Method for reducing distributed simulation data transmission quantity based on dead reckoning algorithm

Abstract

The invention relates to the field of distributed simulation systems, in particular to a method for reducing distributed simulation data transmission quantity based on a dead reckoning algorithm. The method comprises the following steps: acquiring and preprocessing entity state data; local dead reckoning; dead reckoning error calculation; making a data transmission decision; and optimizing algorithm parameters, and in the whole distributed simulation process, continuously performing entity state acquisition, local dead reckoning, error calculation, data transmission decision and position synchronization operation in each simulation time step until the simulation task is finished. According to the method, the change rule and speculability of entity motion can be fully utilized, unnecessary data transmission is effectively reduced, network loads are reduced, meanwhile, the accuracy and stability of distributed simulation are guaranteed, powerful technical support is provided for large-scale and high-precision distributed simulation application, and the method has wide application prospects and remarkable technical advantages.

Description

Technical Field

[0001] The present invention relates to the field of distributed simulation systems, and in particular to a method for reducing the amount of distributed simulation data transmission based on a dead reckoning algorithm. Background Art

[0002] Distributed simulation systems are widely used in many fields such as military simulation, aerospace simulation, and intelligent transportation system simulation. In these distributed simulation environments, multiple simulation nodes need to work together to simulate complex systems or scenarios. As the scale and complexity of simulations continue to increase, data interaction between nodes is required frequently, and the transmission of location information occupies a considerable amount of network bandwidth resources.

[0003] However, traditional distributed simulation systems often adopt a relatively simple and direct data transmission mechanism, that is, the complete location information of the entity is transmitted between nodes according to fixed time intervals or event trigger conditions. This leads to a huge amount of data transmission, which puts tremendous pressure on the network bandwidth and easily causes problems such as data transmission delays and simulation jams, seriously affecting the performance and efficiency of the distributed simulation system. These problems are particularly prominent in large-scale, high-precision simulation scenarios.

[0004] Technical issues solved

[0005] The purpose of the present invention is to provide a method for reducing the amount of distributed simulation data transmission based on a dead reckoning algorithm, so as to overcome the shortcomings of the prior art and significantly reduce the amount of data transmission between nodes while ensuring the accuracy and reliability of distributed simulation. Summary of the invention

[0006] The present invention provides a method for reducing the amount of distributed simulation data transmission based on a dead reckoning algorithm, comprising the following steps:

[0007] Step 1: Entity state data acquisition and preprocessing

[0008] When the distributed simulation system is started, each simulation node obtains the initial state data of the entity it is responsible for simulating, and standardizes and formats the data. At the same time, the initial parameters of the dead reckoning algorithm are set according to the characteristics and accuracy requirements of the simulation scenario.

[0009] Step 2:

[0010] Each simulation node uses an improved dead reckoning algorithm to predict the future state of the entity;

[0011] Step 3: Dead Reckoning Error Calculation

[0012] After completing local dead reckoning, the error between the locally predicted entity state data and the entity state data actually calculated by the high-precision simulation model is calculated. The position error, velocity error, and heading error are calculated using a multi-dimensional error evaluation index. The corresponding weights are assigned according to the importance of different errors to obtain a total dead reckoning error value.

[0013] Step 4: Data transfer decision

[0014] Compare the error of each dimension with the preset error tolerance range. If the errors of all dimensions are within the tolerance range, the current prediction result is determined to be valid, and there is no need to transmit data to other simulation nodes. Only the prediction data is recorded locally for the continuity calculation of subsequent local simulations. If the error of at least one dimension exceeds the tolerance range, it indicates that the dead reckoning result has a large deviation. At this time, the complete and accurate state data of the current entity needs to be compressed and packaged, and sent to other related simulation nodes according to the communication protocol of the distributed simulation system. After receiving the data, the receiving node immediately updates the corresponding entity state in the local simulation model to ensure that each node in the entire distributed simulation system has a consistent understanding of the entity state, thereby maintaining the accuracy and coherence of the simulation.

[0015] Step 5: Algorithm parameter optimization

[0016] After each data transmission is completed or after N local prediction cycles (N can range from 1 to 100 and can be adjusted according to the accuracy required by different simulation systems), the parameters of the improved dead reckoning algorithm are optimized and adjusted according to the data interaction situation and the performance indicators of the entire simulation system;

[0017] Step 6: Loop execution

[0018] Repeat steps 2 to 5. During the entire distributed simulation process, each simulation time step continues to perform entity state acquisition, local dead reckoning, error calculation, data transmission decision, and position synchronization operations until the simulation task is completed.

[0019] Each simulation node obtains the initial state data of the entity it is responsible for simulating, including position coordinates, velocity vector, acceleration vector, and posture information.

[0020] The dead reckoning algorithm predicts the future state of an entity. First, the kinematic models of different entities are combined to constrain and correct the dead reckoning. Second, a dynamic adaptive adjustment mechanism is introduced to adjust the parameters of the prediction model in real time according to the changing trend of the entity's motion during the prediction process.

[0021] The changing trend of the entity movement includes a sudden change in speed and a change in direction.

[0022] The multi-dimensional errors include position errors, velocity, acceleration, and attitude.

[0023] The position error is the deviation between the estimated position and the actual position calculated by the Euclidean distance formula.

[0024] The speed error is the difference between the estimated speed and the actual speed.

[0025] The heading error is the difference between the calculated estimated heading and the actual heading.

[0026] The method for optimizing and adjusting the parameters of the improved dead reckoning algorithm; if the network delay is low and the data packet loss rate is almost zero, the prediction step size is appropriately increased to reduce the number of predictions and the possibility of data interaction; conversely, if the network condition is poor or there is a large deviation in simulation synchronization, the prediction step size is reduced, and the parameters of the prediction model are finely adjusted. At the same time, the parameter information and related performance data of this adjustment are fed back to other simulation nodes so that each node can collaboratively optimize the algorithm parameters to achieve adaptive optimization of the entire distributed simulation system.

[0027] Innovation and advantages

[0028] The method of reducing the amount of distributed simulation data transmission based on the dead reckoning algorithm of the present invention can make full use of the changing laws and predictability of entity motion, effectively reduce unnecessary data transmission, reduce network load, and at the same time ensure the accuracy and stability of distributed simulation, provide strong technical support for large-scale, high-precision distributed simulation applications, and have broad application prospects and significant technical advantages. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 The present invention is a flow chart of a method for reducing the amount of distributed simulation data transmission based on the dead reckoning algorithm. DETAILED DESCRIPTION

[0030] The present invention is described in detail below in conjunction with an aviation distributed simulation scenario:

[0031] like Figure 1 As shown, the method for reducing the amount of distributed simulation data transmission based on the dead reckoning algorithm of the present invention comprises the following steps:

[0032] Step 1: Entity state data acquisition and preprocessing

[0033] When the distributed simulation system is started, each simulation node obtains the initial state data of the entity it is responsible for simulating, including but not limited to position coordinates, velocity vectors, acceleration vectors, attitude information, etc., and standardizes and formats these data for efficient processing by subsequent algorithms. At the same time, according to the characteristics and accuracy requirements of the simulation scene, the initial parameters of the dead reckoning algorithm are set, such as the basic prediction step size, the initial error tolerance range, etc.

[0034] Step 2: Local Dead Reckoning

[0035] Each simulation node uses an improved dead reckoning algorithm to predict the future state of the entity. The improvements are: first, the kinematic models of different entities are combined to constrain and correct the dead reckoning. For example, based on the aerodynamic characteristics of the aircraft, engine performance and other factors, the aircraft's motion equations and constraints are established. During the dead reckoning process, these model information is used to optimize and adjust the calculation results so that the calculated aircraft position and attitude are more consistent with the actual flight situation. Secondly, a dynamic adaptive adjustment mechanism is introduced. During the prediction process, the parameters of the prediction model are adjusted in real time according to the changing trend of the entity's motion (such as sudden changes in speed, changes in direction, etc.) to improve the accuracy and adaptability of the prediction. For example, when it is detected that the rate of change of the entity's speed exceeds a certain threshold, the weight coefficient of the acceleration factor in the prediction model is automatically increased to better capture the entity's motion changes.

[0036] Step 3: Dead Reckoning Error Calculation

[0037] After completing local dead reckoning, calculate the error between the locally predicted entity state data and the entity state data actually calculated by the high-precision simulation model. Use multi-dimensional error evaluation indicators, not only consider the position error, but also comprehensively consider the errors in speed, acceleration, attitude, etc. Error calculation can be done in a variety of ways, such as calculating the position error (the deviation between the estimated position and the actual position can be calculated by the Euclidean distance formula), the speed error (the difference between the estimated speed and the actual speed), and the heading error (the difference between the estimated heading and the actual heading), etc., and assign corresponding weights according to the importance of different errors to comprehensively obtain a total dead reckoning error value.

[0038] Step 4: Data transfer decision

[0039] Compare the error of each dimension with the preset error tolerance range. If the errors of all dimensions are within the tolerance range, the current prediction result is determined to be valid, and there is no need to transmit data to other simulation nodes. Only the prediction data is recorded locally for the continuity calculation of subsequent local simulations. If the error of at least one dimension exceeds the tolerance range, it indicates that there is a large deviation in the dead reckoning result. At this time, the complete and accurate state data of the current entity needs to be compressed and packaged, and sent to other related simulation nodes according to the communication protocol of the distributed simulation system. After receiving the data, the receiving node immediately updates the corresponding entity state in the local simulation model to ensure that each node in the entire distributed simulation system has a consistent understanding of the entity state, thereby maintaining the accuracy and coherence of the simulation.

[0040] Step 5: Algorithm parameter optimization

[0041] After each data transmission is completed (i.e. when the error exceeds the tolerance) or after a certain number of local prediction cycles, the parameters of the improved dead reckoning algorithm are optimized and adjusted according to the data interaction situation and the performance indicators of the entire simulation system (such as network delay, data packet loss rate, overall simulation synchronization, etc.). For example, if the network delay is low and the data packet loss rate is almost zero, the prediction step size can be appropriately increased to reduce the number of predictions and the possibility of data interaction; on the contrary, if the network condition is poor or there is a large deviation in simulation synchronization, the prediction step size is reduced, and the parameters of the prediction model are finely adjusted, such as adjusting the fitting coefficients and weight coefficients in the model, to improve the prediction accuracy and reduce the frequent data interaction caused by excessive errors. At the same time, the parameter information and related performance data of this adjustment are fed back to other simulation nodes so that each node can collaboratively optimize the algorithm parameters to achieve adaptive optimization of the entire distributed simulation system.

[0042] Step 6: Loop execution

[0043] Repeat steps 2 to 5. During the entire distributed simulation process, each simulation time step continues to perform entity state acquisition, local dead reckoning, error calculation, data transmission decision, and position synchronization operations until the simulation task is completed. Through this cyclic iterative method, the dead reckoning results are continuously optimized, and the data transmission strategy is dynamically adjusted to minimize the amount of data transmission while ensuring the simulation accuracy.

[0044] Each simulation node obtains the initial state data of the entity it is responsible for simulating, including position coordinates, velocity vector, acceleration vector, and attitude information.

[0045] Example

[0046] In the aviation distributed simulation system, there are multiple simulation nodes, which simulate different aircraft (such as civil drones, passenger planes, etc.) and related facilities such as ground command centers and radar stations. When the system starts, each node obtains the initial position (longitude, latitude, altitude), speed (horizontal speed, vertical speed), acceleration (linear acceleration, angular acceleration) and attitude (pitch angle, yaw angle, roll angle) of the simulated aircraft, and sets the initial parameters of the dead reckoning algorithm, such as the prediction step length T0 = 0.5 seconds, the position error threshold ε p = 100 meters, speed error threshold ε v =5 m / s, acceleration error threshold ε a =0.5 m / s2, attitude error threshold ε q =2°.

[0047] Taking a simulated civilian drone as an example, it is initially in a uniform straight-line flight state. The simulation node determines to use the uniform straight-line motion sub-prediction model for local prediction based on its historical motion data characteristics. During the prediction process, a prediction is made every 0.5 seconds, and the error between the predicted position, velocity, acceleration, and attitude and the corresponding parameters actually calculated by the high-precision flight mechanics model is calculated. If the prediction errors for multiple consecutive times are within the set threshold range, there is no need to transmit data to other nodes (such as ground command center nodes, other aircraft nodes, etc.).

[0048] However, when a civilian UAV performs maneuvers such as sharp turns, dives, or accelerated climbs, the prediction error may quickly exceed the threshold range. At this time, the simulation node packages the current accurate position, speed, acceleration, attitude and other state data of the civilian UAV and sends them to the relevant nodes. After receiving the data, other nodes update the local civilian UAV simulation model state to ensure that the entire simulation system simulates the state of the civilian UAV consistently.

[0049] During the entire simulation process, the system will dynamically adjust the dead reckoning algorithm parameters according to the network conditions and simulation synchronization. For example, when the network bandwidth is sufficient and the simulation synchronization is good, the prediction step length T can be increased to 1 second; when the network is congested or the simulation has obvious delays, the prediction step length T is reduced to 0.3 seconds, and the weighted coefficients of multi-model fusion are optimized to improve the prediction accuracy of complex maneuvers of civil drones and reduce the transmission of large amounts of data caused by excessive errors, thereby improving the operating efficiency and simulation accuracy of the entire aviation distributed simulation system.

[0050] It can be seen from the above embodiments that the method for reducing the amount of distributed simulation data transmission based on the dead reckoning algorithm proposed in the present invention can effectively reduce unnecessary data interactions in distributed simulation systems, improve system operation efficiency, and at the same time ensure the accuracy and stability of the simulation, and has broad application prospects and significant technical advantages.

Claims

1. A method for reducing the amount of distributed simulation data transmission based on a dead reckoning algorithm, characterized in that: The following steps are involved: Step 1: Entity state data acquisition and preprocessing When the distributed simulation system is started, each simulation node obtains the initial state data of the entity it is responsible for simulating, and standardizes and formats the data. At the same time, the initial parameters of the dead reckoning algorithm are set according to the characteristics and accuracy requirements of the simulation scenario. Step 2: Each simulation node uses an improved dead reckoning algorithm to predict the future state of the entity; Step 3: Dead Reckoning Error Calculation After completing local dead reckoning, the error between the locally predicted entity state data and the entity state data actually calculated by the high-precision simulation model is calculated. The position error, velocity error, and heading error are calculated using a multi-dimensional error evaluation index. The corresponding weights are assigned according to the importance of different errors to obtain a total dead reckoning error value. Step 4: Data transfer decision Compare the error of each dimension with the preset error tolerance range. If the errors of all dimensions are within the tolerance range, the current prediction result is determined to be valid. There is no need to transmit data to other simulation nodes. The prediction data is only recorded locally for the continuity calculation of subsequent local simulations. If the error of at least one dimension exceeds the tolerance range, it indicates that the dead reckoning result has a large deviation. At this time, the complete and accurate state data of the current entity needs to be compressed and packaged, and sent to other related simulation nodes according to the communication protocol of the distributed simulation system. After receiving the data, the receiving node immediately updates the corresponding entity state in the local simulation model to ensure that each node in the entire distributed simulation system has a consistent understanding of the entity state, thereby maintaining the accuracy and consistency of the simulation; Step 5: Algorithm parameter optimization After each data transmission is completed or after N local prediction cycles (N can range from 1 to 100 and can be adjusted according to the accuracy required by different simulation systems), the parameters of the improved dead reckoning algorithm are optimized and adjusted according to the data interaction situation and the performance indicators of the entire simulation system; Step 6: Loop execution Repeat steps 2 to 5. During the entire distributed simulation process, each simulation time step continues to perform entity state acquisition, local dead reckoning, error calculation, data transmission decision, and position synchronization operations until the simulation task is completed.

2. The method for reducing the amount of distributed simulation data transmission based on the dead reckoning algorithm according to claim 1 is characterized in that: Each simulation node obtains the initial state data of the entity it is responsible for simulating, including position coordinates, velocity vector, acceleration vector, and posture information.

3. The method for reducing the amount of distributed simulation data transmission based on the dead reckoning algorithm according to claim 1 is characterized in that: The dead reckoning algorithm predicts the future state of an entity. First, the kinematic models of different entities are combined to constrain and correct the dead reckoning. Second, a dynamic adaptive adjustment mechanism is introduced to adjust the parameters of the prediction model in real time according to the changing trend of the entity's motion during the prediction process.

4. The method for reducing the amount of distributed simulation data transmission based on the dead reckoning algorithm according to claim 1, characterized in that: The changing trend of the entity movement includes a sudden change in speed and a change in direction.

5. The method for reducing the amount of distributed simulation data transmission based on the dead reckoning algorithm according to claim 1, characterized in that: The multi-dimensional errors include position errors, velocity, acceleration, and attitude.

6. The method for reducing the amount of distributed simulation data transmission based on the dead reckoning algorithm according to claim 1, characterized in that: The position error is the deviation between the estimated position and the actual position calculated by the Euclidean distance formula.

7. The method for reducing the amount of distributed simulation data transmission based on the dead reckoning algorithm according to claim 1, characterized in that: The speed error is the difference between the estimated speed and the actual speed.

8. The method for reducing the amount of distributed simulation data transmission based on the dead reckoning algorithm according to claim 1, characterized in that: The heading error is the difference between the calculated estimated heading and the actual heading.

9. The method for reducing the amount of distributed simulation data transmission based on the dead reckoning algorithm according to claim 1, characterized in that: The method for optimizing and adjusting the parameters of the improved dead reckoning algorithm; if the network delay is low and the data packet loss rate is almost zero, the prediction step size is appropriately increased to reduce the number of predictions and the possibility of data interaction; conversely, if the network condition is poor or there is a large deviation in simulation synchronization, the prediction step size is reduced, and the parameters of the prediction model are finely adjusted. At the same time, the parameter information and related performance data of this adjustment are fed back to other simulation nodes so that each node can collaboratively optimize the algorithm parameters to achieve adaptive optimization of the entire distributed simulation system.

Images