Conflict avoidance method and system for multiple grasping targets collaborative work

Abstract

The application provides a conflict avoidance method and system for multi-grabbing target cooperative work, which obtains task information of a current grabbing device, generates an initial prediction trajectory and a state vector, and encapsulates and broadcasts the state beacon, while the current grabbing device extends the initial prediction trajectory of other grabbing devices received to obtain other updated trajectories; the current updated trajectory and the other updated trajectories are mapped to the same space model to determine a conflict grabbing device; then the weight values of the current grabbing device and the conflict grabbing device are called, and the avoidance grabbing device is determined based on the weight difference value, if the weight difference value is less than a preset threshold value, the avoidance scheme of both sides is obtained, and the avoidance grabbing device is obtained through the avoidance scheme, the independent prediction of the conflict grabbing device is realized, the communication data load is reduced, and the conflict avoidance is realized based on the dynamic weight value and the avoidance time.

Description

Technical Field

[0001] This invention relates to data processing technology, and more particularly to a conflict avoidance method and system for collaborative operation of multiple grasping targets. Background Technology

[0002] In modern logistics systems and flexible manufacturing lines, multiple automated gripping devices (such as mobile robots and automated forklifts) are typically deployed to work collaboratively in the same shared space to improve throughput efficiency. As the density of equipment deployment increases, frequent intersections of work paths become the norm, which places extremely high demands on the collaborative control capabilities of multi-agent systems.

[0003] However, existing multi-device conflict handling mechanisms generally suffer from "passive response" and "blind decision-making." Most devices rely solely on local sensors to detect current physical obstacles, lacking the ability to predict the future trajectories of other devices. This results in them only being able to adopt emergency stops or simple waiting strategies when a conflict is imminent, exhibiting significant "shortsightedness." Furthermore, current technologies rely on a central brain (master computer) for all path planning and scheduling. All devices need to continuously report their high-precision real-time status (position, speed) to the central controller while receiving detailed motion commands. This leads to massive network data traffic, and communication delays and packet loss risks directly affect the accuracy and security of control.

[0004] Therefore, how to endow the grasping device with independent trajectory prediction capabilities, enabling it to autonomously predict potential conflicts and make the optimal avoidance decision based on potential conflicts, has become an urgent problem to be solved in this field. Summary of the Invention

[0005] This invention provides a conflict avoidance method and system for collaborative operation of multiple grasping targets. It can endow grasping devices with independent path prediction capabilities, enabling them to autonomously predict potential conflicts and decide on the optimal avoidance subject based on potential conflicts. The distributed decision-making of each device realizes the collaboration of the group of devices, thereby improving work efficiency.

[0006] A first aspect of the present invention provides a conflict avoidance method for collaborative operation of multiple grasping targets, comprising:

[0007] The system acquires the current operating information and task information of the grasping device, generates an initial predicted trajectory based on the operating information and a preset motion model, generates a state vector based on the operating information and task information, encapsulates the state vector and the initial predicted trajectory to obtain a state beacon, and broadcasts the state beacon.

[0008] Receive status beacons sent by other grasping devices, extend the initial predicted trajectories of the current grasping device and other grasping devices to obtain the current updated trajectory and other updated trajectories respectively, compare the current updated trajectory and other updated trajectories to identify conflicting grasping devices;

[0009] Retrieve the weight values ​​of the current capturing device and the conflicting capturing devices, and determine the capturing device to avoid based on the weight values ​​of the current capturing device and the conflicting capturing devices.

[0010] Optionally, in one possible implementation of the first aspect, the step of acquiring the current operating information and task information of the grasping device, and generating an initial predicted trajectory based on the operating information and a preset motion model, includes:

[0011] The operational information includes equipment speed, equipment acceleration, and equipment position. The operational information is input into the motion model, and iterative calculations are performed based on a preset window time and time step to obtain multiple initial spatiotemporal nodes distributed according to the time step within the window time.

[0012] Obtain the time points corresponding to each initial spatiotemporal node, and obtain the confidence level of the initial spatiotemporal node based on the time points corresponding to the initial spatiotemporal nodes and the preset confidence level calculation model;

[0013] The spatial envelope range of each initial spatiotemporal node is determined based on the confidence level, and the initial predicted trajectory is obtained based on the initial spatiotemporal nodes and the spatial envelope range.

[0014] Optionally, in one possible implementation of the first aspect, generating the state vector based on runtime information and task information includes:

[0015] The task information includes the target location and task priority;

[0016] Retrieve the internal state of the current grasping device's internal state machine;

[0017] A state vector is constructed based on internal state, operational information, and task information.

[0018] Optionally, in one possible implementation of the first aspect, extending the initial predicted trajectories of the current grasping device and other grasping devices to obtain the current updated trajectory and other updated trajectories respectively includes:

[0019] The operation information of each grasping device is input into the motion model, and iterative calculations are performed based on the target position and time step to obtain the remaining spatiotemporal nodes of each grasping device distributed according to the time step.

[0020] The remaining predicted trajectories for each grasping device are obtained based on the remaining spatiotemporal nodes. The current updated trajectory is obtained based on the initial predicted trajectory and the remaining predicted trajectory of the current grasping device. Other updated trajectories are obtained based on the initial predicted trajectory and the remaining predicted trajectory of other grasping devices.

[0021] Optionally, in one possible implementation of the first aspect, the comparison of the current update trajectory with other update trajectories to determine the conflict-grabbing device includes:

[0022] Map the current updated trajectory and other updated trajectories to the same spatial model, and calculate the spatial distance between each position of the current updated trajectory and other updated trajectories in the spatial model;

[0023] When it is determined that there is a spatial distance smaller than the preset safety distance, the corresponding spatial distance is obtained as the risk distance, and the positions of the corresponding risk distances on the current update trajectory and other update trajectories are respectively used as the current risk position and other risk positions;

[0024] Extract the time point corresponding to the current risk location as the current risk time, and obtain the time points corresponding to other risk locations as other risk times;

[0025] Calculate the time difference between the current risk time and other risk times. If the time difference is less than a preset time threshold, take the corresponding other updated trajectories as conflict trajectories and obtain the other capturing devices corresponding to the conflict trajectories as conflict capturing devices.

[0026] Optionally, in one possible implementation of the first aspect, retrieving the weight values ​​of the current grasping device and the conflicting grasping devices, and determining the avoidance grasping device based on the weight values ​​of the current grasping device and the conflicting grasping devices, includes:

[0027] Obtain the weight values ​​of the current capturing device and the conflicting capturing device as the current weight value and the conflict weight value, respectively. Calculate the absolute value of the difference between the current weight value and the conflict weight value to obtain the weight difference.

[0028] The weight difference is compared with a preset weight threshold. If the weight difference is greater than the weight threshold, the grasping device corresponding to the lower weight value between the current weight value and the conflict weight value is selected as the avoidance grasping device.

[0029] When the weight difference is less than or equal to the weight threshold, obtain the avoidance scheme of the current grasping device and the conflicting grasping device, and determine the grasping device to avoid based on the avoidance scheme.

[0030] Optionally, in one possible implementation of the first aspect, obtaining the avoidance schemes of the current grasping device and the conflicting grasping devices, and determining the avoidance grasping device based on the avoidance schemes, includes:

[0031] Obtain the avoidance plans assigned by the user to the current capture device and the conflicting capture device respectively, determine the avoidance time based on the avoidance plan, obtain the avoidance time of the current capture device as the current avoidance time, and take the avoidance time of the conflicting capture device as the conflict avoidance time;

[0032] If the current avoidance time is compared with the conflict avoidance time, and it is determined that the current avoidance time is greater than the conflict avoidance time, the conflict capture device will be used as the avoidance capture device.

[0033] If the current avoidance time is determined to be less than the conflict avoidance time, the current grasping device will be used as the avoidance grasping device.

[0034] Optionally, in one possible implementation of the first aspect, determining the avoidance time based on the avoidance scheme includes:

[0035] The avoidance scheme includes a stationary avoidance scheme and a detour avoidance scheme. When the avoidance scheme is determined to be a stationary avoidance scheme, the dwell time corresponding to the stationary avoidance scheme is obtained as the avoidance time.

[0036] When the avoidance plan is determined to be a detour avoidance plan, the detour time corresponding to the detour avoidance plan is obtained, and the avoidance time is obtained based on the detour time.

[0037] Optionally, in one possible implementation of the first aspect, obtaining the weight values ​​of the current crawling device and the conflicting crawling devices includes:

[0038] The weight values ​​are obtained using the following formula.

[0039]

[0040] in, For weight values, As a task priority, Under load conditions As a balance factor, Contribution rate to historical conflicts These are the preset weighting coefficients.

[0041] A second aspect of the present invention provides a conflict avoidance system for collaborative operation of multiple grasping targets, comprising:

[0042] The information sending module is used to acquire the current operating information and task information of the grasping device, generate an initial predicted trajectory based on the operating information and a preset motion model, generate a state vector based on the operating information and task information, encapsulate the state vector and the initial predicted trajectory to obtain a state beacon, and broadcast the state beacon.

[0043] The information processing module is used to receive status beacons sent by other grasping devices, extend the initial predicted trajectories of the current grasping device and other grasping devices to obtain the current updated trajectory and other updated trajectories respectively, and compare the current updated trajectory and other updated trajectories to determine the conflicting grasping device.

[0044] The avoidance confirmation module is used to retrieve the weight values ​​of the current capturing device and the conflicting capturing device, and determine the avoidance capturing device based on the weight values ​​of the current capturing device and the conflicting capturing device.

[0045] The beneficial effects of this invention are as follows:

[0046] 1. This invention reduces the data packet size of a single communication and alleviates the pressure on wireless network bandwidth by acquiring the task information of the capturing device and generating a state beacon containing the initial predicted trajectory and state vector for broadcast transmission. The capturing device, acting as an independent entity, analyzes and processes the received state beacon, extending the initial predicted trajectory using the state vector in the beacon, thereby calculating the complete current updated trajectory and other updated trajectories locally. This invention avoids the capturing device broadcasting the full trajectory data in real time. Furthermore, each capturing device can independently complete trajectory extension and deduction for other devices by exchanging state beacons.

[0047] 2. This invention maps the current updated trajectory and other updated trajectories into the same spatial model. First, it filters out risky locations whose spatial distance is less than a preset safety distance. Then, it further extracts and calculates the time difference between the current risk time and other risk times for reaching the risky location. Only when this time difference is less than a preset time threshold is the corresponding trajectory marked as a conflict trajectory. This invention effectively eliminates situations where two trajectories are close in spatial coordinates but have a safety interval in arrival time, allowing equipment to pass under the premise of meeting the safety time difference, avoiding unnecessary downtime caused by judging solely based on spatial distance.

[0048] 3. This invention calculates weight values ​​based on task priority, load status, balance factor, and historical conflict contribution rate, and determines the avoidance and grabbing devices in layers: When the weight difference is greater than a preset weight threshold, the device with the lower weight value is directly allowed to avoid the grabbing device, ensuring the passage of high-priority or high-load devices; when the weight difference is small, the avoidance time required for both parties to execute a stationary avoidance plan or a detour avoidance plan is further compared, and the device with the shorter avoidance time is selected to avoid the grabbing device. This invention, through dynamic evaluation of weight values ​​and avoidance costs, transforms conflict resolution from rules based on ID identity or fixed priority in existing technologies to evaluation based on real-time value and status, ensuring that each conflict resolution serves the overall efficiency at the current moment and reducing overall operation delays. Attached Figure Description

[0049] Figure 1 A flowchart illustrating a conflict avoidance method for collaborative multi-target grabbing operations provided by this invention;

[0050] Figure 2 A schematic diagram of the structure of a conflict avoidance system for collaborative operation of multiple grasping targets provided by the present invention;

[0051] Figure 3 This is a schematic diagram of the hardware structure of an electronic device provided by the present invention. Detailed Implementation

[0052] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0053] This invention provides a conflict avoidance method for collaborative operations involving multiple grasping targets, such as... Figure 1 As shown, it includes:

[0054] S1. Obtain the current operating information and task information of the grasping device, generate an initial predicted trajectory based on the operating information and a preset motion model, generate a state vector based on the operating information and task information, encapsulate the state vector and the initial predicted trajectory to obtain a state beacon, and broadcast the state beacon.

[0055] It should be noted that in industrial scenarios with high concurrency operations of multiple grabbing devices (such as large automated warehouses or flexible manufacturing workshops), the operation paths of the devices will inevitably intersect during dynamic operation. The devices often adopt conservative strategies (such as stopping and waiting when encountering potential intersections) or get into a "deadlock" state where they wait for each other to pass first. This results in a large amount of non-productive waiting time and low equipment utilization.

[0056] Existing technologies also offer predictive planning for the paths of grasping devices. However, systems that rely on a central brain (such as a main control computer) for all path planning and scheduling are susceptible to paralysis or unpredictable dangerous behavior if the main control center experiences hardware failure, software crashes, or network latency. Furthermore, if each device broadcasts its complete trajectory data from its current location to the task endpoint (typically containing hundreds of spatiotemporal coordinates) in real time, the massive data interaction will instantly exhaust the bandwidth of the wireless communication network as the number of devices increases, causing severe channel congestion and data packet loss, leading to delayed collision detection and even safety incidents. Moreover, existing planning and scheduling strategies are simplistic, such as always prioritizing "device 1 over device 2" or "north-south traffic over east-west traffic." While these rules are effective in specific scenarios, they are unsuitable for all situations.

[0057] Therefore, this invention endows the grasping device with independent path prediction capabilities, enabling it to autonomously anticipate potential conflicts and determine the optimal avoidance agent based on these conflicts. This distributed decision-making among the devices achieves collaboration within the group, improving work efficiency. This invention can describe the future movement trends of the devices while maintaining minimal communication data. Each grasping device, as an intelligent agent with perception, communication, and decision-making capabilities, actively predicts potential conflicts during operation by exchanging spatiotemporal state information in real time, including position, speed, task status, and future intentions. Based on dynamically calculated priority weights, it performs rapid distributed negotiation and autonomously adjusts its behavior to eliminate conflicts.

[0058] Understandably, for the transmission of smaller amounts of communication data, this step combines and encapsulates the local high-precision trajectory and state vector of the grasping device: each grasping device, acting as an independent device, only calculates and broadcasts the trajectory within a very short future time window (initial predicted trajectory), and combines it with information data (state vector) for long-term projection, packaging it into a lightweight state beacon. This invention's design, which trades computing power for bandwidth, aims to minimize the data load of a single communication while ensuring real-time interactive capabilities in large-scale clusters.

[0059] In this invention, the "current grasping device" refers to the automated operating equipment that acts as an independent executor, such as a mobile robot or unmanned forklift. From a first-person perspective, each grasping device is portrayed as an intelligent agent with perception, communication, and decision-making capabilities. During operation, it actively predicts potential conflicts by exchanging spatiotemporal state information in real time, including position, speed, task status, and future intentions. Based on dynamically calculated priority weights, it conducts rapid distributed negotiation and autonomously adjusts its behavior to eliminate conflicts. The "state beacon" refers to the communication data packet generated by the grasping device and periodically broadcast externally, containing the initial predicted trajectory and state vector. The "operational information" refers to the set of physical quantities collected in real time by sensors during operation, characterizing the current physical motion state of the device.

[0060] In some embodiments, step S1 (obtaining the current operating information and task information of the grasping device, and generating an initial predicted trajectory based on the operating information and a preset motion model) includes S11-S13:

[0061] S11, the operation information includes device speed, device acceleration and device position. The operation information is input into the motion model and iteratively calculated based on the preset window time and time step to obtain multiple initial spatiotemporal nodes distributed according to the time step within the window time.

[0062] It should be noted that this invention establishes a global coordinate system for the work area. All grasping devices complete their initial positioning through SLAM mapping or preset markers (such as QR codes), ensuring that each grasping device has a consistent understanding of its spatial location. Furthermore, the system clocks of all grasping devices are synchronized (using NTP or PTP protocols) to ensure unified timestamps, which is the foundation for accurate spatiotemporal state prediction.

[0063] Understandably, each grasping device will integrate data from pre-set sensors, such as joint encoders, IMUs, vision cameras, and LiDAR. This step focuses on the current grasping device and obtains operational information based on the sensor data, including device speed, device acceleration, and device position.

[0064] Specifically, the preset motion model is retrieved, the running information is input into the motion model, and a time window is set. and time step Circular calculations are performed using a motion model, from Initially, increase by one time step each time. until ,in, The cumulative relative time serves as the control variable for iterative calculations, representing the length of time already deduced. In each iteration, the calculated spatial coordinates and time points are combined to generate an initial spatiotemporal node.

[0065] In this context, "device position" refers to the real-time position of the grasping device, "device speed" refers to the real-time speed of the grasping device, "device acceleration" refers to the real-time acceleration of the grasping device, "motion model" refers to a pre-defined set of mathematical equations describing the motion of the grasping device, and "window time" refers to a pre-defined upper limit for acquiring the initial predicted trajectory, such as 3 seconds. "Time step" refers to the time interval between two adjacent initial spatiotemporal nodes in the iterative calculation, and "initial spatiotemporal node" refers to the data nodes that constitute the initial predicted trajectory obtained through the motion model.

[0066] S12, obtain the time points corresponding to each initial spatiotemporal node, and obtain the confidence of the initial spatiotemporal node based on the time points corresponding to the initial spatiotemporal nodes and the preset confidence calculation model.

[0067] Understandably, for each initial spatiotemporal node, the corresponding time point is extracted, and the time point is substituted into the preset confidence calculation model to obtain the confidence level corresponding to each initial spatiotemporal node.

[0068] Among them, the time point corresponding to the initial spatiotemporal node This refers to the time corresponding to the initial spacetime node. , The current system time. Index of iteration steps ( The confidence level calculation model refers to a pre-defined mathematical function used to describe the decay of a system's predictive ability over time. The confidence level represents a probability value between 0 and 1; it indicates whether the capturing device is capable of predicting the outcome of a given event. The moment will indeed appear The degree of certainty regarding the event of "coordinates". For predicted coordinates.

[0069] S13, determine the spatial envelope range of each initial spatiotemporal node based on the confidence level, and obtain the initial predicted trajectory based on the initial spatiotemporal nodes and the spatial envelope range.

[0070] Understandably, by using the confidence level calculated in step S12 as an adjustment lever, the safety volume corresponding to each initial spatiotemporal node is dynamically adjusted to construct a funnel-shaped trajectory that converges at the near end and diverges at the far end, thereby achieving a dynamic balance between passability and safety.

[0071] Specifically, based on a pre-set spatial envelope radius With confidence level From the corresponding table, extract the spatial envelope radius corresponding to each confidence level. For example, when the confidence level is... hour, For 500mm, at a confidence level of hour, The value is 900mm. With each initial spatiotemporal node as the center and the spatial envelope radius as the radius, a spatial sphere of the initial spatiotemporal node is constructed as the spatial envelope range, thereby forming a tubular channel with volume properties that gradually thickens over time, i.e., the initial predicted trajectory.

[0072] The spatial envelope refers to the range of time points. The maximum spatial area that the time-capture device may exist in is the initial predicted trajectory, which refers to the trajectory segment constructed based on the initial spatiotemporal nodes.

[0073] In some embodiments, step S1 (generating a state vector based on runtime information and task information) includes S14-S16:

[0074] S14, the task information includes the target location and task priority.

[0075] It is easy to understand that in this invention, the parent server decomposes the task and distributes the task information of the decomposed sub-tasks to each capture device. The task information includes the target location and the task priority.

[0076] Among them, task information refers to the task data information received by the grasping device, task target location refers to the endpoint location specified in the task information for the grasping device to perform the operation, and task priority refers to the urgency level specified in the task information for the grasping device to perform the operation.

[0077] S15, obtain the internal state of the current grasping device's internal state machine.

[0078] Understandably, obtaining the internal state of the current grasping device's internal state machine provides a time-sensitive and highly reliable basis for acceleration correction in subsequent trajectory extension steps. Examples include MOVING_TO_TARGET, GRASPING, and WAITING.

[0079] The internal state refers to the state information in the internal state machine used to characterize the motion phase of the grasping device.

[0080] S16, construct a state vector based on internal state, operational information, and task information.

[0081] Understandably, by using state vectors, other grasping devices have access to all the data that reproduces the current grasping device's movement patterns, thus achieving the most complete prediction capability at the lowest bandwidth cost.

[0082] Here, the state vector refers to a data set that contains the motion characteristics of the grasping device.

[0083] S2, receive status beacons sent by other grasping devices, extend the initial predicted trajectories of the current grasping device and other grasping devices to obtain the current updated trajectory and other updated trajectories respectively, compare the current updated trajectory and other updated trajectories to determine the conflicting grasping device.

[0084] Understandably, after each grasping device generates a state beacon, it broadcasts it, and other grasping devices receive and analyze it. To ensure real-time communication, the grasping devices only broadcast the initial predicted trajectory (e.g., the trajectory for the next 3 seconds). If only the initial predicted trajectory is compared, conflicts will still occur between grasping devices. Therefore, in step S2, the current grasping device uses the received state vector to supplement the remaining predicted trajectories of other grasping devices, thereby expanding the scope of conflict detection based on the state beacon.

[0085] Other grasping devices refer to grasping devices other than the current grasping device.

[0086] In some embodiments, step S2 (extending the initial predicted trajectories of the current grasping device and other grasping devices to obtain the current updated trajectory and other updated trajectories respectively) includes S21-S22:

[0087] S21, input the operation information of each grasping device into the motion model, and perform iterative calculations based on the target position and time step to obtain the remaining spatiotemporal nodes of each grasping device distributed according to the time step.

[0088] It should be noted that in the distributed communication of this invention, in order to save bandwidth, the initial predicted trajectory broadcast by the capturing device usually only covers a short window time, while the actual task execution cycle is often longer than this window time. In order to achieve collision detection throughout the entire process, the current capturing device needs to predict the trajectory from the window time up to the target location.

[0089] Understandably, extracting the operating information of each grasping device and inputting it into the motion model, based on step S11, and based on the device position, task target position, and time step, is necessary. The motion model is used for iterative calculations until the device position reaches the mission target position. In each iteration, the calculated spatial coordinates and time points are combined to generate a remaining spatiotemporal node.

[0090] S22, based on the remaining spatiotemporal nodes, obtain the remaining predicted trajectory corresponding to each grasping device, obtain the current updated trajectory based on the initial predicted trajectory and remaining predicted trajectory of the current grasping device, and obtain other updated trajectories based on the initial predicted trajectory and remaining predicted trajectory of other grasping devices.

[0091] Understandably, referring to steps S12 and S13, for each remaining spatiotemporal node, the corresponding time point is extracted, and the time point is substituted into the preset confidence calculation model to obtain the confidence level for each remaining spatiotemporal node. Similarly, based on the pre-defined correspondence table between spatial envelope radius and confidence level, the spatial envelope radius corresponding to each confidence level is extracted. With each remaining spatiotemporal node as the center and the spatial envelope radius as the radius, a spatial sphere of the remaining spatiotemporal node is constructed as the spatial envelope range of the remaining spatiotemporal node, thus forming a tubular channel with volume properties that gradually thickens over time, i.e., the remaining predicted trajectory.

[0092] Furthermore, the initial and remaining predicted trajectories of the current grasping device are loaded to obtain a complete predicted trajectory as the current updated trajectory. The initial and remaining predicted trajectories of other grasping devices are loaded to obtain a complete predicted trajectory as the other updated trajectories.

[0093] Among them, the remaining predicted trajectory refers to the trajectory segment constructed by the remaining spatiotemporal nodes, the current updated trajectory refers to the complete trajectory of the current grasping device completing the task, and the other updated trajectories refer to the complete trajectories of other grasping devices besides the current grasping device completing the task.

[0094] In some embodiments, step S2 (comparing the current update trajectory with other update trajectories to determine conflicting capture devices) includes S23-S26:

[0095] S23, map the current updated trajectory and other updated trajectories to the same spatial model, and calculate the spatial distance between each position of the current updated trajectory and other updated trajectories in the spatial model.

[0096] It's easy to understand that, based on the global coordinate system of the work area mentioned in step S11, the current updated trajectory and other updated trajectories are mapped to the same spatial model. A full trajectory scan is performed on the current updated trajectory and other updated trajectories to obtain the spatial distance between each position (the set of all points on the trajectory) on the current updated trajectory and other updated trajectories. For example, if the current updated trajectory and other updated trajectories extend from the same baseline in the same direction, this baseline is translated, and the intersection points of the baseline with the current updated trajectory and other updated trajectories are obtained. The distance between the intersection points on the current updated trajectory and the intersection points on other updated trajectories is the spatial distance. The intersection points on these two trajectories are all close to each other. Of course, other existing technologies can also be used to obtain the spatial distance.

[0097] Among them, the spatial model refers to the three-dimensional environment model used to uniformly describe the positional relationship between the current updated trajectory and other updated trajectories, and the spatial distance refers to the distance between various positions on the current updated trajectory and other updated trajectories.

[0098] S24, when it is determined that there is a spatial distance less than the preset safety distance, the corresponding spatial distance is obtained as the risk distance, and the positions of the corresponding risk distances on the current update trajectory and other update trajectories are respectively used as the current risk position and other risk positions.

[0099] Understandably, the system reads the preset safety distance and compares each spatial distance with the preset safety distance. If the spatial distance is greater than or equal to the preset safety distance, there is no potential conflict between the current update trajectory and other update trajectories. If the spatial distance is less than the preset safety distance, there is a potential conflict between the current update trajectory and other update trajectories. The corresponding safety distance is then taken as the risk distance, and the positions of the corresponding risk distances on the current update trajectory and other update trajectories are taken as the current risk position and other risk positions, respectively.

[0100] It should be noted that adjustments are needed if there is a risk distance between the two trajectories. Adjustments should be made at the point where the first risk distance occurs for greater safety.

[0101] The current risk location refers to the coordinate position on the current updated trajectory that meets the risk distance requirement, while other risk locations refer to the coordinate positions on other updated trajectories that correspond to the current risk location.

[0102] S25, extract the time point corresponding to the current risk location as the current risk time, and obtain the time points corresponding to other risk locations as other risk times.

[0103] It is easy to understand that the current risk time refers to the estimated time when the current grasping device will arrive at the current risk location, while the other risk times refer to the estimated time when other grasping devices will arrive at other risk locations.

[0104] The time point corresponding to the current risk location refers to the moment at the current risk location, while the time point corresponding to other risk locations refers to the moment at other risk locations.

[0105] S26, calculate the time difference between the current risk time and other risk times, and determine that when the time difference is less than a preset time threshold, take the corresponding other updated trajectories as conflict trajectories, and obtain the other capturing devices corresponding to the conflict trajectories as conflict capturing devices.

[0106] Understandably, although potential spatial conflicts between the two trajectories were detected in steps S23-S24, physical trajectory conflicts are not equivalent to instantaneous collisions. If the time interval between the arrival of the two grasping devices at the risk location is sufficiently large (e.g., vehicle A crosses the intersection at 10:00, and vehicle B crosses the intersection at 10:01), then there is absolutely no need for avoidance.

[0107] Specifically, a preset time threshold is retrieved and compared with the time difference. If the time difference is greater than or equal to the preset time, it means that although the two grasping devices have conflicting trajectories, their timing is staggered, as if they are passing through alternately, so there is no risk. If the time difference is less than the preset time threshold, it means that the two grasping devices arrive at their close positions at almost the same time, creating a risk. In this case, other updated trajectories are identified as conflicting trajectories, and other grasping devices corresponding to the conflicting trajectories are identified as conflicting grasping devices.

[0108] Among them, the preset time threshold refers to the minimum time safety interval set in advance, the conflict trajectory refers to the specific other update trajectory that is confirmed to have substantial spatiotemporal intersection with the current update trajectory of the current capturing device after being judged by both spatial distance and time difference, and the time difference refers to the time interval.

[0109] S3 retrieves the weight values ​​of the current capturing device and the conflicting capturing device, and determines the capturing device to avoid based on the weight values ​​of the current capturing device and the conflicting capturing device.

[0110] Understandably, once the conflict-grabbing device is identified, allowing both the current grabbing device and the conflict-grabbing device to choose to avoid each other would lead to unnecessary efficiency reduction. Allowing both devices to rush through would also result in a collision. Therefore, this invention quantifies the urgency of the task and the higher avoidance cost by comparing the weight values ​​of the current grabbing device and the conflict-grabbing device, establishing the device with the lower weight as the one to avoid.

[0111] Among them, the weight value refers to the numerical index used to quantify the priority of the crawling device's right-of-way, and the crawling device to be avoided refers to the crawling device that needs to be avoided in the conflict game.

[0112] In some embodiments, step S3 (retrieving the weight values ​​of the current grasping device and the conflicting grasping devices, and determining the avoidance grasping device based on the weight values ​​of the current grasping device and the conflicting grasping devices) includes S31-S33:

[0113] S31, obtain the weight values ​​of the current capturing device and the conflict capturing device as the current weight value and the conflict weight value respectively, calculate the absolute value of the difference between the current weight value and the conflict weight value, and obtain the weight difference value.

[0114] In some embodiments, step S31 (obtaining the weight values ​​of the current grasping device and the conflicting grasping device) includes S311:

[0115] S311, the weight value is obtained through the following formula.

[0116]

[0117] in, For weight values, As a task priority, Under load conditions As a balance factor, Contribution rate to historical conflicts These are the preset weighting coefficients.

[0118] Understandably, reading task information yields task priority, which can be categorized as 1, 0.8, 0.5, etc.; acquiring motor current values ​​or pressure sensor data provides load status, such as 1.0 for full load, 0.5 for half load, and 0 for no load; reading pre-set balance factors, which are correction coefficients in the weighting formula used to reflect the current environmental state, such as a larger balance factor for a device attempting to leave the area when multiple grasping devices are crowded in the same area, and a smaller balance factor for a device attempting to enter the area; and reading the historical game records of the grasping devices yields the historical conflict contribution rate, for example, a historical conflict contribution rate of 0.8 if the device wins 4 out of 5 games, and 0.2 if it wins 1 out of 5 games.

[0119] Furthermore, As a positive gain term, , , The larger, The higher the value, the greater the priority of the right-of-way. As a negative penalty, The larger, The lower the priority, the more it indicates that the priority is forcibly reduced to achieve polling fairness.

[0120] Among them, load status refers to a value that characterizes the load of the grabbing device, and historical conflict contribution rate refers to the frequency with which the grabbing device is judged to be the dominant party (winner) in historical conflict games.

[0121] It should be noted that a Bayesian optimization framework can also be used to automatically adjust the dynamic weight formula.

[0122] S32, compare the weight difference with a preset weight threshold, and when the weight difference is greater than the weight threshold, select the grasping device corresponding to the lower weight value between the current weight value and the conflict weight value as the avoidance grasping device.

[0123] Understandably, this involves retrieving a preset weight threshold and then comparing the weight threshold with the weight difference. When comparing the weights, if the weight difference is greater than the weight threshold, it indicates that there may be a fundamental difference in the nature of the tasks. For example, one party's grabbing device is for emergency dispatch, while the other party's grabbing device is for routine dispatch. In order to ensure high-value tasks, the grabbing device with the lower weight value between the current weight value and the conflict weight value is selected as the grabbing device to avoid.

[0124] S33, when the weight difference is less than or equal to the weight threshold, obtain the avoidance scheme of the current grasping device and the conflicting grasping device, and determine the avoidance grasping device based on the avoidance scheme.

[0125] Understandably, when the weight difference is less than or equal to the weight threshold, it indicates that the weights of the two devices are similar. If, in this case, one device is ordered to yield based on a smaller weight difference, an unreasonable situation might occur: the device with the slightly lower weight might be navigating a narrow corridor (requiring a 20-meter reversing maneuver, taking 5 minutes), while the device with the slightly higher weight is next to a wide waiting area (requiring only a 0.5-meter lateral movement, taking 1 minute). Therefore, for both the current and conflicting grasping devices, a specific yielding scheme generated by the user is retrieved, and the device to yield is determined based on this scheme.

[0126] Among them, the avoidance scheme refers to the candidate trajectory or speed control command generated by the user terminal for the capturing device to eliminate the conflict.

[0127] In some embodiments, step S33 (obtaining the avoidance schemes of the current grasping device and the conflicting grasping device, and determining the avoidance grasping device based on the avoidance schemes) includes S331-S333:

[0128] S331, obtain the avoidance schemes assigned by the user terminal to the current capturing device and the conflict capturing device respectively, determine the avoidance time based on the avoidance scheme, obtain the avoidance time of the current capturing device as the current avoidance time, and take the avoidance time of the conflict capturing device as the conflict avoidance time.

[0129] Understandably, when the weight values ​​of the current capturing device and the conflict capturing device are similar, it is necessary to determine the conflict capturing device based on the avoidance time cost of each device. First, obtain the specific avoidance plan from the user end (or navigation planning layer), and then calculate the additional time required to execute the avoidance plan (i.e., avoidance time).

[0130] The user end refers to the navigation planning module inside the host computer system or equipment responsible for obstacle avoidance planning, or it can be a planner from existing technologies.

[0131] In some embodiments, step S331 (determining the avoidance time based on the avoidance plan) includes S3311-S3312:

[0132] S3311, the avoidance scheme includes a stationary avoidance scheme and a detour avoidance scheme. When the avoidance scheme is determined to be a stationary avoidance scheme, the dwell time corresponding to the stationary avoidance scheme is obtained as the avoidance time.

[0133] It is understandable that avoidance strategies include stationary avoidance strategies and detour avoidance strategies. A stationary avoidance strategy refers to the strategy where the grabbing device reduces its speed to zero and maintains this speed for a specific period of time to allow other devices to pass first. A detour avoidance strategy refers to the strategy where, while keeping the target position unchanged, the original travel path is actively modified and a path that can bypass the risky location is planned.

[0134] Furthermore, when the avoidance plan is determined to be a stationary avoidance plan, the dwell time corresponding to the stationary avoidance plan is obtained as the avoidance time, where the dwell time refers to the time span during which the grasping device executes the stationary avoidance plan.

[0135] S3312, when it is determined that the avoidance plan is a detour avoidance plan, the detour time corresponding to the detour avoidance plan is obtained, and the avoidance time is obtained based on the detour time.

[0136] Understandably, when the avoidance plan is a detour, the detour time configured by the user is obtained, and the detour starting point corresponding to the detour path is extracted. The estimated completion time from the detour starting point to the task target position on the currently updated trajectory or conflict trajectory is obtained as the comparison time, and the difference between the detour time and the comparison time is calculated to obtain the avoidance time.

[0137] Among them, detour time refers to the total time required to complete the detour path, and comparison time refers to the theoretical time required for the capturing device to travel from the detour starting point to the task target location along the current updated trajectory or conflict trajectory without conflict or avoidance.

[0138] S332, compare the current avoidance time and the conflict avoidance time. If the current avoidance time is greater than the conflict avoidance time, the conflict capture device will be used as the avoidance capture device.

[0139] It is understandable that when the current avoidance time is greater than the conflict avoidance time, it means that the time cost of the current grasping device to avoid conflict is high, while the time cost of the conflict grasping device to avoid conflict is low. Therefore, the conflict grasping device is used as the avoidance grasping device.

[0140] S333: If the current avoidance time is less than the conflict avoidance time, the current grasping device will be used as the avoidance grasping device.

[0141] It is easy to understand that when the current avoidance time is less than the conflict avoidance time, it means that the time cost of the current grasping device avoiding the conflict is lower, while the time cost of the grasping device avoiding the conflict is higher. Therefore, the current grasping device is used as the avoidance grasping device.

[0142] Furthermore, when the current avoidance time equals the conflict avoidance time, the user can directly specify the avoidance capture device.

[0143] It should be noted that if a status beacon broadcast by a certain grasping device is continuously lost, the corresponding grasping device will be assumed to be a static obstacle, and obstacle avoidance will be performed using only local sensors. If the current grasping device and other grasping devices fail to reach an agreement within a predetermined time when determining to avoid the grasping device, both parties will execute the default safety policy (such as both parties slowing down to a stop and then passing through in order of device ID). If the local sensors of each grasping device detect unmodeled static or dynamic obstacles (such as fallen goods or people), they will immediately switch to emergency obstacle avoidance mode and broadcast an emergency status notification. Surrounding devices will treat it as the highest priority dynamic obstacle and avoid it.

[0144] In addition, the master server in this invention not only allocates tasks, but also monitors all crawling devices and stores the data uploaded by all crawling devices to form an experience pool. This allows the use of digital twin technology to test the safety and effectiveness of the trained avoidance scheme in a virtual environment.

[0145] See Figure 2 This is a schematic diagram of a conflict avoidance system for collaborative multi-target grabbing provided in an embodiment of the present invention. The system includes:

[0146] The information sending module is used to acquire the current operating information and task information of the grasping device, generate an initial predicted trajectory based on the operating information and a preset motion model, generate a state vector based on the operating information and task information, encapsulate the state vector and the initial predicted trajectory to obtain a state beacon, and broadcast the state beacon.

[0147] The information processing module is used to receive status beacons sent by other grasping devices, extend the initial predicted trajectories of the current grasping device and other grasping devices to obtain the current updated trajectory and other updated trajectories respectively, and compare the current updated trajectory and other updated trajectories to determine the conflicting grasping device.

[0148] The avoidance confirmation module is used to retrieve the weight values ​​of the current capturing device and the conflicting capturing device, and determine the avoidance capturing device based on the weight values ​​of the current capturing device and the conflicting capturing device.

[0149] See Figure 3 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention. The electronic device 40 includes: a processor 41, a memory 42, and a computer program; wherein,

[0150] The memory 42 is used to store the computer program, and the memory may also be flash memory. The computer program is, for example, an application program or functional module that implements the above method.

[0151] The processor 41 is configured to execute the computer program stored in the memory to implement the various steps performed by the device in the above method. For details, please refer to the relevant descriptions in the preceding method embodiments.

[0152] Alternatively, the memory 42 can be either standalone or integrated with the processor 41.

[0153] When the memory 42 is a device independent of the processor 41, the device may further include:

[0154] Bus 43 is used to connect the memory 42 and the processor 41.

[0155] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A conflict avoidance method for collaborative operations involving multiple grasping targets, characterized in that, include: The system acquires the current operating information and task information of the grasping device, generates an initial predicted trajectory based on the operating information and a preset motion model, generates a state vector based on the operating information and task information, encapsulates the state vector and the initial predicted trajectory to obtain a state beacon, and broadcasts the state beacon. Receive status beacons sent by other grasping devices, extend the initial predicted trajectories of the current grasping device and other grasping devices to obtain the current updated trajectory and other updated trajectories respectively, compare the current updated trajectory and other updated trajectories to determine the conflicting grasping devices; Retrieve the weight values ​​of the current capturing device and the conflicting capturing devices, and determine the capturing device to avoid based on the weight values ​​of the current capturing device and the conflicting capturing devices.

2. The method according to claim 1, characterized in that, The step of acquiring the current operating information and task information of the grasping device, and generating an initial predicted trajectory based on the operating information and a preset motion model, includes: The operational information includes equipment speed, equipment acceleration, and equipment position. The operational information is input into the motion model, and iterative calculations are performed based on a preset window time and time step to obtain multiple initial spatiotemporal nodes distributed according to the time step within the window time. Obtain the time point corresponding to each initial spatiotemporal node, and obtain the confidence level of the initial spatiotemporal node based on the time point corresponding to the initial spatiotemporal node and the preset confidence level calculation model; The spatial envelope range of each initial spatiotemporal node is determined based on the confidence level, and the initial predicted trajectory is obtained based on the initial spatiotemporal nodes and the spatial envelope range.

3. The method according to claim 2, characterized in that, The generation of the state vector based on operational information and task information includes: The task information includes the target location and task priority; Retrieve the internal state of the current grasping device's internal state machine; A state vector is constructed based on internal state, operational information, and task information.

4. The method according to claim 3, characterized in that, The process of extending the initial predicted trajectories of the current grasping device and other grasping devices to obtain the current updated trajectory and other updated trajectories respectively includes: The operation information of each grasping device is input into the motion model, and iterative calculations are performed based on the target position and time step to obtain the remaining spatiotemporal nodes of each grasping device distributed according to the time step. The remaining predicted trajectories for each grasping device are obtained based on the remaining spatiotemporal nodes. The current updated trajectory is obtained based on the initial predicted trajectory and the remaining predicted trajectory of the current grasping device. Other updated trajectories are obtained based on the initial predicted trajectory and the remaining predicted trajectory of other grasping devices.

5. The method according to claim 4, characterized in that, The step of comparing the current update trajectory with other update trajectories to determine conflicting capture devices includes: Map the current updated trajectory and other updated trajectories to the same spatial model, and calculate the spatial distance between each position of the current updated trajectory and other updated trajectories in the spatial model; When it is determined that there is a spatial distance smaller than the preset safety distance, the corresponding spatial distance is obtained as the risk distance, and the positions of the corresponding risk distances on the current update trajectory and other update trajectories are respectively used as the current risk position and other risk positions; Extract the time point corresponding to the current risk location as the current risk time, and obtain the time points corresponding to other risk locations as other risk times; Calculate the time difference between the current risk time and other risk times. If the time difference is less than a preset time threshold, take the corresponding other updated trajectories as conflict trajectories and obtain the other capturing devices corresponding to the conflict trajectories as conflict capturing devices.

6. The method according to claim 5, characterized in that, The step of retrieving the weight values ​​of the current capturing device and the conflicting capturing devices, and determining the avoidance capturing device based on the weight values ​​of the current capturing device and the conflicting capturing devices, includes: Obtain the weight values ​​of the current capturing device and the conflicting capturing device as the current weight value and the conflict weight value, respectively. Calculate the absolute value of the difference between the current weight value and the conflict weight value to obtain the weight difference. The weight difference is compared with a preset weight threshold. If the weight difference is greater than the weight threshold, the grasping device corresponding to the lower weight value between the current weight value and the conflict weight value is selected as the avoidance grasping device. When the weight difference is less than or equal to the weight threshold, obtain the avoidance scheme of the current grasping device and the conflicting grasping device, and determine the grasping device to avoid based on the avoidance scheme.

7. The method according to claim 6, characterized in that, The step of obtaining the avoidance scheme of the current capturing device and the conflicting capturing device, and determining the avoidance device based on the avoidance scheme, includes: Obtain the avoidance plans assigned by the user to the current capture device and the conflicting capture device respectively, determine the avoidance time based on the avoidance plan, obtain the avoidance time of the current capture device as the current avoidance time, and take the avoidance time of the conflicting capture device as the conflict avoidance time; If the current avoidance time is compared with the conflict avoidance time, and it is determined that the current avoidance time is greater than the conflict avoidance time, the conflict capture device will be used as the avoidance capture device. If the current avoidance time is determined to be less than the conflict avoidance time, the current grasping device will be used as the avoidance grasping device.

8. The method according to claim 7, characterized in that, The determination of the avoidance time based on the avoidance scheme includes: The avoidance scheme includes a stationary avoidance scheme and a detour avoidance scheme. When the avoidance scheme is determined to be a stationary avoidance scheme, the dwell time corresponding to the stationary avoidance scheme is obtained as the avoidance time. When the avoidance plan is determined to be a detour avoidance plan, the detour time corresponding to the detour avoidance plan is obtained, and the avoidance time is obtained based on the detour time.

9. The method according to claim 6, characterized in that, The step of obtaining the weight values ​​of the current crawling device and the conflicting crawling devices includes: The weight values ​​are obtained using the following formula. in, For weight values, As a task priority, Under load conditions As a balance factor, Contribution rate to historical conflicts These are the preset weighting coefficients.

10. A conflict avoidance system for multi-target cooperative operation according to any one of claims 1-9, characterized in that, include: The information sending module is used to acquire the current operating information and task information of the grasping device, generate an initial predicted trajectory based on the operating information and a preset motion model, generate a state vector based on the operating information and task information, encapsulate the state vector and the initial predicted trajectory to obtain a state beacon, and broadcast the state beacon. The information processing module is used to receive status beacons sent by other grasping devices, extend the initial predicted trajectories of the current grasping device and other grasping devices to obtain the current updated trajectory and other updated trajectories respectively, and compare the current updated trajectory and other updated trajectories to determine the conflicting grasping device. The avoidance confirmation module is used to retrieve the weight values ​​of the current capturing device and the conflicting capturing device, and determine the avoidance capturing device based on the weight values ​​of the current capturing device and the conflicting capturing device.

Images