Obstacle avoidance navigation and energy scheduling collaborative optimization method and system of hybrid energy supply unmanned system

Abstract

The invention discloses a multi-scale obstacle avoidance navigation and energy scheduling load source collaborative optimization method and system for a hybrid energy supply unmanned system, and belongs to the field of hybrid energy supply unmanned system scheduling optimization. Firstly, an optimal path is searched according to a large-scale detection data reconstruction environment, and dynamic obstacle avoidance navigation is performed in real time according to a dynamic obstacle trajectory in a movement process. The unmanned system performs obstacle detection according to the high-precision environmental data to realize dynamic obstacle avoidance in real time, and when a collision risk is detected, an improved dynamic window method is adopted to perform local adjustment on an optimal path so as to avoid the obstacle. And then feeding back a comprehensive optimal motion strategy according to the energy storage condition, and finally realizing bidirectional online energy scheduling of the load source. And the system predicts a feasible load demand interval and establishes a target function by taking the minimum total equivalent hydrogen consumption as a target, and the optimal output power change rate of each energy supply unit at each prediction moment is solved. The cruising ability of the hybrid energy supply unmanned system can be effectively improved, and safe and efficient movement is achieved.

Description

Technical Field

[0001] The present invention belongs to the field of scheduling optimization of hybrid energy supply unmanned systems, and more specifically, to a method and system for collaborative optimization of obstacle avoidance navigation and energy scheduling of hybrid energy supply unmanned systems. Background Art

[0002] With the increasing attention paid to high-altitude patrols, emergency rescue, ocean monitoring and marine resource exploration by countries around the world, and the increasing demand for military mission reconnaissance and execution, the design and development of technologies that can replace humans in performing various dangerous tasks in complex environments such as underwater and in the air has become the core focus of modern unmanned intelligent system research. Modern exploration and mission execution equipment needs to have the ability to work stably for a long time in extreme environments to support the needs of scientific research, resource management, environmental protection and security. Therefore, the development of a general unmanned system with autonomous long-term endurance can not only reduce human exposure to dangerous environments, but also greatly improve the efficiency, accuracy and responsiveness of missions. Solid oxide fuel cell (SOFC) is a quiet and efficient power generation technology. It is equipped with auxiliary energy such as lithium batteries to form a hybrid power system with strong endurance and rapid adaptation to load requirements. It is very suitable for its multi-scenario application characteristics under complex working conditions and can be widely used in mission scenarios such as underwater unmanned submersibles and high-altitude drones. Unlike urban road environments, unstructured, complex and dynamic environments are changeable and have low map accuracy. It is difficult for autonomous unmanned systems to perform high-precision autonomous navigation and operation like vehicles using high-precision maps and structured road scenes.

[0003] To solve the above problems, energy management strategy has become a powerful tool for in-depth research, which can effectively reduce energy consumption and improve the operating efficiency of the system. However, considering energy management strategy alone is not enough to comprehensively improve the overall performance of fuel cell hybrid autonomous unmanned systems. It is necessary to comprehensively consider multiple factors such as environmental perception, trajectory planning and energy management, and conduct multi-level collaborative optimization to achieve higher mission success rates and longer endurance in complex natural dynamic environments. This multi-dimensional comprehensive optimization method can not only improve energy utilization efficiency, but also enhance the system's autonomous decision-making ability and environmental adaptability, thereby comprehensively improving the overall performance of the unmanned system.

[0004] For fuel cell hybrid unmanned systems, the existing endurance optimization methods are mainly divided into two types: internal energy management and external trajectory planning. The internal energy management method focuses on optimizing the energy distribution between the power sources of the fuel cell hybrid system to improve the energy efficiency of the system and extend the endurance time. Although these methods can optimize energy utilization to a certain extent, they often ignore the dynamic navigation needs of autonomous unmanned systems in complex environments and lack the ability to respond to changes in the external environment in real time, resulting in insufficient energy utilization and limited endurance time in practical applications. On the other hand, the external trajectory planning method mainly focuses on how to plan the optimal path in a complex environment to avoid collisions with obstacles and improve navigation efficiency. These methods can provide more ideal autonomous navigation solutions in static or quasi-static environments through path search and optimization algorithms. However, they usually do not consider the actual energy state of the unmanned system and the characteristics of the power system, and ignore the importance of internal energy management, which may lead to unreasonable energy distribution in long-term missions and interruptions in the middle of mission execution due to energy exhaustion.

[0005] Overall, how to simultaneously improve the navigation accuracy, operational safety and endurance of unmanned systems in complex environments is an urgent problem to be solved. Summary of the invention

[0006] In response to the defects of the prior art and the need for improvement, the present invention provides a method and system for collaborative optimization of obstacle avoidance navigation and energy scheduling of a hybrid-powered unmanned system. Its purpose is to improve the endurance of the hybrid-powered unmanned system and achieve safe and efficient navigation through multi-level system optimization of environmental perception, trajectory planning and energy management.

[0007] To achieve the above object, according to one aspect of the present invention, a three-level optimization method for load-source coordination of a hybrid energy supply unmanned system is provided, comprising:

[0008] Search for the optimal path from the starting point to the target point based on the environmental data collected by large-scale sensors; enable the unmanned system to navigate along the optimal path, and divide the navigation process into multiple prediction time domains, and perform dynamic obstacle avoidance and load-source bidirectional coordinated energy scheduling in each prediction time domain;

[0009] Dynamic obstacle avoidance includes: obstacle detection based on environmental data collected in real time by high-precision sensors; if it is detected that the collision area of ​​an obstacle intersects with the optimal path, the optimal path is partially adjusted to avoid the collision area of ​​each obstacle;

[0010] Energy scheduling includes the following steps:

[0011] (S1) predicting the speed range of the unmanned system at each prediction time in the current prediction time domain according to the current optimal path, and mapping it into a feasible load demand range at the corresponding prediction time, and determining a feasible energy dispatch range corresponding to each feasible load demand range;

[0012] (S2) establishing an objective function with the goal of minimizing the total equivalent hydrogen consumption of the hybrid energy supply system in the current forecast time domain, solving the output power change rate of each energy supply unit in the hybrid energy supply system at each forecast time based on the feasible load demand interval and the corresponding feasible energy scheduling interval, and obtaining the optimal control variable sequence;

[0013] (S3) controlling the output power of each energy supply unit in the hybrid energy supply system at each predicted time according to the optimal control variable sequence;

[0014] Among them, large-scale sensors and high-precision sensors are both carried on unmanned systems, and the large-scale sensors have a wider search range, and the high-precision sensors have higher detection accuracy; the collision area is the area formed by the position where the distance to the obstacle is less than the preset safety distance.

[0015] Furthermore, the optimal path is locally adjusted to avoid the collision area of ​​each obstacle, including:

[0016] (T1) When there is a possibility of collision when continuing to navigate along the optimal path, the current time is taken as the current sampling time;

[0017] (T2) At the current sampling time, sampling is performed in the forward speed sampling space of the unmanned system to obtain multiple sets of candidate speeds; the forward speed sampling space is:

[0018]

[0019] (T3) calculating the position of the unmanned system at the next sampling time according to the position of the unmanned system at the current sampling time and the candidate speed sampling results, and eliminating the position within the collision area;

[0020] (T4) The positions of the current sampling time and the previous sampling times are combined into a current local path, and the optimal local path is selected therefrom;

[0021] (T5) If the real-time trajectory adjustment of the dynamic obstacle is completed, proceed to step (T6); otherwise, proceed to step (T2);

[0022] (T5) using the current optimal local path to adjust the optimal path;

[0023] Where (v, ω) represents a set of candidate velocities, v represents the velocity of the unmanned system, ω represents the angular velocity of the unmanned system; Δt represents the sampling time interval; a maxrepresents the maximum acceleration, Indicates the maximum angular acceleration.

[0024] Further, step (T4) selects the best local path from the current local paths based on the evaluation function;

[0025] The evaluation function is:

[0026]

[0027] Among them, Heading(v,ω) represents the azimuth evaluation function, which is used to evaluate the angular difference between the direction of the trajectory end and the target point at the currently set speed of the unmanned system; Dist(v,ω) represents the distance evaluation function, which is used to evaluate the collision distance between the unmanned system and the obstacle; Velocity(v,ω) represents the speed evaluation function, which is used to evaluate the navigation speed of the unmanned system within the constraint range; α, β, and γ are the coefficients of the azimuth evaluation function Heading(v,ω), the distance evaluation function dist(v,ω) and the speed evaluation function Velocity(v,ω), respectively.

[0028] Furthermore, in step (S2), the objective function is expressed as:

[0029]

[0030] Where J represents the objective function value; P i represents the prediction time domain, i represents the prediction time in the prediction time domain; Indicates that in the prediction domain P i The equivalent hydrogen consumption of the hybrid energy supply system at the i-th prediction moment within .

[0031] Furthermore, the optimal path from the starting point to the target point is searched based on the environmental data collected by the large-scale sensor, including:

[0032] Model the environment based on the environmental data collected by large-scale sensors and determine all paths from the starting point to the target point;

[0033] Each path is a particle in the particle swarm algorithm, and the particle swarm algorithm is used to search for the optimal path from all paths; during the search process, the calculation formula of the particle fitness value is:

[0034] f(p)=ρ·PathLength(p)+σ·ObstacleAvoidance(p)+τ

[0035] EnergyConsumption(p)

[0036] Where p represents a particle, f(p) represents the fitness value of particle p; PathLength(p), ObstacleAvoidance(p) and EnergyConsumption(p) represent the path length, obstacle avoidance capability and energy consumption of the path corresponding to particle p, respectively; ρ, σ and τ represent the coefficients of path length, obstacle avoidance capability and energy consumption, respectively.

[0037] According to another aspect of the present invention, a load-source collaborative three-level optimization system for a hybrid energy supply unmanned system is provided, comprising: a track planning module, a collaborative optimization control module, a dynamic obstacle avoidance module and an energy scheduling module;

[0038] The trajectory planning module is used to search for the optimal path from the starting point to the target point based on the environmental data collected by large-scale sensors;

[0039] Dynamic obstacle avoidance module, used to perform dynamic obstacle avoidance;

[0040] Energy scheduling module, used to perform energy scheduling of load-source bidirectional collaboration;

[0041] A collaborative optimization control module is used to make the unmanned system navigate along the optimal path and divide the navigation process into multiple prediction time domains. In each prediction time domain, the dynamic obstacle avoidance module and the energy scheduling module are used to perform dynamic obstacle avoidance and energy scheduling in turn.

[0042] Dynamic obstacle avoidance includes: obstacle detection based on environmental data collected in real time by high-precision sensors; if it is detected that the collision area of ​​an obstacle intersects with the optimal path, the optimal path is partially adjusted to avoid the collision area of ​​each obstacle;

[0043] Energy scheduling includes the following steps:

[0044] (S1) predicting the speed range of the unmanned system at each prediction time in the current prediction time domain according to the current optimal path, and mapping it into a feasible load demand range at the corresponding prediction time, and determining a feasible energy dispatch range corresponding to each feasible load demand range;

[0045] (S2) establishing an objective function with the goal of minimizing the total equivalent hydrogen consumption of the hybrid energy supply system in the current forecast time domain, solving the output power change rate of each energy supply unit in the hybrid energy supply system at each forecast time based on the feasible load demand interval and the corresponding feasible energy scheduling interval, and obtaining the optimal control variable sequence;

[0046] (S3) controlling the output power of each energy supply unit in the hybrid energy supply system at each predicted time according to the optimal control variable sequence;

[0047] Among them, large-scale sensors and high-precision sensors are both carried on unmanned systems, and the large-scale sensors have a wider search range, and the high-precision sensors have higher detection accuracy; the collision area is the area formed by the position where the distance to the obstacle is less than the preset safety distance.

[0048] According to another aspect of the present invention, an autonomous unmanned system is provided, including a hybrid energy supply unmanned system and a load-source coordination three-level optimization system of the hybrid energy supply unmanned system provided by the present invention;

[0049] Unmanned systems are equipped with large-scale sensors and high-precision sensors; among these sensors, large-scale sensors have a wider search range, and high-precision sensors have higher detection accuracy.

[0050] Furthermore, the large-scale sensor is sonar, and the high-precision sensor is a camera.

[0051] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects:

[0052] (1) The present invention uses large-scale sensors to collect environmental data and plan the optimal path; on this basis, high-precision sensors are used to collect more detailed environmental data in real time to perform more accurate obstacle detection, and the optimal path is dynamically adjusted locally based on the detection results to avoid obstacles, thereby realizing multi-scale environmental perception and effectively improving the safety and efficiency of the operation of the unmanned system; further, according to the real-time speed of the unmanned system, the power distribution between the energy supply units in the hybrid energy supply system is optimized in real time with the goal of minimizing energy consumption, thereby ensuring efficient use of energy and minimizing consumption, and significantly extending the endurance of the unmanned system. In general, the present invention determines the overall range of energy demand of the system load from the load end such as the overall route trajectory, and then re-optimizes the scheduling of the energy system based on the optimized overall load in the dynamic obstacle avoidance and real-time path update process of the unmanned system, obtains a greater degree of systematic optimization performance of the unmanned system through the load-source bidirectional collaborative optimization mechanism, realizes the three-layer optimization strategy of external environment perception, track planning and internal energy management, and can improve the endurance of the hybrid energy supply unmanned system and achieve safe and efficient navigation.

[0053] (2) In the preferred solution of the present invention, when an obstacle is detected and the optimal path is locally adjusted to avoid the obstacle, the feasible speed and trajectory of the submersible within a given time window are calculated in real time based on the improved dynamic window approach (DWA), and local obstacle avoidance and path dynamic adjustment are performed. Specifically, the motion state of the unmanned system is described by the speed, angular velocity and position of the unmanned system. When there is a possibility of collision when navigating along the optimal path, the speed is sampled in the forward speed sampling space, and the position of the unmanned system is updated based on the speed sampling result, while ensuring that the updated position will not collide with the obstacle, thereby efficiently and accurately planning a local path that can avoid obstacles. In a further preferred solution, at each sampling moment, the current optimal path is selected from multiple groups of trajectories based on an evaluation function that simultaneously includes an azimuth evaluation function, a distance evaluation function and a speed evaluation function, ensuring that the highest navigation efficiency can be obtained while avoiding obstacles. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 A flow chart of a three-level optimization method for load-source coordination of a hybrid energy supply unmanned system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0056] In the present invention, the terms "first", "second", etc. (if any) in the present invention and the accompanying drawings are used to distinguish similar objects but not necessarily to describe a specific order or sequence.

[0057] In order to solve the technical problem that the existing autonomous navigation scheme of hybrid energy supply unmanned system cannot simultaneously improve the navigation accuracy, operation safety and endurance of the unmanned system, the present invention provides a method and system for collaborative optimization of obstacle avoidance navigation and energy scheduling of hybrid energy supply unmanned system, the overall concept of which is to realize the organic combination of environmental perception, path planning and energy management, through a multi-level collaborative optimization strategy, so that each subsystem works in coordination at different levels, and utilizes multi-scale environmental perception technology to integrate data from multiple sensors to improve the perception accuracy and real-time performance of the environment, thereby providing reliable environmental information for path planning; through path planning, combined with real-time environmental data, an energy-efficient and safe navigation path is planned to ensure that the unmanned system can efficiently navigate in a complex environment; through a refined energy management strategy, the output distribution relationship of the fuel cell hybrid power system is dynamically adjusted to minimize energy consumption and extend the endurance of the autonomous unmanned system.

[0058] The load-source collaborative three-level optimization method provided by the present invention can be applied to hybrid energy-supply unmanned systems in various complex environments such as underwater and in the air, such as underwater unmanned submersibles, high-altitude drones, etc. Without loss of generality, in the following embodiments, an underwater unmanned submersible powered by a hybrid of fuel cells and lithium batteries is used as an example for explanation.

[0059] For the underwater unmanned submersible, its system model can be constructed, including a fuel cell hybrid power system model composed of fuel cells and lithium batteries used to describe the characteristics of the power system, a dynamics model used to describe the dynamic characteristics of the autonomous unmanned system, and a sensor model used to describe the perception mode of the autonomous unmanned system. Among them:

[0060] The sensor model describes the environmental perception of autonomous unmanned systems, including the detection range, resolution, data update frequency, noise characteristics and other parameters of various sensors. Through the sensor model, it is possible to simulate how underwater unmanned submersibles perceive and model the surrounding environment in actual operation, provide real-time environmental data support, and provide the necessary information basis for dynamic path planning and obstacle avoidance strategies.

[0061] The dynamic model is used to describe the dynamic characteristics of underwater unmanned submersibles, including kinematics and dynamics. The kinematic model involves the motion state parameters of the submersible, such as speed, acceleration, angular velocity, and heading angle, while the dynamic model covers the force conditions, propulsion system characteristics, and motion equations of the submersible in the underwater environment. The dynamic model can be used to predict the motion behavior of underwater unmanned submersibles under various operating conditions, providing basic data support for path planning and track optimization.

[0062] The hybrid power system model describes in detail the working principle and performance characteristics of the hybrid power system composed of fuel cells and lithium batteries. The fuel cell part mainly includes voltage, current, power output characteristics and the energy conversion efficiency of the fuel cell under different operating conditions. The lithium battery part covers parameters such as charge and discharge characteristics, energy density and battery life. Through the comprehensive modeling of fuel cells and lithium batteries, the energy distribution and dynamic performance of the hybrid power system in different mission stages can be accurately simulated to ensure efficient energy utilization and long endurance of the submersible.

[0063] In order to achieve multi-scale environmental perception, optionally, in the following embodiments, sonar is selected as a large-scale sensor, and an underwater camera is selected as a high-precision sensor. Sonar has the characteristics of a wide exploration range and low precision, and the camera has the characteristics of high detection accuracy and a relatively small range. By combining the two, large-scale low-precision environmental perception and small-scale high-precision detail recognition can be achieved. It should be noted that the selection of sensors here is only an optional implementation method and should not be understood as the only limitation of the present invention. In some other embodiments of the present invention, other sensors can also be selected according to the exploration range and accuracy requirements.

[0064] The following are examples.

[0065] Embodiment 1:

[0066] A three-level optimization method for load-source coordination of hybrid energy supply unmanned system, such as Figure 1 As shown, including:

[0067] First, use sonar to collect environmental data and obtain large-scale environmental perception data. Specifically, by sending and receiving sound waves, the sonar system can quickly obtain an overview of the surrounding environment over a large range, including information such as terrain structure, the location and shape of large obstacles, etc. Although the resolution of sonar is relatively low, its wide detection range and strong penetration ability enable it to provide comprehensive environmental data in complex underwater environments, providing basic information for the initial path planning of the submersible.

[0068] After obtaining the large-scale environmental perception data, the optimal path from the starting point to the target point is searched according to the large-scale environmental perception data; optionally, this embodiment uses a particle swarm optimization algorithm (PSO) to search for the optimal path from the starting point to the target, randomly generates N particles, and each particle represents the optimal path from the starting point P start To the end point P goal Calculate the fitness value f(p) of each particle and record the local optimal p of each particle. best and the global optimal g best ;

[0069] The fitness value of a particle is used to measure the quality of each path; in the embodiment, the calculation formula of the fitness value of a particle is:

[0070] f(p)=ρ·PathLength(p)+σ·ObstacleAvoidance(p)+

[0071] τ·EnergyConsumption(p)

[0072] Where p represents a particle, f(p) represents the fitness value of particle p; PathLength(p), ObstacleAvoidance(p) and EnergyConsumption(p) represent the path length, obstacle avoidance capability and energy consumption of the path corresponding to particle p, respectively; ρ, σ and τ represent the coefficients of path length, obstacle avoidance capability and energy consumption, respectively.

[0073] The fitness values ​​designed in this embodiment take into account factors such as path length, obstacle avoidance capability, and energy consumption, and can plan an energy-efficient and safe navigation path to ensure that the unmanned system can navigate efficiently in a complex environment.

[0074] After planning the optimal path, this embodiment will enable the unmanned system to navigate along the optimal path, and divide the navigation process into multiple prediction time domains, and perform dynamic obstacle avoidance and load-source bidirectional coordinated energy scheduling in each prediction time domain in turn.

[0075] Dynamic obstacle avoidance includes:

[0076] Underwater cameras are used to collect environmental data in real time to obtain high-precision environmental perception data. Underwater high-precision cameras can obtain high-resolution image data in a smaller range and capture detailed features in the environment. Underwater camera systems can identify and locate small obstacles, microstructures, and environmental changes through real-time transmission and processing of high-definition images, thereby providing accurate obstacle avoidance information and path adjustment when the submersible approaches obstacles.

[0077] Obstacle detection based on high-precision environmental perception data;

[0078] If it is detected that the collision area of ​​an obstacle intersects with the optimal path, the optimal path is partially adjusted to avoid the collision area of ​​each obstacle;

[0079] Optionally, this embodiment adopts an improved dynamic window approach (DWA) to calculate the feasible speed and trajectory of the submersible in a given time window in real time and perform local obstacle avoidance and dynamic path adjustment;

[0080] It is easy to understand that when the underwater unmanned submersible navigates along the optimal path planned in the above large-scale range, its speed v, angular velocity ω and position Pos are used to describe the basic quantities of its motion state. Due to its own hardware conditions and environmental limitations, the speed, acceleration and angular velocity of the underwater unmanned submersible during operation will be constrained, which can be described by the following formula:

[0081]

[0082] Among them, v min and v max Respectively represent the minimum speed and maximum speed; a min and a max Respectively represent the minimum acceleration and maximum acceleration; α min and α max Respectively represent the minimum angular acceleration and the maximum angular acceleration; ω min and ω max Represent the minimum angular velocity and the maximum angular velocity respectively; and represent the minimum angular acceleration and the maximum angular acceleration respectively.

[0083] In order to perform local obstacle avoidance and dynamic path adjustment, DWA is used to dynamically generate a set of candidate speeds using the current state and constraints, and the optimal speed is selected from them to achieve real-time obstacle avoidance and path optimization. m for:

[0084]

[0085] Where Δt represents the sampling time interval; at each sampling moment, for the forward velocity sampling space V m Each set of velocities (v, ω) obtained by sampling can be used to calculate the position Pos(t+Δt) at the next sampling time, and the positions that may collide with obstacles are eliminated to ensure that the underwater unmanned submersible will not collide with obstacles at the next sampling time;

[0086] Predicting the position at the next sampling moment and ensuring that there is no collision with obstacles can be done using the following formula:

[0087]

[0088] Among them, k and k-1 represent two consecutive sampling moments; (Pos_x k ,Pos_y k ,Pos_θ k ) and (Pos_x k-1 ,Pos_y k-1 ,Pos_θ k-1) represent the states of the underwater unmanned submersible at sampling time k and k-1, respectively, where Pos_x and Pos_y represent coordinates, Pos_θ represents the heading angle; Obs (x,y) Indicates the location coordinates of the obstacle, Dis(Obs (x,y) ,(Pos_x k ,Pos_y k )) represents the distance between the obstacle and the underwater unmanned vehicle, and r represents the preset safety distance.

[0089] According to the dynamic model of the underwater unmanned submersible and its state estimation equation, the underwater unmanned submersible will perform multiple sets of forward trajectory planning within the constraints of dynamic performance, and select the current optimal path according to the path evaluation function. In this embodiment, the path evaluation function can be described by the following formula:

[0090]

[0091] Among them, Heading(v,ω) represents the azimuth evaluation function, which is used to evaluate the angular difference between the direction of the trajectory end and the target point at the currently set speed of the unmanned system; Dist(v,ω) represents the distance evaluation function, which is used to evaluate the collision distance between the unmanned system and the obstacle; Velocity(v,ω) represents the speed evaluation function, which is used to evaluate the navigation speed of the unmanned system within the constraint range; α, β, and γ are the coefficients of the azimuth evaluation function Heading(v,ω), the distance evaluation function dist(v,ω) and the speed evaluation function Velocity(v,ω), respectively.

[0092] Based on the above dynamic obstacle avoidance method, the operating speed of the underwater unmanned submersible in the future can be predicted to obtain an operating speed sequence, and the speed sequence can be converted into a corresponding power demand sequence in combination with the dynamic model of the underwater unmanned submersible, thereby optimizing the energy distribution of the power system and achieving optimal navigation fuel consumption. Accordingly, in this embodiment, energy scheduling includes the following steps:

[0093] (S1) predicting the speed range of the unmanned system at each prediction time within the current prediction time domain according to the current optimal path, and mapping it into a feasible load demand range at the corresponding prediction time, and determining a feasible energy scheduling range corresponding to each feasible load demand range;

[0094] There is a feasible speed range in the dynamic adjustment process of the unmanned system's external path planning. Based on this feasible speed range, the corresponding load demand range can be obtained. For each possible load demand in the load demand range, there is a feasible energy allocation scheme. In this way, each feasible energy allocation scheme together constitutes a feasible energy scheduling range. The minimum external load demand is not necessarily the optimal result. The optimal scheduling result is found from the feasible load demand range to the corresponding feasible scheduling range.

[0095] (S2) establishing an objective function with the goal of minimizing the total equivalent hydrogen consumption of the hybrid energy supply system in the current prediction time domain, solving the output power change rate of each energy supply unit in the hybrid energy supply system at each prediction time based on the feasible load demand interval and the corresponding feasible energy scheduling interval, and obtaining the optimal control variable sequence;

[0096] Optionally, in this embodiment, the objective function established is:

[0097]

[0098] Where J represents the objective function value; P i represents the prediction time domain, i represents the prediction time in the prediction time domain; Indicates that in the prediction domain P i The equivalent hydrogen consumption of the hybrid energy supply system at the i-th prediction moment within ;

[0099] It is easy to understand that there are corresponding constraints on the output power, power change rate, etc. of various batteries in the hybrid energy supply system. In this embodiment, the constraints include the output power constraint and output power change rate constraint of the fuel cell, and the state of charge constraint and output power constraint of the lithium battery, which can be specifically expressed by the following formula:

[0100]

[0101] Among them, P fc and ΔP fc They represent the output power and output power change rate of the fuel cell respectively, and They represent the minimum output power and maximum output power of the fuel cell respectively, and They represent the minimum output power change rate and the maximum output power change rate of the fuel cell respectively; SOC and P BAT Respectively represent the state of charge and output power of the lithium battery, SOC min and SOC max They represent the maximum state of charge and minimum state of charge of the lithium battery respectively. and Respectively represent the maximum charging power and maximum discharging power of lithium batteries;

[0102] In order to meet the requirements of implementing control, this embodiment establishes a state space model describing the hybrid energy supply system as follows:

[0103]

[0104] Where S is the state of charge SOC of the lithium battery and the output power P of the fuel cell. fc The state variable u is composed of the rate of change of the fuel cell output power ΔP fc The control variable is: P is the output power of the hybrid energy supply system composed of fuel cells and lithium batteries; A, B u , C are system matrix, control matrix and output matrix respectively;

[0105] According to the hybrid power system model, the system response prediction equation at the future k+N moment is established as follows:

[0106]

[0107]

[0108] Where N represents the length of the prediction time domain;

[0109] Optionally, in this embodiment, a dynamic programming (DP) solver is specifically used to solve the objective function to obtain an optimal control variable sequence in the prediction time domain;

[0110] (S3) Controlling the output power of each energy supply unit in the hybrid energy supply system at each predicted time according to the optimal control variable sequence.

[0111] In this embodiment, the underwater unmanned submersible obtains the speed sequence according to the optimal path after dynamic obstacle avoidance through the above energy scheduling method, predicts the power of the next N steps, sets constraints under the premise that the equivalent hydrogen consumption of the underwater unmanned submersible is the objective function, and uses a dynamic programming solver to obtain the optimal control variable sequence in the prediction time domain, and applies the first value of the optimal control variable sequence to the hybrid energy supply system, and loops the above process for rolling optimization, thereby realizing the underwater unmanned submersible's obstacle avoidance and extreme endurance. The control method used for energy scheduling in this embodiment is a model predictive control (MPC) method.

[0112] In general, this embodiment proposes a three-level optimization method for load-source coordination of a multi-level autonomous unmanned system covering environmental perception, trajectory planning, and energy management. Through multi-scale dynamic environment perception technology, the autonomous unmanned system can obtain and update detailed information about the surrounding environment in real time, and combine the improved dynamic window method to perform real-time obstacle avoidance and path replanning; according to the dynamic parameters such as the current position, speed, acceleration, etc. of the autonomous unmanned system, multiple feasible paths are generated and evaluated, and the optimal trajectory is selected in real time. In this way, the autonomous unmanned system can not only ensure that the path optimization under a large scale range reduces unnecessary path detours and stagnation time, but also can quickly respond to environmental changes, avoid obstacles, and avoid collision risks. In a complex environment, navigation accuracy and safety are significantly improved. At the same time, the speed sequence in the path planning process is an organic combination of online real-time energy management and optimization. Further, according to the power requirements under different mission stages and environmental conditions, the power output of each energy supply unit in the hybrid energy supply system is intelligently adjusted to ensure efficient use of energy and minimize consumption, significantly extend the endurance of the autonomous unmanned system, and through comprehensive power system modeling, the energy requirements of the autonomous unmanned system and the performance of the fuel cell under different operating conditions can be predicted to prevent potential failures and improve the reliability of the system.

[0113] Embodiment 2:

[0114] A load-source collaborative three-level optimization system for a hybrid energy supply unmanned system, including: a track planning module, a collaborative optimization control module, a dynamic obstacle avoidance module and an energy scheduling module;

[0115] The trajectory planning module is used to search for the optimal path from the starting point to the target point based on the environmental data collected by large-scale sensors;

[0116] Dynamic obstacle avoidance module, used to perform dynamic obstacle avoidance;

[0117] Energy scheduling module, used to perform energy scheduling of load-source bidirectional collaboration;

[0118] A collaborative optimization control module is used to make the unmanned system navigate along the optimal path and divide the navigation process into multiple prediction time domains. In each prediction time domain, the dynamic obstacle avoidance module and the energy scheduling module are used to perform dynamic obstacle avoidance and energy scheduling in turn.

[0119] Dynamic obstacle avoidance includes: obstacle detection based on environmental data collected in real time by high-precision sensors; if it is detected that the collision area of ​​an obstacle intersects with the optimal path, the optimal path is partially adjusted to avoid the collision area of ​​each obstacle;

[0120] Energy scheduling includes the following steps:

[0121] (S1) predicting the speed range of the unmanned system at each prediction time in the current prediction time domain according to the current optimal path, and mapping it into a feasible load demand range at the corresponding prediction time, and determining a feasible energy dispatch range corresponding to each feasible load demand range;

[0122] (S2) establishing an objective function with the goal of minimizing the total equivalent hydrogen consumption of the hybrid energy supply system in the current forecast time domain, solving the output power change rate of each energy supply unit in the hybrid energy supply system at each forecast time based on the feasible load demand interval and the corresponding feasible energy scheduling interval, and obtaining the optimal control variable sequence;

[0123] (S3) controlling the output power of each energy supply unit in the hybrid energy supply system at each predicted time according to the optimal control variable sequence;

[0124] Among them, large-scale sensors and high-precision sensors are both carried on unmanned systems, and the large-scale sensors have a wider search range, and the high-precision sensors have higher detection accuracy; the collision area is the area formed by the position where the distance to the obstacle is less than the preset safety distance.

[0125] In this embodiment, the specific implementation of each module can refer to the description of the above-mentioned embodiment 1 and will not be repeated here.

[0126] Embodiment 3:

[0127] An autonomous unmanned system, comprising a hybrid energy supply unmanned system and a load-source coordination three-level optimization system of the hybrid energy supply unmanned system provided in the above-mentioned embodiment 1;

[0128] Unmanned systems are equipped with large-scale sensors and high-precision sensors; among these sensors, large-scale sensors have a wider search range, and high-precision sensors have higher detection accuracy.

[0129] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A three-level optimization method for load-source coordination of a hybrid energy supply unmanned system, characterized in that: include: Searching for an optimal path from a starting point to a target point based on environmental data collected by large-scale sensors; enabling the unmanned system to navigate along the optimal path, and dividing the navigation process into multiple prediction time domains, and sequentially performing dynamic obstacle avoidance and load-source bidirectional coordinated energy scheduling in each prediction time domain; The dynamic obstacle avoidance includes: performing obstacle detection based on environmental data collected in real time by high-precision sensors; if it is detected that the collision area of ​​an obstacle intersects with the optimal path, locally adjusting the optimal path to avoid the collision area of ​​each obstacle; The energy scheduling comprises the following steps: (S1) predicting the speed range of the unmanned system at each prediction time within the current prediction time domain according to the current optimal path, and mapping it into a feasible load demand range at the corresponding prediction time, and determining a feasible energy scheduling range corresponding to each feasible load demand range; (S2) establishing an objective function with the goal of minimizing the total equivalent hydrogen consumption of the hybrid energy supply system in the current prediction time domain, solving the output power change rate of each energy supply unit in the hybrid energy supply system at each prediction time based on the feasible load demand interval and the corresponding feasible energy scheduling interval, and obtaining the optimal control variable sequence; (S3) controlling the output power of each energy supply unit in the hybrid energy supply system at each predicted time according to the optimal control variable sequence; Among them, large-scale sensors and high-precision sensors are both installed in the unmanned system, and the large-scale sensors have a wider search range, and the high-precision sensors have higher detection accuracy; the collision area is an area formed by positions where the distance between the obstacle and the obstacle is less than a preset safety distance.

2. The load-source coordination three-level optimization method of the hybrid energy supply unmanned system according to claim 1 is characterized in that: The optimal path is locally adjusted to avoid the collision area of ​​each obstacle, including: (T1) When there is a possibility of collision when continuing to navigate along the optimal path, the current time is taken as the current sampling time; (T2) At the current sampling time, sampling is performed in the forward speed sampling space of the unmanned system to obtain multiple sets of candidate speeds; the forward speed sampling space is: (T3) calculating the position of the unmanned system at the next sampling time according to the position of the unmanned system at the current sampling time and the candidate speed sampling results, and eliminating the position within the collision area; (T4) The positions of the current sampling time and the previous sampling times are combined into a current local path, and the optimal local path is selected therefrom; (T5) If the real-time trajectory adjustment of the dynamic obstacle is completed, proceed to step (T6); otherwise, proceed to step (T2); (T5) adjusting the optimal path using the current optimal local path; Wherein, (v, ω) represents a set of candidate speeds, v represents the speed of the unmanned system, ω represents the angular velocity of the unmanned system; Δt represents the sampling time interval; a max represents the maximum acceleration, Indicates the maximum angular acceleration.

3. The load-source coordination three-level optimization method of the hybrid energy supply unmanned system according to claim 2 is characterized in that: The step (T4) selects the optimal local path from the current local path based on the evaluation function; The evaluation function is: Among them, Heading(v,ω) represents the azimuth evaluation function, which is used to evaluate the angular difference between the direction of the end of the trajectory and the target point at the currently set speed of the unmanned system; Dist(v,ω) represents the distance evaluation function, which is used to evaluate the collision distance between the unmanned system and the obstacle; Velocity(v,ω) represents the speed evaluation function, which is used to evaluate the navigation speed of the unmanned system within the constraint range; α, β, γ are the coefficients of the azimuth evaluation function Heading(v,ω), the distance evaluation function dist(v,ω) and the speed evaluation function Velocity(v,ω), respectively.

4. The load-source coordination three-level optimization method of the hybrid energy supply unmanned system according to any one of claims 1 to 3, characterized in that: In the step (S2), the objective function is expressed as: Where J represents the objective function value; P i represents the prediction time domain, i represents the prediction time in the prediction time domain; Indicates that in the prediction domain P i The equivalent hydrogen consumption of the hybrid energy supply system at the i-th prediction moment within .

5. The load-source coordination three-level optimization method of the hybrid energy supply unmanned system according to any one of claims 1 to 3, characterized in that: Search for the optimal path from the starting point to the target point based on the environmental data collected by large-scale sensors, including: Model the environment based on the environmental data collected by large-scale sensors and determine all paths from the starting point to the target point; Each path is a particle in the particle swarm algorithm, and the particle swarm algorithm is used to search for the optimal path from all paths; during the search process, the calculation formula of the particle fitness value is: f(p)=ρ·PathLength(p)+σ·ObstacleAvoidance(p)+τ EnergyConsumption(p) Where p represents a particle, f(p) represents the fitness value of particle p; PathLength(p), ObstacleAvoidance(p) and EnergyConsumption(p) represent the path length, obstacle avoidance capability and energy consumption of the path corresponding to particle p, respectively; ρ, σ and τ represent the coefficients of path length, obstacle avoidance capability and energy consumption, respectively.

6. A three-level optimization system for load-source coordination of a hybrid energy supply unmanned system, characterized in that: include: Trajectory planning module, collaborative optimization control module, dynamic obstacle avoidance module and energy scheduling module; The trajectory planning module is used to search for the optimal path from the starting point to the target point based on the environmental data collected by the large-scale sensor; The dynamic obstacle avoidance module is used to perform dynamic obstacle avoidance; The energy scheduling module is used to perform energy scheduling for load-source bidirectional coordination; The collaborative optimization control module is used to make the unmanned system navigate along the optimal path, and divide the navigation process into multiple prediction time domains, and sequentially use the dynamic obstacle avoidance module and the energy scheduling module to perform dynamic obstacle avoidance and energy scheduling in each prediction time domain; The dynamic obstacle avoidance includes: performing obstacle detection based on environmental data collected in real time by high-precision sensors; if it is detected that the collision area of ​​an obstacle intersects with the optimal path, locally adjusting the optimal path to avoid the collision area of ​​each obstacle; The energy scheduling comprises the following steps: (S1) predicting the speed range of the unmanned system at each prediction time within the current prediction time domain according to the current optimal path, and mapping it into a feasible load demand range at the corresponding prediction time, and determining a feasible energy scheduling range corresponding to each feasible load demand range; (S2) establishing an objective function with the goal of minimizing the total equivalent hydrogen consumption of the hybrid energy supply system in the current prediction time domain, solving the output power change rate of each energy supply unit in the hybrid energy supply system at each prediction time based on the feasible load demand interval and the corresponding feasible energy scheduling interval, and obtaining the optimal control variable sequence; (S3) controlling the output power of each energy supply unit in the hybrid energy supply system at each predicted time according to the optimal control variable sequence; Among them, large-scale sensors and high-precision sensors are both installed in the unmanned system, and the large-scale sensors have a wider search range, and the high-precision sensors have higher detection accuracy; the collision area is an area formed by positions where the distance between the obstacle and the obstacle is less than a preset safety distance.

7. An autonomous unmanned system, characterized in that: A hybrid energy supply unmanned system and a load-source coordination three-level optimization system for the hybrid energy supply unmanned system as claimed in claim 6; The unmanned system is equipped with large-scale sensors and high-precision sensors; among the large-scale sensors and the high-precision sensors, the large-scale sensors have a wider search range, and the high-precision sensors have higher detection accuracy.

8. The autonomous unmanned system according to claim 7, characterized in that: The large-scale sensor is a sonar, and the high-precision sensor is a camera.

Images