Automatic driving multi-agent future behavior topology reasoning method, device, equipment, medium and product

By introducing behavior topology representation and collaborative learning model BeTopNet in the autonomous driving system, combining dense and sparse representation, the problem of unstable behavioral patterns in multi-agent driving scenarios is solved, and more efficient and consistent joint prediction and planning is achieved, improving The safety and accuracy of the autonomous driving system.

CN118966349AActive Publication Date: 2024-11-15SHANGHAI ARTIFICIAL INTELLIGENCE INNOVATION CENT

Patent Information

Application Number
CN202410933027.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-12
Publication Date
2024-11-15
Estimated Expiration
2044-07-12

AI Technical Summary

Technical Problem

In the current autonomous driving system, uncertainty and heterogeneous interactions in multi-agent driving scenarios lead to dense and sparse future behaviors that indicate inefficiency and inconsistency in joint prediction and planning, resulting in instability in multi-agent behavior patterns.

Method used

A behavioral topology (BeTop) representation method based on topological knot theory is adopted, and the collaborative learning model BeTopNet combines dense and sparse representations to generate future trajectory clusters to enhance the consistency and accuracy of joint prediction and planning.

Benefits of technology

By explicitly characterizing the consensus behavior pattern of future interactions of multi-agents, the stability and accuracy of multi-agent behaviors are improved, computing overhead is reduced, and the safety and compliance of autonomous driving systems are enhanced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118966349A_ABST
    Figure CN118966349A_ABST
Patent Text Reader

Abstract

The invention discloses an automatic driving multi-agent future behavior topology reasoning method, device and equipment, a medium and a product. The method comprises the steps of obtaining position information of a current vehicle and position information of each agent in an agent set included in a scene where the current vehicle is located; the position information of the current vehicle and the position information of each agent are input into a collaborative learning model, a future behavior topology reasoning result is obtained, the collaborative learning model is obtained by iteratively training an initial model through a training sample set, and a future behavior topology reasoning result is obtained; the training sample set comprises an agent set sample and a future behavior topology reasoning result sample corresponding to the agent set sample. Through the technical scheme of the invention, the consistency and accuracy of joint prediction and planning of multi-agent behaviors can be enhanced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of autonomous driving technology, and in particular to a method, device, equipment, medium and product for topological reasoning of future behaviors of multi-agents in autonomous driving. Background Art

[0002] Autonomous driving systems aim to achieve safe and socially compliant driving through behavioral integration between interactive agents. However, the uncertainty and heterogeneous interactions of multi-agent driving scenarios remain the biggest challenges currently faced. In order to establish a more compatible representation of multi-agent future behaviors, the current mainstream paradigms are divided into two types, namely dense representation and sparse representation. However, the current dense and sparse future behavior representations are inefficient and inconsistent in multi-agent modeling, resulting in unstable multi-agent behavior patterns during IPP (Integrated Prediction and Planning). Summary of the invention

[0003] Embodiments of the present invention provide a method, device, equipment, medium and product for topological reasoning of future behaviors of multi-agents in autonomous driving, so as to enhance the consistency and accuracy of joint prediction and planning of multi-agent behaviors.

[0004] According to one aspect of the present invention, a method for topological reasoning of future behaviors of multi-agents in autonomous driving is provided, comprising:

[0005] Acquire the position information of the current vehicle and the position information of each agent in the set of agents included in the scene where the current vehicle is located;

[0006] The position information of the current vehicle and the position information of each agent are input into the collaborative learning model to obtain the future behavior topology reasoning result, wherein the collaborative learning model is obtained by iteratively training the initial model through a training sample set, and the training sample set includes: an agent set sample and a future behavior topology reasoning result sample corresponding to the agent set sample.

[0007] According to another aspect of the present invention, a device for topological reasoning of future behaviors of multi-agents in autonomous driving is provided, the device comprising:

[0008] An acquisition module, used to acquire the position information of the current vehicle and the position information of each agent in the set of agents included in the scene where the current vehicle is located;

[0009] An input module is used to input the position information of the current vehicle and the position information of each intelligent agent into a collaborative learning model to obtain a future behavior topology reasoning result, wherein the collaborative learning model is obtained by iteratively training an initial model with a training sample set, and the training sample set includes: an intelligent agent set sample and a future behavior topology reasoning result sample corresponding to the intelligent agent set sample.

[0010] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0011] at least one processor; and

[0012] a memory communicatively connected to the at least one processor; wherein,

[0013] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for future behavior topology reasoning of autonomous driving multi-agents as described in any embodiment of the present invention.

[0014] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method for topological reasoning of future behaviors of multi-agents in autonomous driving as described in any embodiment of the present invention when executed.

[0015] According to another aspect of the present invention, an embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the method for topological reasoning of future behaviors of multi-agents in autonomous driving as described in any embodiment of the present invention.

[0016] The embodiment of the present invention obtains the samples of the agent set and the samples of the future behavior topological reasoning results corresponding to the agent set samples to form a training sample set, iteratively trains the initial model, obtains the collaborative learning model, obtains the position information of the current vehicle and the position information of each agent in the agent set included in the scene where the current vehicle is located, inputs the position information of the current vehicle and the position information of each agent into the collaborative learning model, and obtains the future behavior topological reasoning results. Through the technical solution of the present invention, the consistency and accuracy of joint prediction and planning of multi-agent behavior can be enhanced.

[0017] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 is a flow chart of a method for topological reasoning of future behaviors of multi-agents in autonomous driving according to an embodiment of the present invention;

[0020] Figure 2 is a schematic diagram of the structure of a device for topological reasoning of future behaviors of multi-agents in an autonomous driving according to an embodiment of the present invention;

[0021] Figure 3 It is a schematic diagram of the structure of an electronic device that implements the method for topological reasoning of future behaviors of multi-agents in autonomous driving according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and their meanings are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0024] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0025] Embodiment 1

[0026] Figure 1 is a flow chart of a method for topological reasoning of future behaviors of multi-agents in autonomous driving according to an embodiment of the present invention. This embodiment is applicable to the case of topological reasoning of future behaviors of multi-agents in autonomous driving. The method can be executed by a device for topological reasoning of future behaviors of multi-agents in autonomous driving according to an embodiment of the present invention. The device can be implemented in software and / or hardware, such as Figure 1 As shown, the method specifically comprises the following steps:

[0027] S101, obtaining the position information of the current vehicle and the position information of each agent in the set of agents included in the scene where the current vehicle is located.

[0028] In this embodiment, the current vehicle may be a self-driving car, i.e., a vehicle equipped with an automatic driving system. Exemplarily, the scene where the current vehicle is located may be a driving scene that the current vehicle passes through during driving, such as a crossroads. It should be noted that the agent set may be a set of other motor vehicles, non-motor vehicles, and pedestrians other than the current vehicle included in the driving scene that the current vehicle passes through during driving, wherein each other motor vehicle, non-motor vehicle, or pedestrian is regarded as an agent.

[0029] Specifically, for example, in a driving scene composed of multiple intelligent agents, the position information of the self-driving car and the position information of other intelligent agents except the self-driving car are obtained.

[0030] S102, input the current vehicle's location information and each agent's location information into the collaborative learning model to obtain future behavior topology reasoning results.

[0031] In this embodiment, the collaborative learning model can be a model for predicting IPP (Integrated Prediction and Planning) goals, reasoning about behavioral topology, and generating future trajectory clusters. Preferably, in this embodiment, the collaborative learning model can be a Transformer-based collaborative learning model framework BeTopNet (BeTop, Behavioral Topology, behavioral topology, the core of BeTop is a topological supervision that explicitly characterizes the consensus of future interactions among multiple agents, and establishes reasoning tasks to guide autonomous driving prediction and planning).

[0032] The collaborative learning model is obtained by iteratively training the initial model through a training sample set, and the training sample set includes: agent set samples and future behavior topology reasoning result samples corresponding to the agent set samples.

[0033] Among them, the initial model can be an untrained Transformer-based collaborative learning model framework BeTopNet, the agent set sample can be a set of agents in a driving scene composed of multiple agents collected in advance, and the future behavior topology reasoning result sample corresponding to the agent set sample can be the future behavior topology reasoning result of each agent corresponding to the predetermined agent set sample. This information is collected in advance as a training sample set to train the initial model to obtain a collaborative learning model.

[0034] Specifically, the agent set samples and the future behavior topology reasoning result samples corresponding to the agent set samples are obtained to form a training sample set, and the initial model is iteratively trained to obtain a collaborative learning model. After the model training is completed, the current vehicle position information and the position information of each agent are input into the trained collaborative learning model for predictive reasoning to obtain the future behavior topology reasoning result.

[0035] The embodiment of the present invention obtains the samples of the agent set and the samples of the future behavior topological reasoning results corresponding to the agent set samples to form a training sample set, iteratively trains the initial model, obtains the collaborative learning model, obtains the position information of the current vehicle and the position information of each agent in the agent set included in the scene where the current vehicle is located, inputs the position information of the current vehicle and the position information of each agent into the collaborative learning model, and obtains the future behavior topological reasoning results. Through the technical solution of the present invention, the consistency and accuracy of joint prediction and planning of multi-agent behavior can be enhanced.

[0036] Optionally, iteratively train the initial model using the training sample set, including:

[0037] Build an initial model.

[0038] The agent set samples in the training sample set are input into the initial model to obtain the predicted future behavior topological reasoning results corresponding to the agent set samples.

[0039] Among them, the predicted future behavior topology reasoning result can be the future behavior topology reasoning result obtained after the initial model performs predictive reasoning based on the agent set sample.

[0040] Specifically, the agent set samples in the training sample set are input into the initial model for predictive reasoning to obtain the predicted future behavior topological reasoning results corresponding to the agent set samples.

[0041] Based on the preset fitness function, the fitness value is determined according to the predicted future behavior topological reasoning results corresponding to the agent set samples.

[0042] In this embodiment, the preset fitness function may be a function derived and defined by the user according to the multi-agent behavior topology reasoning task, wherein the fitness value may be a value obtained by calculating the sample of the agent set according to the preset fitness function.

[0043] If the fitness value does not meet the preset target, the parameters of the initial model are trained according to the loss function formed by the predicted future behavior topology reasoning results corresponding to the agent set samples and the future behavior topology reasoning result samples corresponding to the agent set samples.

[0044] Preferably, in this embodiment, the preset goal may be to obtain the maximum fitness value.

[0045] Specifically, during the iterative training of the model, if the fitness value does not reach the preset target, for example, the fitness value is maximum to a certain extent, the weights and other parameters of the initial model are trained according to the loss function formed by the predicted future behavior topological reasoning results corresponding to the agent set samples and the future behavior topological reasoning result samples corresponding to the agent set samples.

[0046] Return to execute the operation of inputting the agent set samples in the training sample set into the initial model to obtain the predicted future behavior topological reasoning results corresponding to the agent set samples, until the fitness value meets the preset target, or the number of iterations is greater than or equal to the preset number, and the collaborative learning model is obtained.

[0047] The preset number of times may be a number threshold pre-set by the user according to actual conditions, and this embodiment does not limit this.

[0048] Specifically, the training is iterated continuously until the obtained fitness value is maximized, or the number of iterations reaches a preset number, then the parameter corresponding to the maximum fitness value is determined as the weight of the collaborative learning model, and finally a trained collaborative learning model is obtained.

[0049] Optionally, the agent set samples in the training sample set are input into the initial model to obtain the predicted future behavior topology reasoning results corresponding to the agent set samples, including:

[0050] Input the agent set samples in the training sample set into the initial model.

[0051] In actual operation, the agent set samples may include self-driving car samples and multi-agent samples such as other motor vehicles, non-motor vehicles or pedestrians.

[0052] Get the position information of each agent sample in the agent set sample.

[0053] Specifically, the position information of the self-driving car sample and the position information of multi-agent samples such as other motor vehicles, non-motor vehicles or pedestrians in the agent set sample are obtained.

[0054] A topological map is constructed based on the location information of each agent sample.

[0055] In this embodiment, the topological map may be a behavioral topological map used to represent the interactive behavioral relationships among multiple agents.

[0056] The topological map is input into the encoder-decoder in the initial model to obtain the predicted future behavior topological reasoning results corresponding to the agent set samples.

[0057] In this embodiment, a Transformer-based collaborative learning model framework BeTopNet is introduced for the joint prediction planning (IPP) goal. Among them, the collaborative decoder (synergisticdecoder) uses an iterative decoding module strategy to collaboratively infer the behavior topology and generate future trajectory clusters. Among them, the decoding module is the codec.

[0058] Specifically, after the topology map is constructed, the topology map is input into the encoder-decoder inference behavior topology in the initial model to obtain the predicted future behavior topology reasoning result corresponding to the agent set sample.

[0059] Optionally, a topology map is constructed based on the location information of each agent sample, including:

[0060] Based on the topological knot theory, each agent sample is regarded as a knot, and a knot interweaving set is generated according to the position information of each agent sample.

[0061] It should be noted that the knot interweaving set can be a knot group formed by all knots after the agent combines each agent sample in the sample as a knot.

[0062] The behavioral topology BeTop originates from the topological knot theory, which infers consistent interaction patterns in multiple trajectories under knot space transformation from intertwine. This enables BeTop to intuitively extract the forward intertwine (occupancy) as a joint topology from the future trajectories of multiple agents abstracted as knot trajectories, combining dense and sparse representations. Specifically, based on the topological knot theory, each agent sample in the agent set sample is regarded as a knot, and the position information of each knot is determined according to the position information of each agent sample, and the knot intertwining set is generated from all knots and their corresponding position information.

[0063] The topological edge set is determined based on the knot interweaving set.

[0064] In this embodiment, every two knots can form a topological edge, and the agent set sample includes multiple agents, that is, there are multiple knots, that is, there are multiple topological edges, wherein the topological edge set can be a set consisting of all topological edges.

[0065] Among them, each topological edge represents the future interaction behavior between the two agents corresponding to the topological edge.

[0066] Specifically, every two knots in the knot interweaving set form a topological edge, and finally form a topological edge set.

[0067] A topological graph is constructed based on each knot and each topological edge.

[0068] Specifically, all knots are regarded as topological nodes, and all topological nodes and topological edges form a topological graph.

[0069] Optionally, the method further includes:

[0070] Get the knot function corresponding to each knot.

[0071] In the actual operation process, knot theory is used to consider a knot group composed of multiple knots, where each knot represents a set of monotonically increasing functions, namely knot functions.

[0072] The edge topology function of the topological edge corresponding to every two knots is determined according to every two knot functions.

[0073] In actual operation, each topological edge element can be defined by two knot functions, that is, the edge topological function of the topological edge composed of the two knots can be represented by the two knot functions.

[0074] The preset fitness function is determined according to the knot function corresponding to each knot and the edge topology function of each topological edge.

[0075] Exemplarily, the preset fitness function may be to obtain the maximum knot function and the maximum edge topology function.

[0076] Optionally, the topology map is input into the encoder / decoder in the initial model to obtain the predicted future behavior topology reasoning result corresponding to the agent set sample, including:

[0077] The topological map is input into the encoder in the initial model to obtain the encoded scene features.

[0078] In this embodiment, the encoder may be, for example, a driving scene encoder, which is used to encode the acquired position information of the agent into scene features of the driving scene, wherein the encoded scene features may be the encoded driving scene features output by the encoder.

[0079] The encoded scene features are input into the decoder in the initial model to obtain the predicted future behavior topology reasoning results corresponding to the agent set samples.

[0080] In this embodiment, the decoder can collaboratively infer the behavior topology based on the encoded scene features generated by the encoder and generate future trajectory clusters.

[0081] As a detailed description of an embodiment of the present invention, the following is a detailed description of a method for topological reasoning of future behaviors of multi-agents in autonomous driving:

[0082] Autonomous driving systems aim to achieve safe and socially compliant driving through behavioral integration between interactive agents. However, uncertainty and heterogeneous interactions in multi-agent driving scenarios remain the biggest challenges currently faced. Current dense and sparse representations of future behaviors in multi-agent modeling are inefficient and inconsistent, resulting in unstable multi-agent behavior patterns when jointly predicting and planning IPP. To address this issue, this embodiment proposes a topological behavior representation paradigm as a compliant behavior prospect to guide downstream trajectory generation. Specifically, this embodiment introduces a behavioral topology (BeTop) that explicitly characterizes the consensus behavior patterns between multi-agent futures. The behavioral topology is derived from the Braid theory and is used to extract compliant interactive topologies from multi-agent future trajectories. The collaborative learning framework (BeTopNet) supervised by the behavioral topology promotes the consistency of behavior prediction and planning in the prediction topology prior. By imitating the emergency learning paradigm, the behavioral topology (BeTop) can effectively handle the behavioral uncertainty of prediction and planning.

[0083] Multi-agent behavior representation: Shaping the collective future behavior of different agents is crucial to achieving socially consistent driving operations. Dense representations are centered on occupancy prediction. Predicting spatial existence under dense BEV (Bird's Eye View) representation provides flexibility and consistency with perception for arbitrary agents. However, fixed resolution and receptive field lead to scene occlusion, resulting in intractable occupancy. At the same time, sparse representations aggregate multi-agent future behaviors into clusters or intentions of future trajectories with multi-modality. The representation of joint future behavior predictions is obtained by recombining from target-based sampling or marginal predictions. However, the collection of joint modal representations is susceptible to modal collapse and brings exponential complexity. Topological representations have gained attention as quantitative indicators of behavioral priors or scene interactions, but topological properties describing collective future behaviors remain unexplored. The behavioral topology (BeTop) of this embodiment targets this problem by combining topologically dense behavior probabilities with sparse motions of joint predictions to present structured future behavior representations. On the other hand, inconsistent multi-agent interaction behaviors inspire the use of future interactions. Implicit methods use attention mechanisms or GNNs (Graph Neural Networks) to obtain implicit behavior interaction information from future trajectory regression learning. However, implicit supervision has unstable learning efficiency in dynamic scenarios. In contrast, explicitly reasoning about mutual behaviors through conditional probability decomposition, relational reasoning, or entropy-based methods can provide consistent behavior priors. However, the huge differences in multi-agent dynamics and driving scene geometry can lead to unstable reasoning. In contrast, behavioral topology designs a compact topological supervision paradigm that can stabilize the consistent consensus interaction between multi-agent future behaviors. It also provides topologically equivalent behavior representations to guide joint prediction and planning.

[0084] Joint Prediction and Planning: Joint Prediction Planning (IPP) systems aim to coordinate the learning of future interactive trajectories between the ego vehicle and multiple agents in the driving scenario. Rule-based approaches integrate manually defined future interactions to evaluate candidate plans, providing remarkable results in rule-driven closed-loop experiments. Nevertheless, the lack of real driving behavior makes rule-based paradigms show a huge gap in real interaction scenarios. Learning-based approaches generate plans that mimic human driving by integrating prediction into the overall model modeling. However, modeling interactions based on historical inputs poses challenges for future consistency among multiple agents. Recently, hybrid paradigms have utilized post-processing and optimization of learning-based models to achieve behavioral interaction between prediction and planning. However, this introduces a large computational overhead, and imitation planning often overestimates the uncertainty in behavior prediction. Work based on tree and contingency planning seeks to balance the active and passive nature of planning behavior in the face of behavioral uncertainty. Nevertheless, paradigms that do not model overall interactions fall into passive planning strategies and lead to exponentially increasing prediction costs. Behavioral Topology (BeTop) provides explicit priors for future interactive behaviors, thereby enhancing the generation of compliant trajectories. Furthermore, the collaborative prediction and contingency planning model based on behavioral topological specifications effectively manages behavioral uncertainty.

[0085] Autonomous driving systems pursue safe, human-like, and compliant driving strategies that are consistent with the driving environment. This drives the representation design of multi-agent future behaviors, future prediction, and multi-agent strategy negotiation between interactive agents and autonomous vehicles (AVs). Learning-based paradigms achieve excellent accuracy, including end-to-end modular design, scene interaction modeling, and motion prediction planning fusion. However, due to the uncertainty of the scene and the unstable interaction pattern of multi-agent future behaviors, there are huge challenges for self-driving systems in actual driving scenarios.

[0086] In order to establish a more compatible representation of multi-agent future behavior, the current mainstream paradigms are divided into two types, namely dense representation and sparse representation. The dense representation paradigm rasterizes the multi-agent behavior centered on self-driving cars and predicts bird's-eye view (BEV) occupancy probability or time flow. It is easy to infer interactions, has strong scalability, can predict the behavior of any agent, and is consistent with BEV perception. However, dense representation is limited by fixed receptive fields and resolution. This greatly affects the robust safety of prediction and makes it difficult to handle occlusions that may interact with driving strategies. In contrast to dense pixel-by-pixel behavior probability modeling, the sparse representation paradigm predicts trajectory clusters or intention distributions anchored by multi-agents. It provides a multimodal representation of each agent, giving it flexible modeling of various behavioral uncertainties under spatial semantics. However, non-aligned behavior conflicts and modal collapse hinder its generation of compliant multi-agent modeling, and the computational overhead grows exponentially with the number of agents. These problems especially lead to unstable and slow behavior paradigm learning when behavior modeling involves joint predictive planning (IPP). Typical solutions through conditional prediction or game-theoretic reasoning often lead to non-strategic planning, as these methods have difficulty ensuring compliance and stability of mutual behavior paradigms in multiple rounds of interactions. This requires a remodeling of the agent's future behavior patterns that is compliant and in line with the IPP optimization objective.

[0087] In the natural decision-making process of humans, human drivers mainly determine the future interactive behaviors of the surrounding rather than the specific states of the surrounding agents to make decisions. Therefore, an effective strategy includes evaluating the impact of the behavior of the surrounding agents on the planning strategy and inferring compliant future interactions. This introduces the core concept of an embodiment of the present invention, that is, multi-agent behaviors that conform to social norms and are consistent exhibit topological structures, which can be identified by extracting consistent or consensus interactions from future behaviors. Previous studies have addressed this challenge using graph neural networks (GNNs) or Transformers through structural design or implicit relationship learning. Other studies have quantified the uncertainty of driving scenarios through topological properties. However, there is currently no method to explicitly formulate a collaborative and consistent supervised representation of the future behavior patterns of agents.

[0088] This embodiment introduces a multi-agent future behavior representation called behavioral topology (BeTop). The core of BeTop is to explicitly characterize a topological supervision with consensus on the future interaction of multi-agents, and to establish reasoning tasks to guide autonomous driving prediction and planning. BeTop originates from the topological knot theory, which infers the consistent interaction paradigm in multiple trajectories under the transformation of knot space from intertwine. This enables BeTop to intuitively extract the forward intertwine (occupancy) as a joint topology from the future trajectories of multi-agents abstracted as knot trajectories, combining dense and sparse representations. With BeTop, this embodiment introduces a Transformer-based collaborative learning model framework BeTopNet for joint prediction planning (IPP) objectives. Among them, the synergistic decoder collaboratively infers the behavioral topology through an iterative decoding module strategy and generates future trajectory clusters. At the same time, the topology-guided local attention module embedded in each decoder layer adaptively uses the reasoned BeTop topological prior to query the agent behavior semantics of future potential interactions and perform information aggregation. In order to alleviate the uncertainty of multi-agents in driving scenarios through the topological prior of BeTop, this embodiment designs and deploys a contingency planning paradigm to optimize the model learning. This embodiment designs a learning paradigm that imitates contingency planning (imitative contingency learning), which standardizes short-term planning with safety guarantees, while solving the uncertainty of long-term planning through joint prediction of BeTop reasoning. Experimental results show that in large-scale public data scenarios, the consistency and accuracy of predictions and planning are enhanced, and state-of-the-art performance is achieved. Simulation tests conducted in the proposed interactive scenario benchmark further highlight the planning capabilities of BeTopNet.

[0089] The technical route of the embodiment of the present invention mainly consists of three parts: first, based on the topological knot theory, a behavioral topology (BeTop) representation is established, and the corresponding topological reasoning task is constructed; then, based on the topological reasoning task, a collaborative learning model BeTopNet and specific module implementation are established; finally, a learning paradigm that imitates emergency planning is introduced to optimize the learning of reasoning tasks.

[0090] First, the behavioral topology representation is constructed:

[0091] Problem construction: Consider the current time t = 0 with N a There are multiple driving scenarios consisting of multiple agents, each of which is represented by Consider the scene graph M at the same time. The corresponding agent is in the historical range T h The states of are represented as X1 of the self-driving car (AV) and X2 of other scene agents. Among them, the future range T f The goal of planning Y1 is n∈[1,N a ] Integrated prediction and planning aims to jointly predict the trajectory of the agent and the vehicle in the scenario

[0092] Topological construction: Using knot theory, the theory is based on future trajectories. Intuitively, it represents a transformation process, According to the local transformation of each agent, the forward interleaving of each future trajectory is collected to represent the future interaction. Formally, consider the N a The original knot σ n The knot group Each of these Represents a set of monotonically increasing functions Represents the Cartesian coordinate system Mapping to the Future of Intelligent Agents The lateral coordinate system Specifically, σ n Functions in Defined as 1≤i,n≤N a , where b n and R n Represents agent A n The left-handed transformation matrix of the local coordinates. In this way, the multi-agent joint interaction behavior is identified as the knot interweaving of one trajectory on another. Different from implicit methods that rely on future trajectory distance heuristics, each intertwining in the knot can be represented as an explicit interactive behavior response, distinguishing active interactions ( Cause other agents to yield) and passive interaction ( To avoid the difficult dynamic knot set reasoning, we restructure the interweaving of multi-agents from the perspective of topological reasoning.

[0093] The goal of BeTop is to infer the topology of the future behavior of multiple agents. Specifically, the node topology Construct the future trajectory of multiple intelligent agents. Reformulated as edge topology of future interaction behavior 1≤i,j≤N aEach topological edge element e ij Both can be achieved through two knot functions To define and evaluate future Y i ,Y j The intertwined relationships of behavior: Here I(.) represents the line segment intersection judgment function under the horizontal coordinate. The reasoning task of behavior topology is further defined as:

[0094]

[0095] in, Represents a topological graph, Represents the predicted node item, that is, the predicted knot, that is, the location information of each agent sample, Represents the predicted edge set, which includes several topological edges e ij .

[0096] Node Item The future state of multi-agents in The optimization is defined by a Gaussian mixture (GMM). Edge topology Framed as a probabilistic reasoning problem:

[0097]

[0098] Among them, e ij Represents edge topology.

[0099] Next, establish a collaborative topological reasoning learning model:

[0100] This example introduces a collaborative learning framework, BeTopNet. Its main body is an encoder-decoder deep network based on Transformer as the backbone. By encoding the scene semantics X;M, the proposed network can achieve simultaneous reasoning and guidance of BeTop using the designed collaborative decoder module. The inference output header and The output of the Joint Prediction Planning (IPP) head constitutes the behavioral topology graph

[0101] Driving scene encoder: This embodiment follows the principle of planning-centricity and uses a scene-centric coordinate system to construct the model input. The scene feature attributes include the historical state X of the multi-agent and the map vector input M. In this embodiment, a length of L is segmented from the complete scene map. m Length N m Map segment. The two types of input attributes are encoded as S A ,S M , and concatenate them into scene features. S=[S A ; SM ]The Transformer encoder stack with local attention is directly used to encode scene semantics S A ,S M Capture the information interaction of historical local areas.

[0102] Collaborative decoder: Based on the encoded scene feature S A ,S M , this embodiment focuses on the decoding strategy. This strategy requires: 1) to infer the same edge topology of BeTop nodes in an interactive manner; 2) to adaptively aggregate the agent features of future interactions using the inferred edge topology prior. To this end, this embodiment follows an iterative decoding strategy consisting of N Transformer decoder layers for all agents. In order to eliminate the uncertainty of the scene, a set of multimodal M decoding queries are initialized for the future trajectories of multiple agents At the same time, the relative relationship feature S R Deployed as topological features via multi-layer perceptron (MLP) Next, this embodiment designs a dual information flow for future trajectories. and future topologies The iterative decoding process of agent A n The decoding process at layer l is as follows:

[0103]

[0104]

[0105] Among them, the future trajectory in BeTop and interactive topology ij ∈ε implements decoding in a collaborative way. Reasoning edge topology The topology decoder is queried by Obtain; This embodiment further As a priori, a Transformer decoder with topology-guided local attention mechanism is built with inference output nodes.

[0106] Topology-guided local attention: Querying the semantic features of all agents in the entire scene will lead to inconsistency with future interactions and sparse attention distribution. This motivates this embodiment to design local attention to infer edge topology. Specifically, this embodiment retrieves the edge topology from the reasoning The first K indices of the prior For the final n The behavior of the interaction. The index with the interaction Directly used to adaptively aggregate S ATo perform local cross attention modeling. This process is formed as follows:

[0107]

[0108] in, Then each layer of decoder has interactive agent feature aggregation represented as

[0109] Inference output head: given each layer’s decoded features This embodiment adds inference output heads, which correspond to the edge and node topology outputs of the behavior topology graph. The topology output head, planning output head, and prediction output head (IPP) are output simultaneously by the stacked MLP when inferring the BeTop result. n , the inference head decodes the future trajectory state The components modeled by generating Gaussian mixture distribution (GMM) (including μ x ,μ y ,σ x ,σ y ,ρ), where the mixed distribution fraction And generate edge topology

[0110] Finally, simulate emergency planning learning:

[0111] BeTopNet learns to mimic the end-to-end goals of human multi-agent behavior and integrates scenario uncertainty compliance behaviors through contingency planning.

[0112] The imitation goal is first established by regulating the behavior state of multiple agents while maximizing their interaction distribution The imitation target is the negative log-likelihood (NLL) of the best inference component m* that is closest to the true value; the behavior distribution of the edge topology is calculated by the binary cross entropy (BCE). In order to integrate uncertainty in multi-agent scenarios By combining the prediction with the immediate safety planning τ M Connect to branch planning trajectory cluster It can postpone uncertain decisions and ensure the safety of actual planning. Direct joint prediction may lose diversity, while the inferred behavior topology can serve as a suitable medium to refine future interactive agents and achieve effective joint combination. Given the imitative self-driving car planning output And the planning branch time is t b ∈(1,T f), the learning of joint contingency planning first uses all marginal predictions Guide safe short-term planning And by joint prediction Guided M branch planning set It is defined as:

[0113]

[0114] Among them, maxC M represents the worst-case cost function (joint prediction With scenario probability By K M Subset recombination of marginal predictions of interacting multi-agents, fast indexing from sorted AV topology It is composed of a joint cost function C that guides the branch planning strategy J Specifically, both types of cost functions are defined by repelling potential fields, which reduce the proximity of plans to their respective prediction paradigms.

[0115] The technical solution implemented by the present invention introduces the concept of behavioral topology (BeTop), which is a multi-agent future behavior paradigm for topological reasoning. It explicitly supervises the consistent future collaborative interaction behavior of the joint prediction planning (IPP) system. In addition, a collaborative learning framework BeTopNet is designed to provide joint planning and prediction guided by topological reasoning. Topologically guided local attention and simulated emergency planning can solve scenario compliance and multi-agent uncertainty. This solution has achieved the best performance in both planning strategy and prediction accuracy.

[0116] Currently, there is no method to explicitly formulate a collaborative supervisory representation for the future behavior patterns of intelligent agents. In this embodiment, a supervisory representation of a topological form with consensus on the future interaction of multiple intelligent agents is explicitly established, namely the proposed behavior topology. The technical solution of this embodiment, by designing the model BeTopNet, realizes the collaborative reasoning of the behavior topology graph, and at the same time uses the interaction edge node reasoning prior to enhance the state generation of predictive planning, and uses the imitation emergency planning learning paradigm to organically integrate the topological reasoning results to solve uncertainty.

[0117] Embodiment 2

[0118] Figure 2is a schematic diagram of a structure of a multi-agent future behavior topology reasoning device for autonomous driving in an embodiment of the present invention. This embodiment is applicable to the case of multi-agent future behavior topology reasoning for autonomous driving. The device can be implemented in software and / or hardware. The device can be integrated in any device that provides the function of multi-agent future behavior topology reasoning for autonomous driving, such as Figure 2 As shown, the autonomous driving multi-agent future behavior topology reasoning device specifically includes: an acquisition module 201 and an input module 202.

[0119] The acquisition module 201 is used to acquire the position information of the current vehicle and the position information of each agent in the agent set included in the scene where the current vehicle is located;

[0120] The input module 202 is used to input the position information of the current vehicle and the position information of each intelligent agent into the collaborative learning model to obtain the future behavior topology reasoning result, wherein the collaborative learning model is obtained by iteratively training the initial model through the training sample set, and the training sample set includes: the intelligent agent set sample and the future behavior topology reasoning result sample corresponding to the intelligent agent set sample.

[0121] Optionally, the input module 202 includes:

[0122] Establishing unit, used to establish initial model;

[0123] A first input unit is used to input the agent set samples in the training sample set into the initial model to obtain the predicted future behavior topological reasoning results corresponding to the agent set samples;

[0124] A first determination unit, configured to determine a fitness value according to a predicted future behavior topology reasoning result corresponding to the agent set sample based on a preset fitness function;

[0125] A training unit, configured to train the parameters of the initial model according to a loss function formed by the predicted future behavior topology reasoning result corresponding to the agent set sample and the future behavior topology reasoning result sample corresponding to the agent set sample if the fitness value does not meet the preset target;

[0126] The execution unit is used to return to execute the operation of inputting the agent set samples in the training sample set into the initial model to obtain the predicted future behavior topological reasoning results corresponding to the agent set samples, until the fitness value meets the preset target, or the number of iterations is greater than or equal to the preset number, to obtain a collaborative learning model.

[0127] Optionally, the input module 202 includes:

[0128] A second input unit, used to input the agent set samples in the training sample set into the initial model;

[0129] A first acquisition unit, configured to acquire location information of each agent sample in the agent set sample;

[0130] A construction unit, used for constructing a topological map according to the location information of each agent sample;

[0131] The third input unit is used to input the topological map into the codec in the initial model to obtain the predicted future behavior topological reasoning result corresponding to the agent set sample.

[0132] Optionally, the construction unit is specifically used for:

[0133] Based on the topological knot theory, each agent sample is regarded as a knot, and a knot interweaving set is generated according to the position information of each agent sample;

[0134] Determine a topological edge set according to the knot interweaving set, wherein each topological edge represents a future interaction behavior between two agents corresponding to the topological edge;

[0135] A topological graph is constructed based on each knot and each topological edge.

[0136] Optionally, the device further comprises:

[0137] A second acquisition unit is used to acquire a knot function corresponding to each knot;

[0138] A second determining unit is used to determine the edge topology function of the topological edge corresponding to every two knots according to every two knot functions;

[0139] The third determination unit is used to determine the preset fitness function according to the knot function corresponding to each knot and the edge topology function of each topological edge.

[0140] Optionally, the third input unit is specifically used for:

[0141] Inputting the topological map into an encoder in the initial model to obtain encoded scene features;

[0142] The encoded scene features are input into the decoder in the initial model to obtain the predicted future behavior topology reasoning results corresponding to the agent set samples.

[0143] The above-mentioned product can execute the future behavior topology reasoning method of autonomous driving multi-agent provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0144] Embodiment 3

[0145] Figure 3 A schematic diagram of the structure of an electronic device 30 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0146] like Figure 3 As shown, the electronic device 30 includes at least one processor 31, and a memory connected to the at least one processor 31, such as a read-only memory (ROM) 32, a random access memory (RAM) 33, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 31 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 32 or the computer program loaded from the storage unit 38 to the random access memory (RAM) 33. In the RAM 33, various programs and data required for the operation of the electronic device 30 can also be stored. The processor 31, the ROM 32, and the RAM 33 are connected to each other through a bus 34. An input / output (I / O) interface 35 is also connected to the bus 34.

[0147] A number of components in the electronic device 30 are connected to the I / O interface 35, including: an input unit 36, such as a keyboard, a mouse, etc.; an output unit 37, such as various types of displays, speakers, etc.; a storage unit 38, such as a disk, an optical disk, etc.; and a communication unit 39, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 39 allows the electronic device 30 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0148] The processor 31 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 31 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The processor 31 executes the various methods and processes described above, such as the future behavior topology reasoning method of autonomous driving multi-agent:

[0149] Acquire the position information of the current vehicle and the position information of each agent in the set of agents included in the scene where the current vehicle is located;

[0150] The position information of the current vehicle and the position information of each agent are input into the collaborative learning model to obtain the future behavior topology reasoning result, wherein the collaborative learning model is obtained by iteratively training the initial model through a training sample set, and the training sample set includes: an agent set sample and a future behavior topology reasoning result sample corresponding to the agent set sample.

[0151] In some embodiments, the autonomous driving multi-agent future behavior topology reasoning method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 38. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 30 via the ROM 32 and / or the communication unit 39. When the computer program is loaded into the RAM 33 and executed by the processor 31, one or more steps of the autonomous driving multi-agent future behavior topology reasoning method described above may be executed. Alternatively, in other embodiments, the processor 31 may be configured to execute the autonomous driving multi-agent future behavior topology reasoning method by any other appropriate means (e.g., by means of firmware).

[0152] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0153] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0154] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0155] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0156] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0157] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0158] In one embodiment, the embodiment of the present invention also includes a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the method for future behavior topology reasoning of autonomous driving multi-agents of any embodiment of the present invention.

[0159] In the process of implementation, the computer program product can be written in one or more programming languages ​​or a combination thereof to perform the computer program code of the present invention, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0160] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0161] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for topological reasoning of future behaviors of multi-agents in autonomous driving, characterized in that: include: Acquire the position information of the current vehicle and the position information of each agent in the set of agents included in the scene where the current vehicle is located; The position information of the current vehicle and the position information of each agent are input into the collaborative learning model to obtain the future behavior topology reasoning result, wherein the collaborative learning model is obtained by iteratively training the initial model through a training sample set, and the training sample set includes: an agent set sample and a future behavior topology reasoning result sample corresponding to the agent set sample.

2. The method according to claim 1, characterized in that The initial model is iteratively trained using the training sample set, including: Build an initial model; Inputting the agent set samples in the training sample set into the initial model to obtain the predicted future behavior topological reasoning results corresponding to the agent set samples; Based on a preset fitness function, determining a fitness value according to a predicted future behavior topological reasoning result corresponding to the agent set sample; If the fitness value does not meet the preset target, the parameters of the initial model are trained according to the loss function formed by the predicted future behavior topology reasoning result corresponding to the agent set sample and the future behavior topology reasoning result sample corresponding to the agent set sample; Return to execute the operation of inputting the agent set samples in the training sample set into the initial model to obtain the predicted future behavior topological reasoning results corresponding to the agent set samples, until the fitness value meets the preset target, or the number of iterations is greater than or equal to the preset number, to obtain a collaborative learning model.

3. The method according to claim 2, characterized in that Inputting the agent set samples in the training sample set into the initial model to obtain the predicted future behavior topology reasoning results corresponding to the agent set samples, including: Inputting the agent set samples in the training sample set into the initial model; Obtaining location information of each agent sample in the agent set sample; Construct a topological map based on the location information of each agent sample; The topological graph is input into the codec in the initial model to obtain the predicted future behavior topological reasoning result corresponding to the agent set sample.

4. The method according to claim 3, characterized in that Construct a topological map based on the location information of each agent sample, including: Based on the topological knot theory, each agent sample is regarded as a knot, and a knot interweaving set is generated according to the position information of each agent sample; Determine a topological edge set according to the knot interweaving set, wherein each topological edge represents a future interaction behavior between two agents corresponding to the topological edge; A topological graph is constructed based on each knot and each topological edge.

5. The method according to claim 2, characterized in that: Also includes: Get the knot function corresponding to each knot; Determine the edge topology function of the topological edge corresponding to each two knots according to each two knot functions; The preset fitness function is determined according to the knot function corresponding to each knot and the edge topology function of each topological edge.

6. The method according to claim 3, characterized in that Inputting the topological graph into the codec in the initial model to obtain the predicted future behavior topological reasoning result corresponding to the agent set sample includes: Inputting the topological map into an encoder in the initial model to obtain encoded scene features; The encoded scene features are input into the decoder in the initial model to obtain the predicted future behavior topology reasoning results corresponding to the agent set samples.

7. An autonomous driving multi-agent future behavior topology reasoning device, characterized in that: include: An acquisition module, used to acquire the position information of the current vehicle and the position information of each agent in the set of agents included in the scene where the current vehicle is located; An input module is used to input the position information of the current vehicle and the position information of each intelligent agent into a collaborative learning model to obtain a future behavior topology reasoning result, wherein the collaborative learning model is obtained by iteratively training an initial model with a training sample set, and the training sample set includes: an intelligent agent set sample and a future behavior topology reasoning result sample corresponding to the intelligent agent set sample.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the autonomous driving multi-agent future behavior topology reasoning method described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, which are used to enable a processor to implement the autonomous driving multi-agent future behavior topology reasoning method according to any one of claims 1 to 6 when executed.

10. A computer program product, comprising a computer program, which, when executed by a processor, implements the autonomous driving multi-agent future behavior topology reasoning method according to any one of claims 1-6.

Citation Information

Patent Citations

  • Topological reasoning method, system and equipment for driving scene and storage medium

    CN116386009A

  • Multi-agent co-evolution Internet of Things topology robust optimization method

    CN116647459A

  • Multi-agent reinforcement learning power grid energy management method and system

    CN117035007A

  • Road topology detection method and related device

    CN117115766A

  • Intelligent agent driving prediction model construction method, prediction method, equipment and medium

    CN118306428A

Cited By

  • Intelligent driving network topology reconstruction method, system and device and storage medium

    CN120768773A

  • Automatic driving decision-making method based on three-dimensional scene reconstruction and multi-modal large model

    CN120894490A

  • Autonomous driving decision-making method based on three-dimensional scene reconstruction and multi-modal large model

    CN120894490B