Cooperative path game and planning system of unmanned aerial vehicle and autonomous vehicle
By generating a pool of collaborative possibilities through probabilistic trajectory options and collaborative guarantee contracts, and combining distributed negotiation with portfolio management, the problem of rigid decision-making and interest balancing in dynamic environments for drones and autonomous vehicles is solved, achieving efficient collaborative path planning.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHONGQING PULIN TECHNOLOGY CO LTD
- Filing Date
- 2026-04-09
- Publication Date
- 2026-07-03
AI Technical Summary
Existing collaborative path planning methods for drones and autonomous vehicles are rigid in decision-making and lack adaptability in dynamic and uncertain environments. Furthermore, they are difficult to effectively balance the individual goals of each agent with the overall interests of the team in a decentralized architecture.
A collaborative possibility pool generation module is introduced to generate probabilistic trajectory options and collaborative guarantee contracts. Combined with distributed negotiation and portfolio management modules, dynamic adjustments are achieved through real-time monitoring and trajectory execution modules, and trajectory switching is performed using model prediction control methods.
It improves the system's adaptability and robustness in dynamic environments, reduces the risk of communication bottlenecks, realizes fair negotiation and collaborative value distribution among intelligent agents, and enhances the system's decision-making efficiency and collaborative depth.
Smart Images
Figure CN122334494A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of multi-agent cooperative control technology, specifically a cooperative path game and planning system for unmanned aerial vehicles and autonomous vehicles. Background Technology
[0002] With the development of drone and autonomous vehicle technologies, their collaborative application has shown broad prospects in fields such as logistics delivery, environmental monitoring, and emergency rescue. By combining the high-altitude wide-area perspective of drones with the long-endurance ground capabilities of autonomous vehicles, collaborative systems can accomplish complex tasks that are difficult for a single intelligent agent to handle.
[0003] Existing cooperative path planning methods are typically based on a shared environment model, using various optimization algorithms to calculate a set of cooperative trajectories for each agent to achieve a specific task objective. These methods can be broadly categorized into centralized and distributed architectures. Centralized architectures rely on a central node to collect information and make decisions, while distributed architectures depend on local communication and negotiation between agents.
[0004] However, these methods often assume environmental determinism during planning, generating a static, pre-defined cooperative trajectory. When the system operates in a complex and dynamic real-world environment, frequent unforeseen events, such as sudden traffic congestion, temporary communication interruptions, or limited sensor perception, can quickly render the pre-defined trajectory ineffective or suboptimal. In such cases, the system must perform global or local online replanning, a process with enormous computational overhead and high response latency, making it difficult to meet high real-time requirements and thus reducing the adaptability and robustness of the entire cooperative system.
[0005] Furthermore, within a decentralized collaborative framework, effectively coordinating the individual goals of different agents with the overall interests of the team presents an inherent technical challenge. Traditional collaborative mechanisms typically rely on predefined rules or simple information sharing, lacking a flexible mechanism to quantify and negotiate the additional costs and potential benefits generated during collaborative behavior. This makes it difficult for agents to proactively provide opportunistic assistance or risk protection to other members, as such actions may harm their own interests, ultimately limiting the depth of collaboration and the improvement of overall system performance. Summary of the Invention
[0006] To address the shortcomings of existing technologies, this invention provides a collaborative path game and planning system for drones and autonomous vehicles. This system solves the problems of rigid decision-making and insufficient adaptability in dynamic and uncertain environments, as well as the difficulty in effectively balancing the individual goals of each agent and the overall interests of the team under a decentralized architecture.
[0007] To achieve the above objectives, the present invention provides the following technical solution: a cooperative path game and planning system for unmanned aerial vehicles and autonomous vehicles, comprising: The collaborative possibility pool generation module is used to generate a collaborative possibility pool containing probabilistic trajectory options and collaborative guarantee contracts based on preset system information and environment models. The distributed negotiation and portfolio management module is used to execute a decentralized negotiation protocol, enabling the drones and autonomous vehicles to trade probabilistic trajectory options and collaborative guarantee contracts in the collaborative possibility pool based on their respective utility assessments, and to construct their respective asset portfolios. The real-time monitoring and trajectory execution module is deployed on the drone and the autonomous vehicle to monitor the triggering conditions of each asset in the asset portfolio, and dynamically adjust and execute their respective trajectories when the conditions are met.
[0008] Preferably, the probabilistic trajectory option is a type of structured data, including: the trajectory of the underlying asset, the triggering conditions, the execution cost, and the option value.
[0009] Preferably, the collaborative protection contract is a type of structured data, including: protection trajectory, risk conditions, contract premium, and performance costs.
[0010] Preferably, the collaborative possibility pool generation module generates the probabilistic trajectory option or collaborative guarantee contract by injecting a preset disturbance event into the benchmark collaborative trajectory and planning alternative trajectories or guarantee trajectories for the disturbance event.
[0011] Preferably, in the distributed negotiation and portfolio management module, the utility evaluation of the drone and the autonomous vehicle is based on a dynamic utility function, which performs a weighted calculation of the benefits and costs of the probabilistic trajectory option or collaborative guarantee contract according to the current state of the intelligent agent.
[0012] Preferably, the real-time monitoring and trajectory execution module uses a model predictive control method for trajectory tracking, and dynamically adjusts the trajectory by switching the reference trajectory of the model predictive controller from the baseline trajectory to the underlying trajectory of the contingent trajectory option or the guaranteed trajectory of the collaborative guarantee contract when the triggering condition is met.
[0013] A collaborative path game and planning method for drones and autonomous vehicles includes the following steps: Generation steps: Based on the preset system information and environment model, generate a collaborative possibility pool that includes probabilistic trajectory options and collaborative guarantee contracts; Negotiation steps: Execute a decentralized negotiation protocol so that the drone and the autonomous vehicle can trade assets in the collaborative possibility pool based on their respective utility assessments and construct their respective asset portfolios; Execution steps: Monitor the trigger conditions of each asset in the asset portfolio, and dynamically adjust and execute the respective trajectories of the drone and the autonomous vehicle when the conditions are met.
[0014] Preferably, the probabilistic trajectory option is a type of structured data, including: the trajectory of the underlying asset, triggering conditions, execution costs, and option value; the collaborative protection contract is a type of structured data, including: protection trajectory, risk conditions, contract premium, and performance costs.
[0015] Preferably, in the negotiation step, the utility evaluation is based on a dynamic utility function, which performs a weighted calculation of the benefits and costs of the probabilistic trajectory option or collaborative guarantee contract based on the agent's current state.
[0016] Preferably, the execution steps specifically include: using a model predictive control method to track the trajectory, and dynamically adjusting the trajectory by switching the reference trajectory of the model predictive controller from the baseline trajectory to the underlying trajectory of the contingent trajectory option or the guaranteed trajectory of the collaborative guarantee contract when the triggering condition is met.
[0017] This invention provides a cooperative path game and planning system for drones and autonomous vehicles. It has the following beneficial effects: 1. This invention introduces contingent trajectory options and collaborative guarantee contracts to pre-convert collaborative possibilities under future uncertainties into structured solutions. Combined with the event triggering mechanism of the real-time monitoring and trajectory execution module, the system can quickly activate corresponding alternative or guarantee trajectories when the environment changes, thus eliminating the need for time-consuming online replanning and significantly improving the system's adaptability and robustness to dynamic environments.
[0018] 2. The distributed negotiation and portfolio management module employed in this invention enables each agent to make decisions and trade autonomously through a decentralized negotiation protocol. This mechanism avoids dependence on a single central computing or decision-making node, effectively reducing system communication bottlenecks and single-point-of-failure risks, resulting in a highly scalable system architecture that maintains decision-making efficiency even as the number of collaborative agents increases.
[0019] 3. The dynamic utility function introduced in this invention provides a quantitative basis for each agent to evaluate the costs and benefits of collaborative behavior based on its own state. Combined with a negotiation model based on asset transactions, it allows agents to fairly negotiate and allocate collaborative value, thereby guiding system decisions towards a better overall direction while ensuring the rationality of each agent. This effectively solves the technical problem of coordinating individual and group goals in decentralized systems. Attached Figure Description
[0020] Figure 1 This is a system architecture diagram of the present invention. Detailed Implementation
[0021] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0022] Please see the appendix Figure 1 This invention provides a cooperative path game and planning system for drones and autonomous vehicles, comprising: The collaborative possibility pool generation module is used to generate a collaborative possibility pool containing probabilistic trajectory options and collaborative guarantee contracts based on preset system information and environment models. The distributed negotiation and portfolio management module is used to execute a decentralized negotiation protocol, enabling drones and autonomous vehicles to trade probabilistic trajectory options and collaborative guarantee contracts in the collaborative possibility pool based on their respective utility assessments, and to construct their respective asset portfolios. The real-time monitoring and trajectory execution module, deployed on drones and autonomous vehicles, is used to monitor the triggering conditions of each asset in the asset portfolio and dynamically adjust and execute their respective trajectories when the conditions are met.
[0023] The collaborative possibility pool generation module generates a collaborative possibility pool containing probabilistic trajectory options and collaborative guarantee contracts based on preset system information and environmental models. This module runs before task execution or in a low-frequency online update mode, providing a decision-making basis for the entire collaborative system.
[0024] A contingent trajectory option (COTO) is a type of structured data that defines the right to execute an alternative trajectory under specific conditions. Its data structure is defined as follows: ; in: This is a unique identifier for the probabilistic trajectory option; The target trajectory is a pre-calculated, dynamically feasible alternative path or sequence of behaviors. The trigger condition is a Boolean function that determines the trigger condition based on the external environment state. and the internal state of the agent The input is a true value, which determines whether the option can be exercised. For execution cost, it is a cost vector that quantifies the trajectory of the execution target. The expected additional resource consumption relative to the baseline trajectory; Option value is a payoff vector that quantifies the trajectory of the underlying asset. Expected benefits; The expiration date is the time at which the option becomes valid.
[0025] A Synergistic Assurance Contract (SAC) is another type of structured data that defines an assurance obligation established to hedge synergistic risks. Its data structure is defined as follows: ; in: This serves as a unique identifier for the collaborative guarantee contract; The guaranteed trajectory is the sequence of protective actions that the insured intelligent agent must perform when the contract is triggered. The risk condition is a Boolean function that determines whether the insured agent has encountered a specific risk based on its state Sinsured. The premium is a cost vector representing the cost that the insured must pay to the insurer to purchase this coverage. For fulfillment costs, it is a cost vector representing the resource consumption that the insurer needs to bear in fulfilling its protection obligations; This refers to the contract's expiration date.
[0026] The distributed negotiation and portfolio management module interacts with the decision-making units of each agent (UAV, AV) to execute a decentralized negotiation protocol, enabling each agent to trade assets in the collaborative possibility pool based on its own utility assessment and to construct its own asset portfolio.
[0027] Each agent i evaluates the asset value through a dynamic utility function Ui. For a COTO, its utility function is: ; in: The current state of agent i; Let be the state-dependent weight vector of agent i for the reward, reflecting its emphasis on different types of rewards; Let be the state-dependent weight vector of agent i in relation to cost, reflecting its sensitivity to different types of cost.
[0028] The real-time monitoring and trajectory execution module, deployed on each agent, is used to frequently monitor the agent's own state and the external environment, and dynamically adjust and execute trajectories based on its asset portfolio. This module continuously monitors the trigger conditions of its COTO and SAC holdings. and risk conditions .
[0029] In the initial stage of system operation, the collaborative possibility pool generation module is executed first to generate an asset pool containing various COTOs and SACs.
[0030] Subsequently, the distributed negotiation and portfolio management module is activated. Each agent evaluates the assets in the pool based on its own dynamic utility function and conducts transactions through a distributed negotiation protocol, ultimately forming its own asset portfolio.
[0031] During task execution, the real-time monitoring and trajectory execution modules on each agent run continuously. By default, the agent tracks a baseline trajectory. Once the module detects that the trigger condition for a certain COTO it holds has been met and the immediate utility of executing that COTO is positive, it switches the reference trajectory of the underlying trajectory controller to the target trajectory of that COTO. If it detects that a risk condition of a SAC for which it is the policyholder is met, it will send a trigger signal to the insurer. Upon receiving the signal, the insurer's real-time monitoring and trajectory execution module will switch the reference trajectory to the insured trajectory of that SAC. This completes a full, event-triggered, dynamic collaborative planning and execution loop.
[0032] This invention also provides a cooperative path game and planning method for drones and autonomous vehicles, comprising the following steps: Generation steps: Based on the preset system information and environment model, generate a collaborative possibility pool that includes probabilistic trajectory options and collaborative guarantee contracts; The generation step is a preprocessing phase performed before task execution or through low-frequency online updates. This step first receives pre-defined system information and an environmental model. The system information includes the dynamics models of the UAV and autonomous vehicle, sensor performance parameters, communication range, and other intrinsic attributes. The environmental model includes digital maps, known static obstacles, and dynamic information based on historical data statistics, such as the probability of traffic congestion on specific road segments or the distribution of communication signal strength in specific areas. Based on these inputs, this step proactively identifies potential future opportunities or risks through scenario analysis or perturbation injection. For example, by analyzing maps and traffic data, it identifies a potentially smooth shortcut under specific conditions and generates a probabilistic trajectory option around this shortcut; or by analyzing the communication model, it identifies signal blind spots that the autonomous vehicle will enter and plans a guaranteed trajectory for the UAV to provide relay communication over that area, thereby generating a cooperative guarantee contract. Finally, this step outputs a cooperative possibility pool containing multiple structured data objects, providing a decision-making basis for subsequent negotiation steps.
[0033] Negotiation steps: Execute a decentralized negotiation protocol that enables drones and autonomous vehicles to trade assets in the collaborative possibility pool based on their respective utility assessments and build their respective asset portfolios; The negotiation step is a decentralized decision-making process executed among intelligent agents. In this step, each intelligent agent, whether a drone or an autonomous vehicle, acts as an independent decision-making unit, accessing assets in a pool of collaborative possibilities. For each probabilistic trajectory option or collaborative guarantee contract in the pool, the agent evaluates its value based on its own dynamic utility function. This evaluation process is not static but closely related to the agent's current state; for example, a drone with low battery power might assign a higher utility evaluation to a probabilistic trajectory option that includes a charging opportunity. After the evaluation, the agents interact through a pre-defined decentralized negotiation protocol, such as a distributed auction or contract network protocol. Under this protocol, agents can bid for desired assets as buyers or set prices for the guarantee services they can provide as sellers. Through a series of offers, bids, and the exchange of winning bid messages, assets are allocated among the agents. The final product of this step is that each agent holds a personalized asset portfolio consisting of probabilistic trajectory options and collaborative guarantee contracts acquired through successful transactions.
[0034] Execution steps: Monitor the trigger conditions of each asset in the asset portfolio, and dynamically adjust and execute the respective trajectories of the drone and the autonomous vehicle when the conditions are met.
[0035] The execution phase is the real-time control stage that operates frequently and cyclically during task execution. By default, each agent travels or flies along a baseline trajectory. Simultaneously, its real-time monitoring and trajectory execution module continuously compares its real-time status and external environmental information acquired from sensors with the trigger conditions of each asset in the asset portfolio. This comparison is a logical judgment process. Once the trigger condition of a contingent trajectory option is detected as being met—for example, a drone confirming through sensors that a shortcut is indeed clear—or the risk condition of a collaborative guarantee contract is met—for example, an autonomous vehicle about to enter a signal blind spot—the module immediately performs a dynamic trajectory adjustment. Specifically, the adjustment seamlessly switches the reference trajectory input of the underlying trajectory tracking controller, such as the model prediction controller, from the current baseline trajectory to the underlying trajectory or guarantee trajectory defined in the triggered asset. This event-driven execution method allows the agent to instantly utilize opportunities or mitigate risks without needing to replan.
[0036] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A cooperative path game and planning system for UAVs and autonomous vehicles, characterized in that, include: The collaborative possibility pool generation module is used to generate a collaborative possibility pool containing probabilistic trajectory options and collaborative guarantee contracts based on preset system information and environment models. The distributed negotiation and portfolio management module is used to execute a decentralized negotiation protocol, enabling the drones and autonomous vehicles to trade probabilistic trajectory options and collaborative guarantee contracts in the collaborative possibility pool based on their respective utility assessments, and to construct their respective asset portfolios. The real-time monitoring and trajectory execution module is deployed on the drones and autonomous vehicles to monitor the triggering conditions of each asset in the asset portfolio, and dynamically adjust and execute their respective trajectories when the conditions are met.
2. The collaborative path game and planning system for drones and autonomous vehicles according to claim 1, characterized in that, The probabilistic trajectory option is a type of structured data, including: the trajectory of the underlying asset, the triggering conditions, the execution cost, and the option value.
3. The cooperative path game and planning system for drones and autonomous vehicles according to claim 1, characterized in that, The collaborative protection contract is a type of structured data, including: protection trajectory, risk conditions, contract premium, and performance costs.
4. The cooperative path game and planning system for drones and autonomous vehicles according to claim 1, characterized in that, The collaborative possibility pool generation module generates the probabilistic trajectory option or collaborative guarantee contract by injecting preset disturbance events into the benchmark collaborative trajectory and planning alternative trajectories or guarantee trajectories for the disturbance events.
5. The collaborative path game and planning system for drones and autonomous vehicles according to claim 1, characterized in that, In the distributed negotiation and portfolio management module, the utility evaluation of the drone and the autonomous vehicle is based on a dynamic utility function. This dynamic utility function calculates the benefits and costs of the probabilistic trajectory option or collaborative guarantee contract in a weighted manner according to the current state of the intelligent agent.
6. The cooperative path game and planning system for drones and autonomous vehicles according to claim 1, characterized in that, The real-time monitoring and trajectory execution module uses a model predictive control method for trajectory tracking, and dynamically adjusts the trajectory by switching the reference trajectory of the model predictive controller from the baseline trajectory to the underlying trajectory of the contingent trajectory option or the guaranteed trajectory of the collaborative guarantee contract when the triggering condition is met.
7. A method for cooperative path game and planning between unmanned aerial vehicles (UAVs) and autonomous vehicles, used in the cooperative path game and planning system for UAVs and autonomous vehicles according to any one of claims 1-6, characterized in that, Includes the following steps: Generation steps: Based on the preset system information and environment model, generate a collaborative possibility pool that includes probabilistic trajectory options and collaborative guarantee contracts; Negotiation steps: Execute a decentralized negotiation protocol so that the drone and the autonomous vehicle can trade assets in the collaborative possibility pool based on their respective utility assessments and construct their respective asset portfolios; Execution steps: Monitor the trigger conditions of each asset in the asset portfolio, and dynamically adjust and execute the respective trajectories of the drone and the autonomous vehicle when the conditions are met.
8. The method for cooperative path game and planning between unmanned aerial vehicles and autonomous vehicles according to claim 7, characterized in that, The probabilistic trajectory option is a type of structured data, including: the trajectory of the underlying asset, triggering conditions, execution costs, and option value; the collaborative protection contract is a type of structured data, including: protection trajectory, risk conditions, contract premium, and performance costs.
9. A method for cooperative path game and planning between unmanned aerial vehicles and autonomous vehicles according to claim 7, characterized in that, In the negotiation step, the utility evaluation is based on a dynamic utility function, which calculates the benefits and costs of the probabilistic trajectory option or cooperative guarantee contract in a weighted manner according to the agent's current state.
10. A method for cooperative path game and planning between unmanned aerial vehicles and autonomous vehicles according to claim 7, characterized in that, The execution steps specifically include: using a model predictive control method to track the trajectory, and dynamically adjusting the trajectory by switching the reference trajectory of the model predictive controller from the baseline trajectory to the underlying trajectory of the contingent trajectory option or the guaranteed trajectory of the collaborative guarantee contract when the triggering condition is met.