on model prediction. The anomaly monitoring and autonomous response module monitors potential risks in real time and decides what safety measures the shuttle should take. The modules mentioned above are connected via an onboard bus (preferably a CAN bus or an industrial Ethernet in this embodiment) to achieve data interaction and collaborative control.

[0026] Based on the above control system, the four-way shuttle safety control method with autonomous recovery decision-making function proposed in this application includes the following stages: Stage (1), receiving and parsing instructions from the host computer; Receive order task instructions from the upper-level scheduling system, including task type (pickup / delivery), target pallet / bin ID, target location row and column coordinates, and path point sequence (composed of preset nodes in a grid-like aisle, each node containing instructions for reversing or picking up / placing goods). Phase (2), Pre-operation safety verification; Based on a multimodal environmental perception module, the system proactively combines the characteristics of the sensing track and the cargo location to verify the safety and feasibility of the shuttle vehicle's travel path; this includes... The track area covered by the path points ahead is scanned using lidar to detect obstacles and verify the feasibility of the path; If the task type is pickup, use a depth camera to identify whether the target pallet and / or bin are in place and whether the pallet is neatly arranged, and verify that the size of the storage location meets the vehicle entry requirements; If the task type is delivery, use a depth camera to detect whether the target storage location is available. If all the above test results are satisfactory, proceed to the next stage; otherwise, the shuttle bus will be stopped and the abnormality will be reported. Stage (3), Path tracking and motion control; The controller drives a four-way shuttle to travel according to a sequence of waypoints, and the optimization problem is solved using the following formula at preset time intervals of 20ms:

[0027] Where X is the predicted state vector, X ref Here, M is the desired state vector, N is the control time domain, U is the prediction time domain, Q is the control input vector, and R is the state weight matrix (such as the optimization objective of position and velocity). Specifically, R is used to adjust the optimization objective of each control input. By solving and optimizing the following formula, a set of optimal control input sequences can be obtained. Under the premise of satisfying dynamic constraints and state constraints, the reference trajectory can be tracked and its energy consumption can be predicted to be within a reasonable range. Subsequently, by controlling the input sequence, starting from the current time k, the optimal control sequence is obtained by rolling the solution according to each consecutive control cycle, thus realizing rolling time-domain calculation.

[0028] The real-time detection and calculation formula for the distance deviation of the shuttle car traveling on the track is as follows: Wherein, Δp is the distance deviation value; X meas、 Y meas are the measured X-axis and Y-axis coordinate values, respectively. X ref、 Y ref represents the expected reference coordinates for the X and Y axes, respectively. If Δp is greater than the preset threshold, such as Δp > 10mm, the shuttle car is triggered to reposition itself. The system monitors the vehicle's status in real time during operation and autonomously selects appropriate responses when malfunctions occur; this includes... (3.1) Conditions for triggering fault detection and recovery; The autonomous recovery process is triggered when the shuttle vehicle experiences one of the following conditions: Communication interruption with the upper control system exceeds the threshold time T comm = 5s; During task execution, recoverable anomalies were detected, including walking distance deviation exceeding the threshold Δp and motor overload; The upper control system periodically requires position calibration self-testing, such as triggering it every 8 hours; Received an initialization command issued remotely by a human; (3.2) Self-inspection and status assessment; The vehicle immediately stops and performs a self-check procedure, including: Derailment detection; Real-time detection and calculation of the lateral offset and vertical height of the traveling wheels in contact with the track. h The reference height is determined based on the current condition of the vehicle body. h The base is used to calculate the derailment coefficient; if the calculation result exceeds the preset threshold, it is determined that the vehicle is in a derailment state and an emergency stop is initiated. Taking the straight-moving wheel as an example, the distance measured by the laser sensor on the left is dL The right side is dR When driving normally, the straight-going wheels are centered. dL ≈ dR ≈ d 0 ( d 0 represents the standard gap), defining the lateral offset Δ. y The lateral offset Δy is calculated using the following formula, representing the distance between the wheel center and the track center: Δ y = K ⋅( dL - dR ) Where K is the calibration coefficient; Set the horizontal offset threshold Δy max as follows: in, W r For track width, Ww For wheel width, δ s For safety margin; for example, track width Wr =100mm, wheel width Ww =80mm, safety margin δ s =2mm, Δ is calculated. y max=8mm; Real-time measurement of the height of the vehicle body relative to the top surface of the track. h The reference heights for different operating states of the shuttle are set as follows: When the shuttle car body is in a low position, the lateral wheels contact the lateral rails, and the reference height of the car body is... h base,lat; When the shuttle car body is in the center position, the longitudinal wheels are in contact with the longitudinal rails, the pallet is in a low position and not raised, and the reference height of the car body is [missing information]. h base,lon,mid; When the shuttle car body is in a high position, the longitudinal wheels contact the longitudinal rails, and the pallet is in a raised position for loading and unloading. The base height of the car body is... h base,lon,high= h base,lon,mid+Δ H, where Δ H This is the increment of the lifting height; the aforementioned reference height value is calibrated statically in the vehicle and stored in the controller; Based on the aforementioned reference height of the vehicle body, the reference height of each traveling wheel is set for different operating conditions. h base ; Calculate the derailment coefficient using the following formula. δ derail : in, T h This is the vertical threshold, which can be 5mm. like δ derail If the value is >1, it is determined that the shuttle car has derailed, and the vehicle is stopped and the incident is reported. Vehicle integrity inspection; The vehicle tilt angle is detected in real time via a six-axis inertial measurement unit (IMU). When the pitch angle or roll angle exceeds the preset threshold, such as 3°, the vehicle body attitude will change abnormally (e.g., the IMU detects the vehicle body shaking or increased tilting). However, the value of the lifting shaft encoder of the lifting mechanism remains unchanged for a period of time. At this time, the control system has determined that the vehicle body has tilted and uses the position sensor set in the lifting mechanism to provide feedback and determine whether the lifting mechanism is stuck. If an exception occurs, proceed to phase (3.7). Power status detection; Get current battery level Ecurr r, calculates the energy required to move the RFID tag to the nearest location. Ereq, Based on distance and energy consumption per unit distance, energy consumption can be estimated according to the predicted state vector of the path tracking and motion control stage. That is, the electricity consumed in the previous cycle and the actual energy consumed have been fully recorded, and the energy consumed in the subsequent stage can be predicted based on this. like E curr− E req≤ E min (safety margin) E If the power consumption is less than 10% of the rated power, then the power is insufficient, and stage (3.7) is executed; otherwise, stage (3.6) is executed. (3.3) Environmental safety verification; This includes detecting and scanning the path taken to the nearest RFID tag location; Specifically, the system scans the area in front of the track, identifies other shuttles or objects encroaching on the track, and predicts their movement trends; if an obstacle is detected and the obstacle is within a certain distance... dobs < minimum safe distance D If s, then execute phase (3.7); Minimum safe distance Ds The calculation formula is as follows: in, v recover refers to the recovery speed. t react refers to reaction time. a max For maximum braking deceleration, d margin is the safety distance allowance; For example, take the recovery speed. v recover=0.5m / s, reaction time t react=0.3s, maximum braking deceleration a max =1.0 (m / s) 2 safety distance margin d With margin = 0.1m, the calculated value is... Ds ≈0.35m; if d obs < Ds If the path is blocked, the operation should be stopped. Traffic condition detection: real-time acquisition of the operating status of other shuttles in the surrounding area. If there is an oncoming vehicle approaching in the direction of the route and there is no way to pass, a traffic conflict is determined and the execution phase (3.7) is initiated. (3.4) Autonomous movement decision-making; Based on the evaluation results of stage (3.3), a multi-dimensional comprehensive decision function is performed using the following judgment criteria. C Execution of move: C move=( δ derail≤1) and (normal posture) and ( E curr− E req> E min) and ( d obs≥ Ds (and (no traffic conflict)); like C If the move condition is True, then execute phase (3.5). If a temporary traffic conflict is caused by other shuttle vehicles or dynamic obstacles, then proceed to phase (3.6). Otherwise, during the execution phase (3.7), the shuttle will stop operating; (3.5) Autonomous path planning and movement; Determine the location of the nearest RFID tag; Based on the current vehicle location coordinates ( X curr ,Y curr), calculate the location distance to all RFID tags, and select the target point with the smallest distance ( X RFID Y RFID): Autonomous route planning; The A* algorithm is used to search for the shortest path in the track grid map and generate a sequence of path points, including reversal points and vehicle status switching instructions. Execute the move command; The shuttle operates at a low speed, such as when the walking speed decreases to... v recover=0.5m / s, travel along the planned path; during the journey, execute phase (3.2) in real time, and stop immediately if a new safety risk occurs; (3.6) Wait for retry; Entering a waiting state, continuously monitoring environmental changes, and setting a maximum waiting time. T wait,max =30s; If environmental conditions improve during the waiting period, proceed to stage (3.5); if they do not improve after the timeout, proceed to stage (3.7). (3.7) Moving or reporting is prohibited; The shuttle stops in place, reports the abnormal status and cause to the upper control system, and waits for manual intervention. (3.8) RFID reading and status initialization; Once the shuttle arrives at the nearest target RFID tag location, the following steps will be executed sequentially: The RFID reader reads the chip information and obtains the absolute coordinate code. zone , level , row , col ), which is converted to global physical coordinates through a mapping function. X , Y , Z ); Reset the current row and column coordinates of the shuttle to the RFID coordinate values, record the initialization timestamp, and update the status to "initialized"; If recovery is triggered due to communication interruption, attempt to re-establish the connection with the upper control system; after successful connection, report the current location and status; if the connection still fails or the order task has been lost, wait for the upper control system to reschedule. Phase (4), picking up and placing goods and status feedback; Once the shuttle arrives at the target location, it will perform either a pickup or delivery operation depending on the task type. If the task type is pickup, the vehicle body enters the area below the storage location while maintaining the center position; after confirming the presence of the pallet, the pallet is raised to the high position; after the goods are retrieved, the pallet is lowered to the low position, and the shuttle vehicle exits the storage location. If the task type is delivery, the vehicle body remains in a high position and enters the space above the storage location while loaded; after confirming that the storage location is vacant, the pallet is lowered to the middle position to place the pallet, and the shuttle car exits the storage location.

[0029] As described above, the embodiments given in conjunction with the accompanying drawings are merely preferred solutions for achieving the objectives of this invention. Those skilled in the art can draw inspiration from this and directly derive other alternative structures that conform to the design concept of this invention. Other structural features derived therefrom should also fall within the scope of the solutions described in this invention.

Claims

1. A four-way shuttle safety control system with autonomous recovery decision-making function, characterized in that: The system is installed on a four-way shuttle to implement safety control methods. The system includes, The controller is connected and interacts with the multi-sensor fusion track recognition and positioning device and the walking motor that drives the four-way shuttle through signals; The four-way motion chassis and actuator include two sets of mutually perpendicular longitudinal wheels and transverse wheels. The lifting mechanism enables the longitudinal wheels and transverse wheels to switch their running states along tracks in different directions. A position sensor is installed on one side of the lifting mechanism. The multimodal environmental perception module includes a forward-facing LiDAR installed at the front of the vehicle body to scan the track area ahead and detect obstacles and track features; a rearward-facing LiDAR installed at the rear of the vehicle body for rearward environmental perception when reversing; front and rear binocular depth cameras installed at both ends of the vehicle body to identify cargo location status, pallet position, and label information; wheel encoders installed on the axles of the longitudinal and lateral wheels to measure travel distance and speed; and a six-axis inertial measurement unit (IMU) installed at the center of the vehicle body to measure vehicle attitude and acceleration. The sensor data from the above units are aligned through a time synchronization mechanism and input to the controller. The dedicated derailment detection sensor module uses multiple sensors combined with vehicle height status to detect and distinguish between normal reversal and abnormal derailment in real time, triggering a timely safety response. This module includes two sets of laser displacement sensors symmetrically mounted on both sides of each longitudinal and transverse wheel to measure the distance between the wheel flange and the side of the track, and a sensor mounted on the bottom of the vehicle body to measure the height of the vehicle body relative to the top surface of the track. h Point laser sensor; The RFID positioning module is an RFID reader / writer installed at the bottom of the vehicle body, with an RFID chip pre-embedded in each cargo location to record the unique code information of that cargo location. The autonomous recovery decision module is used to determine whether the vehicle can be moved to a designated RFID location for recovery based on whether the vehicle has derailed, environmental obstacles, traffic conditions, and / or remaining battery power. The precision motion control and execution module predicts and controls the shuttle to achieve trajectory tracking and state switching based on model prediction. The anomaly monitoring and autonomous response module monitors potential risks in real time and decides what safety measures the shuttle should take. The modules mentioned above are connected via an onboard bus to enable data interaction and collaborative control.

2. A four-way shuttle safety control method with autonomous recovery decision-making function, applying the four-way shuttle safety control system with autonomous recovery decision-making function as described in claim 1, characterized in that: Includes the following stages, Stage (1), receiving and parsing instructions from the host computer; Receive order task instructions issued by the upper-level scheduling system; Phase (2), Pre-operation safety verification; Based on the multimodal environment perception module, the characteristics of the perception track and the cargo location are actively combined to verify the safety and feasibility of the shuttle vehicle's travel path. Stage (3), Path tracking and motion control; The system monitors the vehicle's status in real time during operation and autonomously selects and responds to malfunctions; this includes... (3.1) Conditions for triggering fault detection and recovery; The autonomous recovery process is triggered when the shuttle vehicle experiences one of the following conditions: Communication interruption with the upper control system exceeds the threshold time T comm = 5s; During task execution, recoverable anomalies were detected, including walking distance deviation exceeding the threshold Δp and motor overload; The upper control system periodically requires position calibration self-testing, such as triggering it every 8 hours; Received an initialization command issued remotely by a human; (3.2) Self-inspection and status assessment; The vehicle immediately stops and performs a self-inspection procedure, including derailment status detection, vehicle body integrity detection, and power status detection; if an abnormality is found, proceed to stage (3.7). (3.3) Environmental safety verification; This includes detecting and scanning the path taken to the nearest RFID tag location; Traffic condition detection: real-time acquisition of the operating status of other shuttles in the surrounding area. If there is an oncoming vehicle approaching in the direction of the route and there is no way to pass, a traffic conflict is determined and the execution phase (3.7) is initiated. (3.4) Autonomous movement decision-making; Based on the evaluation results of stage (3.3), a multi-dimensional comprehensive decision function is performed using the following judgment criteria. C Execution of move: C move=( δ derail≤1) and (normal posture) and ( E curr− E req> E min) and ( d obs≥ Ds (and) (no traffic conflict); like C If the move condition is True, then execute phase (3.5). If a temporary traffic conflict is caused by other shuttle vehicles or dynamic obstacles, then proceed to phase (3.6). Otherwise, during the execution phase (3.7), the shuttle will stop operating; (3.5) Autonomous path planning and movement; Determine the location of the nearest RFID tag; Based on the current vehicle location coordinates ( X curr ,Y curr), calculate the location distance to all RFID tags, and select the target point with the smallest distance ( X RFID Y RFID): Autonomous route planning; The A* algorithm is used to search for the shortest path in the track grid map and generate a sequence of path points, including reversal points and vehicle status switching instructions. Execute the move command; The shuttle operates at a low speed, such as when the walking speed decreases to... v recover=0.5m / s, travel along the planned path; during the journey, execute phase (3.2) in real time, and stop immediately if a new safety risk occurs; (3.6) Wait for retry; Entering a waiting state, continuously monitoring environmental changes, and setting a maximum waiting time. T wait,max If environmental conditions improve during the waiting period, proceed to stage (3.5); if they do not improve after the timeout, proceed to stage (3.7). (3.7) Moving or reporting is prohibited; The shuttle stops in place, reports the abnormal status and cause to the upper control system, and waits for manual intervention. (3.8) RFID reading and status initialization; Once the shuttle arrives at the nearest target RFID tag location, the following steps will be executed sequentially: The RFID reader reads the chip information and obtains the absolute coordinate code. zone , level , row , col ), which is converted to global physical coordinates through a mapping function. X , Y , Z ); Reset the current row and column coordinates of the shuttle to RFID coordinate values; If recovery is triggered due to communication interruption, attempt to re-establish the connection with the upper control system; after successful connection, report the current location and status; if the connection still fails or the order task has been lost, wait for the upper control system to reschedule. Phase (4), picking up and placing goods and status feedback; Once the shuttle arrives at the target location, it will perform either a pickup or delivery operation depending on the task type.

3. The four-way shuttle safety control method with autonomous recovery decision-making function according to claim 2, characterized in that: The stage (2) includes using lidar to scan the track area covered by the path points ahead, detecting obstacles, and verifying the feasibility of the path; If the task type is pickup, use a depth camera to identify whether the target pallet and / or bin are in place and whether the pallet is neatly arranged, and verify that the size of the storage location meets the vehicle entry requirements; If the task type is delivery, use a depth camera to detect whether the target storage location is available. If all the above test results are satisfactory, proceed to the next stage; otherwise, the shuttle will stop operating and an abnormality will be reported.

4. The four-way shuttle safety control method with autonomous recovery decision-making function according to claim 2, characterized in that: In stage (3), the controller drives the four-way shuttle to travel according to the path point sequence, and solves the optimization problem at preset time intervals using the following formula: Where X is the predicted state vector, X ref is the desired state vector, M is the control time domain, N is the prediction time domain, U is the control input vector, Q is the state weight matrix, and R is the control weight matrix; By solving and optimizing the following formula, a set of optimal control input sequences can be obtained. Under the premise of satisfying dynamic constraints and state constraints, the reference trajectory can be tracked and its energy consumption can be predicted to be within a reasonable range. Subsequently, by controlling the input sequence, starting from the current time k, the optimal control sequence is obtained by rolling the solution according to each consecutive control cycle, thus realizing rolling time-domain calculation.

5. The four-way shuttle safety control method with autonomous recovery decision-making function according to claim 2, characterized in that: The distance deviation calculation formula for stage (3) is as follows: Where Δp is the distance deviation value; X meas、 Y meas are the measured X-axis and Y-axis coordinate values, respectively. X ref、 Y ref represents the expected reference coordinates for the X and Y axes, respectively. If Δp is greater than the preset threshold, such as Δp > 10mm, the shuttle is triggered to reposition itself.

6. The four-way shuttle safety control method with autonomous recovery decision-making function according to claim 2, characterized in that: In stage (3.2), the derailment detection involves real-time detection and calculation of the lateral offset and vertical height of the wheels in contact with the track. h The reference height is determined based on the current condition of the vehicle body. h The base is used to calculate the derailment coefficient; if the calculation result exceeds the preset threshold, it is determined that the vehicle is in a derailment state and an emergency stop is initiated. Calculate the lateral offset Δy: D y = K ⋅( dL − dR ) Where K is the calibration coefficient; Set the horizontal offset threshold Δy max as follows: in, W r For track width, Ww For wheel width, δ s For safety margin; Real-time measurement of the height of the vehicle body relative to the top surface of the track. h Based on the aforementioned reference height of the vehicle body, the reference height of each traveling wheel under different operating conditions is set. h base ; Calculate the derailment coefficient using the following formula. δ derail : in, T h This is the vertical threshold, which can be 5mm. like δ derail If the value is greater than 1, it is determined that the shuttle car has derailed, and the vehicle is stopped and the incident is reported.

7. The four-way shuttle safety control method with autonomous recovery decision-making function according to claim 2, characterized in that: In the aforementioned stage (3.3), the area in front of the track is scanned to identify other shuttles or objects that have intruded into the track and to predict their movement trends; If an obstacle is detected and the obstacle is within a certain distance... d obs < minimum safe distance D If s, then execute phase (3.7); Minimum safe distance Ds The calculation formula is as follows: in, v recover refers to the recovery speed. t react refers to reaction time. a max For maximum braking deceleration, d margin is the safety distance allowance.

Images