AGV robot two-way obstacle avoidance method and system for narrow storage channel

Abstract

The invention discloses an AGV robot two-way obstacle avoidance method and system for a narrow storage channel, and relates to the technical field of robot obstacle avoidance. The method comprises the following steps: when entering a narrow channel, acquiring opposite AGV data; conflict prediction based on opposite AGV data is executed through occupied space prediction, and if conflicts exist, a preset cooperation rule is adopted to dynamically determine a traffic priority sequence of the AGVs of all conflicting parties; according to the traffic priority sequence and the opposite AGV data, an obstacle avoidance planning module is triggered to make a traffic obstacle avoidance decision, and a bidirectional obstacle avoidance strategy is determined; and according to the bidirectional obstacle avoidance strategy, AGV robot bidirectional obstacle avoidance control of the narrow channel is carried out. The technical problem that in the prior art, the AGV robot is prone to collision during two-way passing, an effective cooperative obstacle avoidance mechanism is lacked, and the passing efficiency is low is solved, the passing priority is generated based on conflict prediction of the opposite AGV, the cooperative obstacle avoidance decision is implemented, and the technical effect of improving the passing efficiency of a narrow channel is achieved.

Description

Technical Field

[0001] This invention relates to the field of robot obstacle avoidance technology, specifically to a bidirectional obstacle avoidance method and system for AGV robots facing narrow warehouse passages. Background Technology

[0002] With the increasing level of warehouse automation, AGVs (Automated Guided Vehicles) are being used more and more widely in logistics and warehousing systems. However, in space-constrained warehousing environments, narrow passages typically only allow single-lane traffic. When AGVs from two directions enter the area simultaneously, traffic conflicts are highly likely to occur. Existing technologies mostly rely on single-vehicle obstacle avoidance strategies or simple yield rules, lacking a holistic and coordinated assessment of the status of vehicles in both directions. They cannot accurately predict the space occupied by oncoming AGVs and their future movement trends, resulting in a high degree of randomness in the obstacle avoidance process and a tendency for confrontations or blockages. Especially in warehousing scenarios with high-frequency traffic of multiple vehicles, when AGVs lack an effective priority coordination mechanism, conflicts can lead to passage congestion, reduced traffic efficiency, and even system outages. Summary of the Invention

[0003] This application provides a bidirectional obstacle avoidance method and system for AGV robots in narrow warehouse passages, which solves the technical problems of low passage efficiency caused by collisions and lack of effective collaborative obstacle avoidance mechanisms when AGV robots travel in both directions in the prior art.

[0004] The first aspect of this application provides a bidirectional obstacle avoidance method for AGV robots facing narrow warehouse aisles, the method comprising:

[0005] When entering a narrow passage, the AGV acquires data from oncoming AGVs via onboard sensors and wireless communication. It then predicts space occupancy and performs conflict prediction based on the oncoming AGV data. If a conflict exists, a preset coordination rule is used to dynamically determine the passage priority sequence of the conflicting AGVs. Based on the passage priority sequence and the oncoming AGV data, the obstacle avoidance planning module is triggered to make passage and obstacle avoidance decisions, determining a two-way obstacle avoidance strategy. The planning for higher-priority AGVs is based on their original planned path, while lower-priority AGVs are planned using the available space beside the passage for active avoidance. Based on this two-way obstacle avoidance strategy, the AGV robot performs two-way obstacle avoidance control in the narrow passage.

[0006] A second aspect of this application provides a bidirectional obstacle avoidance system for AGV robots in narrow warehouse aisles, the system comprising:

[0007] Data Acquisition Component: When entering a narrow passage, the AGV acquires data from oncoming AGVs via onboard sensors and wireless communication; Conflict Prediction Component: Based on space occupancy prediction, it performs conflict prediction based on oncoming AGV data. If a conflict exists, it dynamically determines the passage priority sequence of the AGVs on each side using preset coordination rules; Obstacle Avoidance Decision Component: Based on the passage priority sequence and oncoming AGV data, it triggers the obstacle avoidance planning module to make passage obstacle avoidance decisions and determine a two-way obstacle avoidance strategy. The planning of the higher priority AGV is based on the original planned path, while the planning of the lower priority AGV is based on the active avoidance of available space beside the passage; Obstacle Avoidance Control Component: Based on the two-way obstacle avoidance strategy, it performs two-way obstacle avoidance control of the AGV robot in the narrow passage.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] When entering a narrow passage, the AGV acquires data from oncoming AGVs via onboard sensors and wireless communication. Through space occupancy prediction, it performs conflict prediction based on the oncoming AGV data. If a conflict exists, it dynamically determines the passage priority sequence of the conflicting AGVs using preset collaborative rules. Based on the passage priority sequence and the oncoming AGV data, the obstacle avoidance planning module is triggered to make passage obstacle avoidance decisions and determine a two-way obstacle avoidance strategy. The planning of the higher-priority AGV is based on its original planned path, while the lower-priority AGV plans by actively avoiding obstacles using available space beside the passage. Based on the two-way obstacle avoidance strategy, the AGV robot performs two-way obstacle avoidance control in narrow passages. This solves the technical problems of low passage efficiency caused by frequent collisions and the lack of effective collaborative obstacle avoidance mechanisms in existing AGV robots traveling in both directions. By generating passage priorities based on oncoming AGV conflict prediction and implementing collaborative obstacle avoidance decisions, it achieves the technical effect of improving the passage efficiency in narrow passages. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic diagram of a bidirectional obstacle avoidance method for AGV robots in narrow warehouse aisles, provided in an embodiment of this application.

[0012] Figure 2 This is a schematic diagram of the bidirectional obstacle avoidance system for AGV robots in narrow warehouse passages, provided in an embodiment of this application.

[0013] Figure labeling: Data acquisition component 11, conflict prediction component 12, obstacle avoidance decision component 13, obstacle avoidance control component 14. Detailed Implementation

[0014] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0015] Example 1, as Figure 1 As shown, this application provides a bidirectional obstacle avoidance method for AGV robots in narrow warehouse aisles, wherein the method includes:

[0016] When entering a narrow passage, the AGV obtains data from the oncoming AGV through onboard sensors and wireless communication.

[0017] When an AGV approaches a narrow passage entrance within a preset trigger distance (e.g., 1-3 meters), it activates a passage cooperative perception mode. It utilizes its onboard forward sensors (LiDAR, depth camera, or ultrasonic rangefinder) and lateral sensors to continuously scan moving objects within the passage, extracting the position point cloud, contour boundary, and motion characteristic parameters of the opposing AGV. Simultaneously, it broadcasts its status request signal to the scheduling system or nearby AGVs via its onboard wireless communication module (Wi-Fi, UWB, or a dedicated warehouse scheduling network), and receives structured data proactively reported by the opposing AGV, including identifiers, real-time position coordinates, velocity vectors, task path points, and task urgency indicators. The system timestamps, fuses, and transforms the data obtained from the sensors and wireless communication to form opposing AGV data, providing fundamental data support for subsequent space occupancy prediction and conflict assessment.

[0018] Furthermore, the vehicle-mounted sensors include forward-facing sensors and lateral sensors; the wireless communication transmits data including the identifier of the opposing AGV, its real-time position, speed vector, current task path point, and task urgency indicator; by integrating the sensor data and the transmitted data, the opposing AGV data is determined.

[0019] Before entering a narrow passage, the AGV activates forward-facing sensors (including LiDAR, depth cameras, millimeter-wave radar, or ultrasonic ranging modules) and lateral sensors deployed at the front of the vehicle to continuously scan the geometry, boundaries, and dynamic changes within the passage, obtaining the spatial position point cloud, contour boundaries, and relative motion trends of the opposing AGV. Simultaneously, the system receives structured status data actively broadcast by the opposing AGV or forwarded by the scheduling system via a wireless communication module. This data includes at least the opposing AGV's unique identifier, real-time position coordinates based on a global coordinate system, velocity vector, path point sequence of the currently executing task, and task urgency indicator. The system synchronously processes the environmental perception data acquired by the sensors and the status data acquired by wireless communication, constructing a unified data model through timestamp alignment, coordinate transformation, error compensation, and data fusion algorithms. The final result is opposing AGV data containing motion parameters, path parameters, and task attributes.

[0020] Furthermore, by using kinematic prediction, the space occupied by the opposing AGV within a preset time window is determined and added to the opposing AGV data.

[0021] Based on the fused data of the opposing AGVs, the system invokes a kinematic prediction model to extrapolate the spatial trajectory of the opposing AGVs within a preset future time window. The kinematic prediction model comprehensively utilizes the current velocity vector, acceleration estimation, pathpoint pointing information, and task constraints to generate a position sequence of the opposing AGVs across consecutive time slices. Simultaneously, it combines the AGV's own dimensions, dynamic safety boundaries, and channel geometry to generate a corresponding occupied space region for each time slice. This occupied space region can be represented as an expanded rectangle, a polygonal coverage area, or a grid-based occupied segment. The system appends the predicted future occupied space set as an extension parameter to the opposing AGV data, ensuring that the opposing AGV data simultaneously includes the current state and the spatiotemporal regions that may be occupied in the future. This supports subsequent conflict determination, priority calculation, and obstacle avoidance planning.

[0022] By predicting the occupied space, conflict prediction based on the data of opposing AGVs is performed. If a conflict exists, a preset coordination rule is used to dynamically determine the passage priority sequence of the AGVs of the conflicting parties.

[0023] After acquiring data from opposing AGVs and predicting future space occupancy, the system constructs a spatiotemporal occupancy model for both AGVs within a preset time window. By comparing whether the occupied sections of the two vehicles overlap or intersect at the same channel node, it determines whether a traffic conflict will occur in the future. When an overlapping area or potential obstruction trend is detected in the occupied space of any time slice, the system triggers the conflict prediction module, including the AGVs entering the conflict chain within the scope of collaborative decision-making. Subsequently, preset collaborative rules are invoked to perform hierarchical weighted calculations on the task attribute weights (such as task urgency and cargo priority) and status weights (such as remaining battery power, waiting time, and distance to the conflict point) of the conflicting AGVs, obtaining a traffic priority weight for each AGV, and generating a traffic priority sequence according to the weight. This priority sequence serves as a constraint for the subsequent obstacle avoidance planning module, enabling coordination before traffic conflicts occur, ensuring that high-priority AGVs maintain their original routes, while low-priority AGVs actively avoid obstacles during planning, thereby achieving efficient two-way collaborative obstacle avoidance decision-making.

[0024] Furthermore, by employing preset coordination rules, the passage priority sequence of the AGVs of the conflicting parties is dynamically determined, including:

[0025] The system sets dynamic passage priority weights and defines the preset coordination rules, wherein the priority weights are defined based on task attributes and AGV status; the priority weights are determined by: weighting and summing the task urgency and the priority of the carried goods to determine the task attribute weights; weighting and summing the remaining battery power, the waiting time, and the distance from the current position to the conflict point to determine the AGV status weights; and weighting and summing the task attribute weights and the AGV status weights to determine the priority weights.

[0026] After identifying a potential traffic conflict within a narrow passage, the system first sets a priority weight for dynamic traffic coordination and defines preset coordination rules based on this. The priority weight consists of task attribute weights and AGV status weights, reflecting the importance and intensity of traffic demand for each conflicting AGV in the current scenario.

[0027] The methods for determining priority include:

[0028] First, based on the time requirements and task level of each AGV's current task, a weighted sum is calculated based on the task urgency and the priority of the goods carried, generating a task attribute weight to characterize the degree of impact of the task on passage priority. Second, based on the AGV's operating status information, a weighted sum is calculated based on the remaining battery power, waiting time, and distance from the current position to potential conflict points, obtaining the AGV status weight to reflect the equipment's sustainable operating capacity and the degree of passage obstruction. Finally, the task attribute weight and AGV status weight are merged according to a preset ratio, and a comprehensive priority weight for each AGV is obtained by weighting and summing again. The system sorts the priority weights from high to low, dynamically generating a passage priority sequence for conflicting AGVs, which serves as the core basis for subsequent obstacle avoidance path planning and passage coordination, realizing a controllable, transparent, and executable priority passage strategy for the AGVs of the conflicting parties.

[0029] Based on the passage priority sequence and the opposing AGV data, the obstacle avoidance planning module is triggered to make passage obstacle avoidance decisions and determine the two-way obstacle avoidance strategy. The planning of the higher priority AGV is based on the original planned path, while the planning of the lower priority AGV is based on the active avoidance of the available space on the side of the channel.

[0030] After generating a priority sequence for the AGVs involved in the conflict, the system calls the obstacle avoidance planning module to establish a bidirectional passage decision model constrained by priority. For the highest-priority AGV in the sequence, the obstacle avoidance planning module makes minor adjustments to some path nodes based on its original planned path to ensure that it can maintain a basically straight or slightly modified passage trajectory in narrow passages without causing significant deviation from the overall task path. For lower-priority AGVs, the obstacle avoidance planning module automatically generates an active avoidance strategy based on available space data beside the passage (including gap areas, lateral buffer zones, and local widening areas of the passage). Through methods such as lateral movement, deceleration, or short-term stopping, it reserves sufficient passage space for high-priority AGVs while ensuring a safe distance. Based on the prediction of the future occupied space of the opposing AGVs, the system dynamically adjusts the avoidance interval and avoidance magnitude to generate lateral or longitudinal obstacle avoidance paths for lower-priority AGVs. Ultimately, the obstacle avoidance planning module, driven by priority rules, performs multiple rounds of spatiotemporal consistency checks on the high-priority path-keeping strategy and the low-priority obstacle avoidance strategy, and outputs a bidirectional obstacle avoidance strategy that satisfies both bidirectional passage and safety constraints, thereby achieving collaborative and efficient bidirectional passage control of AGVs in narrow passages.

[0031] Furthermore, triggering the obstacle avoidance planning module to make passage and obstacle avoidance decisions includes:

[0032] The temporary obstacle avoidance methods for non-priority AGVs include deceleration, lateral movement, and stopping. The lateral movement avoidance area can be any one of a gap area, an intersection buffer zone, or a channel widening area. Based on the passage priority sequence and the temporary obstacle avoidance method, a progressive obstacle avoidance planning for AGVs with decreasing priority is executed to generate the bidirectional obstacle avoidance strategy.

[0033] When the system determines a two-way traffic conflict based on the traffic priority sequence, the obstacle avoidance planning module first selects an executable temporary obstacle avoidance method for non-priority AGVs to ensure the continuity of passage for high-priority AGVs. The temporary obstacle avoidance methods include at least three modes: deceleration, lateral movement, and stopping. The avoidance zone used for lateral movement is a pre-marked available area in the narrow passage structure, including any type such as a gap area, intersection buffer zone, or passage widening area, used to create a safe passage for high-priority AGVs within a limited space. Based on this, the obstacle avoidance planning module executes a progressive planning process from the highest priority AGV to the lowest priority AGVs according to the traffic priority sequence. The system uses the temporary obstacle avoidance method as a constraint benchmark and dynamically generates a corresponding avoidance strategy for each priority level: high-priority AGVs maintain their original path or make minor adjustments, while low-priority AGVs execute avoidance actions matching their priority, such as decelerating in advance, laterally entering the buffer zone, or making a short stop. The obstacle avoidance planning module ensures that the movement trajectories of each AGV are executable, conflict-free, and safe across multiple priority levels through multi-round progressive obstacle avoidance path generation and spatiotemporal consistency checks. Finally, the system integrates the planning results from each level and outputs a bidirectional obstacle avoidance strategy that satisfies bidirectional passage constraints, enabling controllable, collaborative obstacle avoidance and efficient passage of AGVs in narrow passages.

[0034] Furthermore, the progressive obstacle avoidance planning for AGVs with decreasing priority includes:

[0035] Based on the priority sequence, a first-priority AGV and a second-priority AGV are determined; fine-tuning of the first-priority AGV under the original planned path and obstacle avoidance planning of the second-priority AGV under the available space beside the channel are performed to determine the first spatiotemporal obstacle avoidance strategy; constraint of the available space beside the channel is applied to the first spatiotemporal obstacle avoidance strategy, and obstacle avoidance decisions are made for the third-priority AGV and the fourth-priority AGV. Multiple rounds of planning are performed with repeated steps to generate the bidirectional obstacle avoidance strategy.

[0036] After obtaining the passage priority sequence, the system first determines the first-priority AGV and the second-priority AGV in the current conflict scenario based on the priority sequence, and initiates the first round of obstacle avoidance planning. For the first-priority AGV, the obstacle avoidance planning module uses its original planned path as the core, making only minor adjustments to the path at necessary locations to maintain its primary right-of-way in narrow passages. For the second-priority AGV, the obstacle avoidance planning module generates an executable avoidance path based on the distribution of available space beside the passage (including gap areas, lateral buffer zones, or widening areas), releasing passage space for the first-priority AGV through methods such as lateral movement, deceleration, or short-term stops. Based on the planning results of the above two approaches, the system constructs a first spatiotemporal obstacle avoidance strategy to ensure that the passage trajectories of both AGVs do not conflict in time and space.

[0037] After the initial spatiotemporal obstacle avoidance strategy is determined, the system updates the remaining capacity and usable area of ​​the adjacent space, using this as a new planning constraint to execute the next round of obstacle avoidance decisions. In this round, the obstacle avoidance planning module selects the third-priority AGV and the fourth-priority AGV based on their priority sequence, planning according to the same obstacle avoidance logic as described above: the third-priority AGV prioritizes traffic stability with minimal path adjustments, while the fourth-priority AGV performs active obstacle avoidance actions based on the remaining available avoidance area. After each round of planning, the system performs a spatiotemporal consistency check on the generated local strategy and accumulates the results into the global obstacle avoidance strategy.

[0038] Through multiple rounds of progressive planning, the system resolves all potential conflicts layer by layer from high priority to low priority, ultimately forming a two-way obstacle avoidance strategy that meets the safety passage requirements of all AGVs and the constraints of the channel structure, achieving efficient and coordinated multi-AGV conflict resolution and obstacle avoidance scheduling in narrow channel scenarios.

[0039] Based on the described two-way obstacle avoidance strategy, two-way obstacle avoidance control is implemented for AGV robots in narrow passages.

[0040] After generating a bidirectional obstacle avoidance strategy that satisfies spatiotemporal consistency and priority constraints, the system issues the corresponding obstacle avoidance instruction set to the controlled AGVs, initiating the bidirectional passage control process within the narrow passage. For AGVs with higher passage priority, the system instructs them to execute a finely adjusted driving trajectory according to the planned path to maintain passage continuity. For low-priority AGVs, the system schedules them to perform avoidance actions such as lateral movement, deceleration, stopping, or re-entering the passage buffer zone according to the strategy requirements, ensuring that they complete safe avoidance within the specified time slot. At the same time, the system monitors the execution status of each AGV in real time through onboard sensors and compares the monitoring results with the planned trajectory, dynamically correcting trajectories with large deviations to maintain path tracking accuracy and safe distance control.

[0041] During obstacle avoidance control, the system continuously monitors the passage conditions within the channel, the relative distance between AGVs, and changes in occupied space. Based on real-time feedback, it makes necessary local updates to the bidirectional obstacle avoidance strategy, achieving dynamic obstacle avoidance and adaptive control. When a conflict is detected to be resolved or the channel is reopened, the system instructs the affected AGVs to gradually resume their original task paths and normal travel speeds, ensuring the entire obstacle avoidance process remains efficient, stable, and safe.

[0042] Furthermore, a flexible safety boundary is introduced to manage obstacle avoidance for AGVs passing through narrow passages. Specifically, a compact safety boundary is used for AGVs traveling towards each other to avoid obstacles; a standard safety boundary is used for AGVs traveling in opposite directions to avoid obstacles; and an expanded lateral safety boundary is used for AGVs traveling side by side and in staggered directions to avoid obstacles.

[0043] When executing bidirectional obstacle avoidance planning in narrow passages, the system constructs a dynamically adjustable flexible safety boundary model to adapt to the safety requirements and passage space conditions under different traffic scenarios. This flexible safety boundary is based on parameters such as AGV dimensions, travel speed, passage width, lateral available space, and the distribution of environmental obstacles. It generates corresponding safety distance ranges through real-time calculations, serving as constraints for obstacle avoidance planning and control execution.

[0044] In the scenario of AGVs traveling in opposite directions and avoiding each other, since both are in motion and the risk of axial collision in the passage is high, the system shrinks the lateral margin and forward spacing of the safety boundary to form a compact safety boundary, so as to free up more passage space for high-priority AGVs, while ensuring a safe distance between the two in short-term encounters.

[0045] In scenarios where AGVs avoid stationary targets (such as parked AGVs or temporary obstacles), the system adopts standard safety boundaries and plans based on normal driving safety distances to ensure that AGVs have a stable and controllable safety margin during avoidance or detour.

[0046] In scenarios where AGVs travel side-by-side or partially intersecting, due to the small lateral distance and the susceptibility to scratches or lateral interference, the system expands the lateral range of the safety boundary to form an expanded lateral safety boundary.

[0047] Through the dynamic switching and automatic adaptation of the aforementioned flexible safety boundaries, the system can achieve more refined and safer obstacle avoidance behavior control in different passage scenarios, improving the passage stability and obstacle avoidance robustness of AGVs in narrow passages.

[0048] In summary, the embodiments of this application have at least the following technical effects:

[0049] When entering a narrow passage, the AGV acquires data from oncoming AGVs via onboard sensors and wireless communication. Through space occupancy prediction, it performs conflict prediction based on the oncoming AGV data. If a conflict exists, it dynamically determines the passage priority sequence of the conflicting AGVs using preset collaborative rules. Based on the passage priority sequence and the oncoming AGV data, the obstacle avoidance planning module is triggered to make passage obstacle avoidance decisions and determine a two-way obstacle avoidance strategy. The planning of the higher-priority AGV is based on its original planned path, while the lower-priority AGV plans by actively avoiding obstacles using available space beside the passage. Based on the two-way obstacle avoidance strategy, the AGV robot performs two-way obstacle avoidance control in narrow passages. This solves the technical problems of low passage efficiency caused by frequent collisions and the lack of effective collaborative obstacle avoidance mechanisms in existing AGV robots traveling in both directions. By generating passage priorities based on oncoming AGV conflict prediction and implementing collaborative obstacle avoidance decisions, it achieves the technical effect of improving the passage efficiency in narrow passages.

[0050] Example 2, based on the same inventive concept as the bidirectional obstacle avoidance method for AGV robots facing narrow warehouse passages in the previous examples, such as... Figure 2 As shown, this application provides a bidirectional obstacle avoidance system for AGV robots in narrow warehouse aisles, wherein the system includes:

[0051] Data acquisition component 11: When entering a narrow passage, the AGV acquires data from the oncoming AGV via onboard sensors and wireless communication; Conflict prediction component 12: By predicting occupied space, it performs conflict prediction based on the oncoming AGV data. If a conflict exists, it uses preset coordination rules to dynamically determine the passage priority sequence of the AGVs on each side of the conflict; Obstacle avoidance decision component 13: Based on the passage priority sequence and the oncoming AGV data, it triggers the obstacle avoidance planning module to make passage obstacle avoidance decisions and determine a two-way obstacle avoidance strategy. The planning of the higher priority AGV is based on the original planned path, while the planning of the lower priority AGV is based on the active avoidance of the available space beside the passage; Obstacle avoidance control component 14: Based on the two-way obstacle avoidance strategy, it performs two-way obstacle avoidance control of the AGV robot in the narrow passage.

[0052] Furthermore, the data acquisition component 11 is used to perform the following methods:

[0053] The vehicle-mounted sensors include forward-facing and lateral-facing sensors; the wireless communication transmits data including the identifier of the opposing AGV, its real-time position, speed vector, current task path point, and task urgency indicator; by integrating the sensor data and the transmitted data, the opposing AGV data is determined.

[0054] Furthermore, the data acquisition component 11 is used to perform the following methods:

[0055] By using kinematic prediction, the space occupied by opposing AGVs within a preset time window is determined and added to the opposing AGV data.

[0056] Furthermore, the conflict prediction component 12 is used to perform the following method:

[0057] The system sets dynamic passage priority weights and defines the preset coordination rules, wherein the priority weights are defined based on task attributes and AGV status; the priority weights are determined by: weighting and summing the task urgency and the priority of the carried goods to determine the task attribute weights; weighting and summing the remaining battery power, the waiting time, and the distance from the current position to the conflict point to determine the AGV status weights; and weighting and summing the task attribute weights and the AGV status weights to determine the priority weights.

[0058] Furthermore, the obstacle avoidance decision component 13 is used to perform the following method:

[0059] The temporary obstacle avoidance methods for non-priority AGVs include deceleration, lateral movement, and stopping. The lateral movement avoidance area can be any one of a gap area, an intersection buffer zone, or a channel widening area. Based on the passage priority sequence and the temporary obstacle avoidance method, a progressive obstacle avoidance planning for AGVs with decreasing priority is executed to generate the bidirectional obstacle avoidance strategy.

[0060] Furthermore, the obstacle avoidance decision component 13 is used to perform the following method:

[0061] Based on the priority sequence, a first-priority AGV and a second-priority AGV are determined; fine-tuning of the first-priority AGV under the original planned path and obstacle avoidance planning of the second-priority AGV under the available space beside the channel are performed to determine the first spatiotemporal obstacle avoidance strategy; constraint of the available space beside the channel is applied to the first spatiotemporal obstacle avoidance strategy, and obstacle avoidance decisions are made for the third-priority AGV and the fourth-priority AGV. Multiple rounds of planning are performed with repeated steps to generate the bidirectional obstacle avoidance strategy.

[0062] Furthermore, the obstacle avoidance control component 14 is used to perform the following methods:

[0063] A flexible safety boundary is introduced to manage obstacle avoidance for AGVs passing through narrow passages. Specifically, a compact safety boundary is used for AGVs traveling towards each other to avoid obstacles; a standard safety boundary is used for AGVs traveling in opposite directions to avoid obstacles; and an expanded lateral safety boundary is used for AGVs traveling side by side and in staggered directions to avoid obstacles.

[0064] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for AGV robot bidirectional obstacle avoidance facing narrow aisle of warehouse, characterized in that, The method includes: When entering a narrow passage, the AGV obtains data from the oncoming AGV through onboard sensors and wireless communication. By predicting the occupied space, conflict prediction based on the data of the opposing AGVs is performed. If a conflict exists, a preset coordination rule is used to dynamically determine the passage priority sequence of the AGVs of the conflicting parties. Based on the passage priority sequence and the opposing AGV data, the obstacle avoidance planning module is triggered to make passage obstacle avoidance decisions and determine the bidirectional obstacle avoidance strategy. The planning of the higher priority AGV is based on the original planned path, while the planning of the lower priority AGV is based on the active avoidance of the available space on the side of the channel. Based on the described two-way obstacle avoidance strategy, two-way obstacle avoidance control is implemented for AGV robots in narrow passages.

2. The AGV robot bidirectional obstacle avoidance method for the narrow channel of the warehouse of claim 1, wherein, The vehicle-mounted sensors include forward-facing sensors and lateral sensors; The data transmitted via wireless communication includes the identifier of the opposing AGV, its real-time position, speed vector, current task path point, and task urgency indicator. By integrating sensor data and transmitted data, the data of the opposing AGV is determined. 3.The warehouse-oriented narrow aisle AGV robot bidirectional obstacle avoidance method of claim 2, wherein, By using kinematic prediction, the space occupied by opposing AGVs within a preset time window is determined and added to the opposing AGV data.

4. The AGV robot bidirectional obstacle avoidance method for the narrow channel of warehouse, according to claim 1, characterized in that, Using preset coordination rules, the passage priority sequence of AGVs from conflicting parties is dynamically determined, including: Set dynamic passage priority weights and define the preset coordination rules, wherein the priority weights are defined based on task attributes and AGV status; The methods for determining priority include: The weights of task attributes are determined by weighting and summing the urgency of the task and the priority of the cargo carried. The AGV status weight is determined by weighting and summing the remaining battery power, the waiting time, and the distance from the current position to the conflict point. Priority weights are determined by weighting and summing the task attribute weights and the AGV state weights.

5. The bidirectional obstacle avoidance method for AGV robots in narrow warehouse passages as described in claim 1, characterized in that, Trigger the obstacle avoidance planning module to make passage obstacle avoidance decisions, including: Temporary obstacle avoidance methods for non-priority AGVs include deceleration, lateral movement, and stopping. The lateral movement avoidance zone can be any one of the following: a gap zone, an intersection buffer zone, or a widened passage zone. Based on the passage priority sequence and taking the temporary obstacle avoidance method as a benchmark, the AGV progressive obstacle avoidance planning under the priority decrease is executed to generate the bidirectional obstacle avoidance strategy.

6. The bidirectional obstacle avoidance method for AGV robots in narrow warehouse passages as described in claim 5, characterized in that, Execution priority-decreasing AGV progressive obstacle avoidance planning includes: Based on the priority sequence, the first priority AGV and the second priority AGV are determined; Fine-tuning of the first priority AGV based on the original planned path, and obstacle avoidance planning of the second priority AGV based on the available space beside the passage, to determine the first spatiotemporal obstacle avoidance strategy. Using the first spatiotemporal obstacle avoidance strategy, constraints are imposed on the available space beside the passage. Obstacle avoidance decisions are made for the third-priority AGV and the fourth-priority AGV. Multiple rounds of planning are performed with repeated steps to generate the bidirectional obstacle avoidance strategy.

7. The bidirectional obstacle avoidance method for AGV robots in narrow warehouse passages as described in claim 1, characterized in that, Introduce flexible safety boundaries to manage obstacle avoidance when AGVs pass through narrow passages; In the case of AGVs traveling in opposite directions and avoiding each other, a compact safety boundary is adopted; For scenarios where AGVs need to avoid each other while their stationary and moving parts are in opposition, standard safety boundaries should be used. For scenarios where AGVs pass side-by-side and in a staggered manner, an extended lateral safety boundary is adopted.

8. A bidirectional obstacle avoidance system for AGV robots operating in narrow warehouse aisles, characterized in that: For implementing the bidirectional obstacle avoidance method for AGV robots facing narrow warehouse passages as described in any one of claims 1-7, the system comprises: Data acquisition component: When entering a narrow passage, the AGV acquires data from the oncoming AGV through onboard sensors and wireless communication; Conflict prediction component: By predicting the occupied space, it performs conflict prediction based on the data of the opposing AGVs. If a conflict exists, it uses preset coordination rules to dynamically determine the passage priority sequence of the AGVs of the conflicting parties. Obstacle avoidance decision component: Based on the passage priority sequence and the opposing AGV data, the obstacle avoidance planning module is triggered to make passage obstacle avoidance decisions and determine the two-way obstacle avoidance strategy. The planning of the higher priority AGV is based on the original planned path, while the lower priority AGV is planned by actively avoiding the available space on the side of the channel. Obstacle avoidance control component: Based on the bidirectional obstacle avoidance strategy, it performs bidirectional obstacle avoidance control for AGV robots in narrow passages.

Images