Virtual simulation control system for amphibious unmanned platform

Abstract

The invention relates to the field of underwater unmanned platforms, and discloses a virtual simulation control system specially designed for an amphibious unmanned platform. Comprising the steps that a virtual part is constructed by using a Gazebo visual platform, such as a virtual environment, completes the construction of a land and water complex environment, and loads a physical engine; the platform model is an amphibious platform structure model, and a multi-body dynamic model and a buoyancy and hydrodynamic model are added; the sensor simulates and carries an IMU, a GPS and a depth sensor, and the camera feeds back platform state information in real time. Performing behavior control and state management to realize construction of a behavior tree and a state machine; through artificial intelligence and machine learning, autonomous navigation and path planning functions of the platform are realized, and an autonomous operation instruction is issued to a control system. The system provides a comprehensive tool for verifying and optimizing platform design and control strategies. Through comprehensive pre-deployment testing in a simulation environment, the safety, efficiency and accuracy of amphibious platform operation are improved.

Description

Technical Field

[0001] The present invention belongs to the field of underwater unmanned platforms, and in particular relates to a virtual simulation control system suitable for amphibious unmanned platforms. Background Art

[0002] As the demand for amphibious unmanned platforms in environmental monitoring, disaster relief and military applications continues to increase, their operational capabilities and reliability in complex environments have become key issues. Traditional physical testing methods are costly and easily limited by environmental and equipment conditions, making it difficult to fully cover a variety of possible operating scenarios. Virtual simulation technology provides a low-risk, high-efficiency testing method by simulating real-world physical and operational conditions in a computer environment. However, current virtual simulation technology still has limitations in terms of environmental complexity, dynamic simulation accuracy and control strategy optimization, making it difficult to fully meet the needs of amphibious unmanned platforms. Summary of the invention

[0003] In order to solve the deficiencies in the prior art, the present invention proposes a virtual simulation control system for the design, testing and optimization of an amphibious unmanned platform.

[0004] A virtual simulation control system for an amphibious unmanned platform, characterized by comprising:

[0005] Virtual construction part: Virtual environment construction based on the Gazebo visualization platform, used to build and manage complex water and land environments, and load the physical engine to achieve high-precision dynamic simulation; platform model construction, used to build the structural model of the amphibious platform, add multi-body dynamics model as well as buoyancy and hydrodynamic models to simulate the motion characteristics of the platform in water and on land; sensor simulation, used to simulate IMU, GPS, depth sensor and camera, and provide real-time feedback of the platform's attitude, position and environmental perception information;

[0006] Control function part: includes control system, behavior control and state management, artificial intelligence and machine learning modules, and task simulation and collaboration.

[0007] The virtual environment construction of the Gazebo visualization platform integrates the ODE (Open Dynamics Engine) physics engine to simulate the physical behavior of the platform in complex water and land environments. It is further configured to simulate the flow characteristics of water bodies and various terrain environments on land, including flat land, slopes and obstacle areas.

[0008] Among them, the platform model is constructed based on the Bullet physics engine, which constructs a multi-body dynamics model of the amphibious platform under different motion states.

[0009] The movement state is divided into an underwater stealth state and an underwater crawling state.

[0010] Among them, the mathematical model of the underwater stealth state is constructed based on the nonlinear kinematic model and dynamic model of the platform:

[0011]

[0012] Where M = M RB +M AM , C(v)=C RB (v)+C AM (v), τ=τ B +τ d .

[0013] M RB : mass inertia matrix; M AM : Additional water quality matrix; C RB : "Coriolis force-centripetal force" matrix; C AM : Additional Coriolis force and moment matrix; D(v) is the damping matrix; g(η): restoring force matrix; τ d : the interference force and torque generated by the external environment; τ B :Thrust and torque in six degrees of freedom.

[0014] Crawling in water: The platform can be simplified into a two-wheel differential car model. Ignoring the roll and pitch motions, the four-degree-of-freedom model of the platform is obtained:

[0015]

[0016] Where d 11 =X u +X u|u| |u|,d 22 =Y v +Y v|v| |v|, X u|q| ,Y ν ,Y ν|q| ,Z w ,N r ,N rr They are the hydrodynamic parameters and damping terms of the underwater amphibious obstacle-breaking platform during navigation.

[0017] Among them, sensor simulation uses UUV Simulator (ROS simulation tool for underwater vehicles) to simulate GPS sensors, IMU, and depth gauges. The configuration can add multiple noise items to dynamically simulate the actual noise level and accuracy.

[0018] Among them, the attitude control algorithm in the control system module includes attitude control algorithms based on nonlinear PID control and sliding mode control. The algorithm combines the platform's dynamic model with real-time sensor data, and generates simulated thrust instructions through the control plug-in in Gazebo.

[0019] Among them, the behavior control and state management module uses the BehaviorTree.CPP library to build a control framework based on the behavior tree, autonomous behavior switching and state management in multi-task mode, and schedules each behavior node through the state machine

[0020] Among them, artificial intelligence and machine learning modules are used to simulate and verify algorithm functions such as platform path tracking, and establish global navigation, local path planning and automatic obstacle avoidance algorithms to achieve autonomous navigation.

[0021] Among them, the task simulation and collaboration module is based on the Multi-Robot Communication framework of ROS to realize task allocation and collaborative simulation among multiple amphibious platforms, and share task status and environmental information in real time through the network communication plug-in in Gazebo, thereby optimizing the platform's task collaboration capabilities.

[0022] The virtual simulation control system for an amphibious unmanned platform provided in the present application has the following beneficial effects and advantages:

[0023] 1. Multi-state precise dynamic modeling: By integrating ODE and Bullet physics engines, the present invention realizes precise dynamic modeling of amphibious platforms in different motion states, covering detailed simulation of multiple states such as underwater diving and underwater crawling.

[0024] 2. Advanced control and path planning algorithms: Combining nonlinear PID control, sliding mode control and a variety of path planning algorithms (such as A*, DWA, artificial potential field method), the system has high-precision attitude control and adaptive path planning capabilities.

[0025] 3. Powerful task collaboration and management capabilities: Through the BehaviorTree.CPP library and the Multi-RobotCommunication framework of ROS, the system performs well in the collaborative execution of multiple tasks and multiple platforms, realizing dynamic task management and real-time collaborative simulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 : Overall architecture diagram of virtual simulation control system;

[0027] Figure 2 : Amphibious platform coordinate system;

[0028] Figure 3:Kinematic analysis of underwater crawling of amphibious platform;

[0029] Figure 4 : Control system block diagram;

[0030] Figure 5 : Path tracking technology route;

[0031] Figure 6 : Three-dimensional path tracking simulation example;

[0032] Figure 7 : Path planning ROS node graph. DETAILED DESCRIPTION

[0033] 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 are within the scope of protection of the present invention.

[0034] The present invention discloses a virtual simulation control system for an amphibious unmanned platform, which aims to achieve precise control and task execution of the amphibious unmanned platform in a complex water and land environment through simulation technology and advanced control algorithms. Figure 1 As shown: The system integrates the virtual construction part (virtual environment construction, platform model construction, sensor simulation, communication and network) and the control function part

[0035] Virtual build part:

[0036] Gazebo visualization platform and ODE physics engine integration: The virtual environment construction part of the virtual simulation control system is based on the Gazebo visualization platform and integrated with the ODE physics engine. The ODE engine is responsible for simulating the physical behavior of the platform in the water and land environment, including the flow characteristics of the water body and the various terrain environments on land. The simulation environment includes not only flat land, but also slopes and obstacle areas, simulating the complex terrain conditions that may be encountered in real scenes.

[0037] Platform model construction and application of Bullet physics engine: The platform model is constructed using the Bullet physics engine, which is used to build the multi-body dynamics model of the amphibious platform. The URDF (Unified Robot Description Format) is used to build the platform's physical model, including the multi-body dynamics characteristics under different motion states, and detailed modeling is performed for the underwater diving state and the underwater crawling state. First, the following is established: Figure 2The amphibious platform coordinate system is described in Table 1.

[0038] Table 1 Description of basic motion symbols

[0039]

[0040]

[0041] Underwater diving state model: When the platform dives in water, the model is based on the nonlinear kinematics and dynamics model of the platform, taking into account factors such as the platform's inertia matrix, additional water mass matrix, Coriolis force matrix, damping matrix, and the interference force and torque of the external environment. Its motion equation is as follows:

[0042]

[0043] Where M = M RB +M AM , C(v)=C RB (v)+C AM (v), τ=τ B +τ d .

[0044] M RB : mass inertia matrix; M AM : Additional water quality matrix; C RB : "Coriolis force-centripetal force" matrix; C AM : Additional Coriolis force and moment matrix; D(v): Damping matrix; g(η): Restoring force matrix; τ d : the interference force and torque generated by the external environment; τ B :Thrust and torque in six degrees of freedom.

[0045] Water crawling state model: Figure 3 When the platform shown crawls in the water, its motion can be simplified to the model of a two-wheel differential car. The parameters in the figure represent:

[0046] W: The distance between the two driving wheels, from the center point O, unit: m.

[0047] V x : Target forward speed at point O, forward is positive, unit: m / s.

[0048] V z : The target rotation speed around point O, counterclockwise is positive, unit: rad / s.

[0049] R: The turning radius R generated by moving forward and rotating at the same time, unit: m.

[0050] V L 、V R: Left and right wheel speeds, coordinated to achieve target speed V x 、V z , forward movement is positive, unit: m / s.

[0051] Arc L 、Arc M 、Arc R : The path traveled by the left wheel, point O, and right wheel within a certain time t, unit: m.

[0052] θ: The angle of rotation within a certain time t, unit: rad.

[0053] The integral of speed over time equals distance:

[0054]

[0055] The arc length divided by the radius equals the radian:

[0056]

[0057] Integrating over time gives:

[0058]

[0059] Decomposition can be obtained:

[0060]

[0061] The inverse kinematics formula can be obtained by combining the target speed V x 、V Z Find the target speed V of the driving wheel L 、V R :

[0062]

[0063] The kinematic solution formula can be obtained by deformation. The current speed V of the driving wheel is known. L 、V R Find the real-time speed V of the current platform x 、V Z :

[0064]

[0065] According to the actual situation, the platform model needs to be simplified to a certain extent. This study ignores the influence of the heeling motion on the platform. Since the depth change of the platform is mainly studied in the vertical direction, the influence of the longitudinal motion on the platform is ignored. In addition, in the dynamic modeling of the platform, the influence of the higher-order nonlinear hydrodynamic coefficients on the model parameters is ignored. In summary, the four-degree-of-freedom model of the platform is obtained.

[0066] The simplified four-degree-of-freedom kinematic model of the platform is:

[0067]

[0068] Where d 11 =X u +X u|u| |u|,d 22 =Y v +Y v|v| |v|, X u|q| ,Y ν ,Y ν|q| ,Z w ,N r ,N rr They are the hydrodynamic parameters and damping terms of the underwater amphibious obstacle-breaking platform during navigation.

[0069] Sensor simulation and UUV Simulator application: The sensor simulation part integrates the UUV Simulator tool, which is used to simulate the platform's GPS, IMU, depth sensor and camera. The simulation system introduces multiple noise items to dynamically simulate the noise level and accuracy changes that may be encountered in actual operation, ensuring the consistency of simulation data with actual operation data.

[0070] Attitude control algorithm in control system: such as Figure 4 The system shown first receives the desired position and attitude (such as position x, y, z and attitude angles θ, ψ, φ), and calculates the corresponding desired velocity vector Vd = [ud, vd, wd, qd, rd] T These velocity components represent the target velocity of the platform in each degree of freedom, including linear velocity and angular velocity along different axes. The system then compares the actual position and attitude with the desired state and calculates the position and attitude error e p , and compare the actual speed with the expected speed to get the speed error e v These errors are input to the dynamic speed regulation module, which performs speed regulation through a controller (such as a PID controller) to generate a control output signal τ, which is used to correct the motion trajectory of the platform so that it can accurately follow the predetermined trajectory. This control signal is transmitted to the dynamic model module, which takes into account the hydrodynamic effect and calculates the actual force and torque τ acting on the platform. o These outputs are further passed to the kinematic model, which simulates the platform's motion state in the water based on the current speed and acceleration, calculates the platform's actual speed, position, and attitude, and feeds them back to the system input to form a closed-loop control. The entire control method ensures that the platform can accurately navigate in complex underwater environments, and achieves attitude control and real-time compensation for external disturbances.

[0071] Behavior control and state management module: Based on the BehaviorTree.CPP library, a flexible behavior tree control framework is built. This framework implements the platform's autonomous behavior switching and state management in multi-task mode. The state machine is responsible for scheduling each behavior node in the behavior tree, dynamically adjusting according to task requirements and platform status to ensure efficient execution of tasks.

[0072] Artificial Intelligence and Machine Learning Module: used to simulate and verify the path tracking function of the platform. Figure 5 Global navigation, local path planning and automatic obstacle avoidance algorithms are established to achieve autonomous navigation. A certain working depth is added to the two-dimensional global path planned by the surface satellite map to map it to a three-dimensional path to achieve global path planning. On the basis of global path planning, the surrounding environment is perceived through sonar, a discrete waypoint space is established, a cost function is defined, and a local path planning algorithm based on dynamic planning is designed to complete obstacle avoidance and give reference waypoints. At the same time, the path construction of the given reference waypoint is carried out through the LIPB method, and then the trajectory planning is implemented based on the finite difference method, and the trajectory planning results are sent to the trajectory tracking controller. The global navigation adopts the A* algorithm, and the local path planning combines DWA (dynamic window method) and artificial potential field method to ensure the platform's autonomous navigation capability in complex environments. The machine learning part is used to optimize the path planning algorithm so that it can adapt to environmental changes and improve the accuracy of path tracking.

[0073] like Figure 6 , Figure 7 It is a path tracking instance, which is realized through the collaborative work between ROS nodes. dp_wpts indicates that the path planning node contains global path and local path planning, / reference_wpts is the reference waypoint message, / yaml_modify_node is the node for modifying the reference waypoint, / launch_waypoint_node is the node for publishing the reference waypoint to trajectory planning, and / rexrov / rov_pid_controller is the trajectory tracking control node.

[0074] Task simulation and collaboration module: The Multi-Robot Communication framework based on ROS allows task allocation and collaborative simulation between multiple amphibious platforms. Through the network communication plug-in in Gazebo, each platform can share task status and environmental information in real time to ensure maximum collaborative efficiency during multi-platform task execution. Collaborative simulation further realizes dynamic adjustment and load balancing of tasks between platforms through task priority scheduling algorithm.

[0075] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit the same. Although these embodiments have been described in detail, it is still possible for those skilled in the art to modify the above technical solutions or to perform equivalent replacements on some of the technical features thereof. These modifications or replacements will not deviate from the core spirit and scope of the technical solutions claimed for protection by the present invention.

[0076] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A virtual simulation control system for an amphibious unmanned platform, characterized in that: include: Virtual construction part: Virtual environment construction based on the Gazebo visualization platform, used to build and manage complex water and land environments, and load the physical engine to achieve high-precision dynamic simulation; platform model construction, used to build the structural model of the amphibious platform, add multi-body dynamics model as well as buoyancy and hydrodynamic models to simulate the motion characteristics of the platform in water and on land; sensor simulation, used to simulate IMU, GPS, depth sensor and camera, and provide real-time feedback of the platform's attitude, position and environmental perception information; Control function part: includes control system, behavior control and state management, artificial intelligence and machine learning modules, and task simulation and collaboration.

2. A virtual simulation control system for an amphibious unmanned platform according to claim 1, characterized in that: The virtual environment construction of the Gazebo visualization platform integrates the ODE (Open Dynamics Engine) physics engine to simulate the physical behavior of the platform in complex water and land environments, and is further configured to simulate the flow characteristics of water bodies and various terrain environments on land, including flat land, slopes and obstacle areas.

3. A virtual simulation control system for an amphibious unmanned platform according to claim 1, characterized in that: The platform model is constructed based on the Bullet physics engine, and a multi-body dynamics model of the amphibious platform under different motion states is constructed.

4. A virtual simulation control system for an amphibious unmanned platform according to claim 3, characterized in that: The movement state is divided into an underwater stealth state and an underwater crawling state. Among them, the mathematical model of the underwater stealth state is constructed based on the nonlinear kinematic model and dynamic model of the platform: where M = M RB + M AM , C(v) = C RB (v) + C AM (v), τ = τ B + τ d . M RB : mass inertia matrix; M AM : Additional water quality matrix; C RB : "Coriolis force-centripetal force" matrix; C AM : Additional Coriolis force and moment matrix; D(v) is the damping matrix; g(η): restoring force matrix; τ d : the interference force and torque generated by the external environment; τ B :Thrust and torque in six degrees of freedom. Crawling in water: The platform can be simplified into a two-wheel differential car model, ignoring the roll and pitch motions, and the four-degree-of-freedom model of the platform is obtained: Where d 11 =X u +X u|u| |u|,d 22 =Y v +Y v|v| |v|,d 33 =Z w +Z w|u| |w|,d 44 =N r +N r|v| |r|, X u|q| ,Y ν ,Y ν|q| ,Z w ,N r ,N rr They are the hydrodynamic parameters and damping terms of the underwater amphibious obstacle-breaking platform during navigation.

5. The virtual simulation control system for an amphibious unmanned platform according to claim 1, characterized in that: The sensor simulation utilizes UUV Simulator (ROS simulation tool for underwater vehicles) to simulate GPS sensors, IMUs, and depth gauges, and is configured to add multiple noise items to dynamically simulate actual noise levels and accuracy.

6. A virtual simulation control system for an amphibious unmanned platform according to claim 1, characterized in that: The attitude control algorithm in the control system module includes an attitude control algorithm based on nonlinear PID control and sliding mode control, etc. The algorithm combines the platform's dynamic model with real-time sensor data and generates simulated thrust instructions through the control plug-in in Gazebo.

7. The virtual simulation control system for an amphibious unmanned platform according to claim 1, characterized in that: The behavior control and state management module uses the BehaviorTree.CPP library to build a control framework based on the behavior tree, autonomous behavior switching and state management in multi-tasking mode, and schedules each behavior node through a state machine.

8. The virtual simulation control system for an amphibious unmanned platform according to claim 1, characterized in that: The artificial intelligence and machine learning modules are used to simulate and verify algorithm functions such as platform path tracking, and to establish global navigation, local path planning and automatic obstacle avoidance algorithms to achieve autonomous navigation.

9. A virtual simulation control system for an amphibious unmanned platform according to claim 1, characterized in that: The task simulation and collaboration module is based on the Multi-Robot Communication framework of ROS to realize task allocation and collaborative simulation among multiple amphibious platforms, and shares task status and environmental information in real time through the network communication plug-in in Gazebo, thereby optimizing the task collaboration capability of the platform.

Images