CN102279568A - Data control method used for formation flying - Google Patents

Abstract

The invention discloses a data control method for formation flight and belongs to the technical field of simulation. The method consists of pilot behavior collection, behavior prediction and evaluation, behavior data control, control state switching and so on. The present invention adopts the idea of feedback control, takes the "interval", "distance", "height difference" and the "maneuvering" of the formation flight as the control targets, and adopts the " The data control of maneuvering, when the next "maneuvering" is not predicted, the automatic control oriented to "separation", "distance" and "altitude difference" is adopted. The invention can effectively improve the authenticity of formation flight simulation, and has direct reference value for both military and civilian simulation fields.

Description

A data control method for formation flight

technical field

The invention belongs to the technical field of simulation and comes from the engineering practice of Computer Generated Forces (CGFs, computer generated forces). The invention can be used in both the field of military simulation and the field of civilian simulation.

Background technique

In CGFs engineering practice, the effect of simulating aircraft formation is not that the smaller the error, the better, but the more realistic the better. Because pilots will inevitably produce errors when flying an aircraft, effectively reflecting such errors is a manifestation of "anthropomorphism". In other words, the formation flight simulation has certain errors, indicating that it is more like the effect of real pilots driving. And these "real errors" can only come from the collection of real pilots' driving behavior. We invented a formation flight simulation method based on "data control" to take advantage of this feature, that is, save the pilot's driving behavior in a data file for future use in the same situation.

On the other hand, for a dynamical system, a set of input vector sequences (as a function of time t) will produce a set of output effects. If we "play by fruit", save the sequence of input vectors that produces the "perfect" effect. Then, in a specific time period, using the saved input vector sequence to control the original dynamic system will produce the effect we want. This is another motivation and basis for our invention of formation flight simulation based on "data control".

The formation flight simulation based on "data control" needs the interaction of two states "automatic control state" and "data control state". During formation flight, the "prediction and evaluation" module is used to predict the behavior of the lead aircraft (maneuver to be performed). The "prediction and evaluation" module works continuously. When the next maneuver of the lead plane is correctly predicted, it enters the "data control state", otherwise, it enters the "automatic control state". In the "data control state", the aircraft is controlled by the control quantity pre-stored in the data file; in the "automatic control state", the aircraft is controlled by the automatic control program.

Compared with "automatic control", "data control" has the following advantages and disadvantages. Advantages: (1) High accuracy of data control, suitable for complex aerobatic flight simulation; (2) High anthropomorphic data control, suitable for CGFs simulation; (3) Good reusability of data control, can reuse pre-stored data file. Disadvantages: (1) The flexibility of data control is poor, and only predetermined maneuvers can be performed; (2) The robustness of data control is poor, and it is only suitable for simulation reproduction.

The formation flight method based on "data control" invented by us is a method combining "data control" and "automatic control", which is suitable for CGFs simulation engineering.

For example, let the lead plane fly freely first, and the wingman pilot drives the wingman to follow the lead plane to fly in formation; during the formation flight, record the driving behavior of the wingman pilot and save it in a data file. Then, let the lead plane perform scheduled maneuvers, and also let the wingman pilots fly in formation and record their driving behavior. Finally, under the CGFs simulation environment, let the wingman follow the lead plane to fly under the control of the automatic control program and data files. Among them, when performing predetermined special formation maneuvers, it is controlled by a pre-saved data file; when performing conventional formation flight, it is controlled by an automatic control program.

In conclusion, this method of data control for formation flight is a valuable approach. The present invention has the following characteristics: (1) high anthropomorphism, by saving the pilot's driving behavior in the data file, and then controlling the aircraft formation flight by the data file is the reproduction of the pilot's driving behavior; (2) the method is easy to implement, "data file "control method" involves the establishment, preservation, reading and writing of data files, so it is easy to realize through programming; (3) the connection is good, it is easy to connect the data control method and the automatic control method, so that the data control method "controls" Special maneuvering formation flight, automatic control method to control conventional formation flight; (4) It has wide application value in simulation exercises, not only for small-scale tactical exercises but also for large-scale interactive simulation exercises.

Contents of the invention

The invention discloses a data control method for formation flight, which is derived from CGFs engineering practice. This method improves the traditional formation flight simulation method, and controls the whole process of aircraft formation flight simulation by combining automatic control and data control. This method is of great value in improving the authenticity of CGFs simulation.

The traditional formation flight simulation method is a formation simulation method completely based on automatic control. The traditional method regards the aircraft participating in the formation (such as the lead plane and the wingman) as the controlled objects, regards the interval, distance and height difference between the formation aircraft as the target of control, and designs the control law to maintain the predetermined interval, distance and height difference. However, the traditional method has the following problems: (1) The formation flight maneuver has a predetermined action essentials, and there is no need to maintain the formation through deviation control; (2) Some difficult maneuvers or special maneuvers can easily cause the system to diverge, Cause deviation control to be ineffective, for example, for serpentine maneuvers, it is difficult to maintain formation through deviation control; (3) Formation flight performed by real pilots is not to follow the team through the idea of "deviation control", but to do it at the same time For scheduled maneuvers, try to (but not necessarily) consider formation maintenance during the maneuver. We improved the traditional formation flight simulation method and invented a formation flight simulation method based on "data control". This method combines the automatic control method and the "data control method" closely, and gives full play to the advantages of both. That is, under normal circumstances, the formation formation is maintained through the automatic control program; and when performing predetermined "difficult" maneuvers (predetermined special formation maneuvers), the formation formation is maintained through data control.

This method is described in detail below.

1. Evaluation of prediction cost

The evaluation of prediction cost includes two aspects: one is that the number of correct predictions is greater than the specified threshold; the other is that the cost of prediction is less than the specified threshold.

(1) The number of correct predictions is greater than the specified threshold

Correct prediction means that the wingman correctly predicts the behavior of the lead plane at the next moment according to the situation of the lead plane in the most recent period. Since it is a prediction of behavior that has not occurred, the accuracy of the prediction is very important. In order to evaluate the accuracy of the prediction, we introduced the "threshold of the correct number of predictions" - minCorrectness. Therefore, the forecast count evaluation formula is:

{NumCorrPredict} _{t ∈ [ti, tj]} ≥ minCorrectness

where {NumCorrPredict} _{t ∈ [ti, tj]} is the actual number of predictions between time t _i and t _j . t _i and t _j are any two time points. minCorrectness is the pre-specified threshold of correct prediction times.

When we specify the forecast period, we can calculate the actual number of forecasts during this period.

(2) The predicted cost is less than the specified threshold

The predicted cost function is built from the relationship between the current state of affairs and the desired state of affairs.

The calculation formula of the predicted cost function is:

J＝(1/n)*∑ _i＝1 ⁿ [W _i -Ave(W)] ²

Among them, W _i ∈ {W ₁ , W ₂ ,…, W _n }, W=<w _x , w _y , w _z , w _v , w _theta , w _gamma , w _psi >, n represents the time point of prediction N times of sampling calculations are carried out consecutively.

in addition,

w _x ＝x _cur -x _des ; (x _cur is the current coordinate of x, x _des is the desired x coordinate value)

w _y = y _cur -y _des ; (y _cur is the current coordinate of y, y _des is the expected y coordinate value)

w _z ＝z _cur -z _des ; (z _cur is the current coordinate of z, z _des is the expected z coordinate value)

w _v ＝v _cur -v _des ; (v _cur is the current value of speed v, v _des is the desired speed value)

w _theta = θ _cur - θ _des ; (θ _cur is the current value of the pitch angle θ, θ _des is the expected value of the pitch angle)

w _gamma = γ _cur - γ _des ; (γ _cur is the current value of roll angle γ, γ _des is the expected value of roll angle)

w _psi ＝ψ _cur -ψ _des ; (ψ _cur is the current value of yaw angle ψ, ψ _des is the desired yaw angle value)

If (1) n sampling calculations are performed at a certain time point; (2) the threshold of the cost function is specified as minCost, then the method of calculating the prediction cost is:

First, calculate the value of W ₁ , W ₂ ... W _n ;

Then, calculate Ave(W), that is, calculate the average value of W ₁ , W ₂ ... W _n ;

Next, calculate the value of the cost function J;

Finally, compare the value of J with minCost, if J≤minCost, the predicted cost is "extremely small"; otherwise, the predicted cost is too high.

To sum up, if the number of correct predictions is greater than the specified threshold and the cost of prediction is less than the specified threshold, the prediction is considered valid and put into practice, otherwise, the prediction is considered invalid and remedial adjustments are made immediately.

2. Data control method

When the prediction is valid, enter the data control stage.

Data control refers to a technology applied in flight simulation. It saves the control quantity of the aircraft in a data file in advance, and in some cases directly reads the control quantity from the data file to control the aircraft.

The maneuvers involved in the data control method are: "right pull up", "left pull up", "combat turn", "fall", "high speed yaoyao", "low speed yaoyao", "half catty flip", "half roll Reversal", "sharp circling", "accelerating turn", "level flight increasing" (as shown in Table 1).

Table 1 is an illustration of the maneuvers involved in the data control method.

Table 1. The maneuvers involved in the data control method

In Table 1, 11 maneuvers commonly used in data control methods are listed.

The basic control quantities involved in the data control method are: "throttle stick control quantity", "elevator control quantity", "rudder control quantity", "aileron control quantity", that is, what is commonly called "two levers and one rudder" control information.

Table 2 lists the meanings of the basic control quantities involved in the data control method.

Table 2 Description of the basic control quantities involved in the data control method

In Table 2, the four basic control quantities involved in the data control method are listed. Through the regular assignment of these four kinds of control quantities, the aircraft can be kept at a predetermined attitude, thereby completing various maneuvers.

For the 11 kinds of maneuvers commonly used in data control methods, the following methods are adopted to collect control quantities:

(1) Select the lead pilot and wingman pilot. Wingmen are flown by experienced pilots.

(2) For a specific maneuver, such as a "combat turn", let the lead aircraft lead the flight.

(3) Place a recording device in the wingman's control system. This device is used to record the amount of control the pilot is manipulating the aircraft.

(4) The wingman pilot drives the aircraft to follow the lead aircraft to perform a certain maneuver and maintain formation.

(5) Record the control amount of the wingman pilot and save it in the data file.

Notice:

(1) The above process can also be carried out on the simulator.

(2) The lead plane can also be automatically controlled by software (unmanned driving).

(3) The control volume of the wingman pilot is stored in the data file according to a certain format, which is usually arranged in the order of "throttle stick control volume", "elevator control volume", "rudder control volume", and "aileron control volume" .

Figure 1 shows the principle of the data control method.

In Figure 1, "prediction and evaluation", "controller", "aircraft object" and "data file" together constitute the "data control system". Among them, "prediction and evaluation" is a module that predicts the next maneuver to be performed; "controller" is a module that performs flight control on the aircraft object and completes switching between conventional control methods and data control methods; "aircraft "object" is a six-degree-of-freedom aircraft dynamics model used for flight simulation; "data file" is a file of control quantities related to maneuvering stored in advance.

Under normal circumstances, the "switch" on the right side of the "controller" is in the No. 1 switch position. At this point, the "aircraft object" is in normal flight under the control of the "controller". When performing a predetermined maneuver, the "controller" controls the "switch" on the right side to switch on the No. 2 switch position, and enters the "data control" state at this time, that is, the flight control of the aircraft is stored in the "data file" to control the amount of control in .

The "forecast and evaluation" module works continuously, and its working principle is: to make continuous predictions based on the current situation, and the cost of the predictions is evaluated by the aforementioned method. Right now,

(1) The number of correct predictions NumCorrPredict is greater than or equal to minCorrectness;

(2) The value of prediction cost J (1/n)*∑ _i=1 ⁿ [W _i -Ave(W)] ² is less than or equal to minCost;

When (1) and (2) are satisfied, it can be determined that the data control state should be entered.

3. Flying in formation

Formation flight refers to the flight of two or more aircrafts forming a certain formation. Flying in formation requires a certain separation, distance and height difference between aircraft.

In the present invention, we mainly consider that two aircraft maintain a predetermined separation, distance and altitude difference while performing the same maneuver.

Fig. 2 is a schematic diagram of "separation", "distance" and "altitude difference" of two aircrafts flying in formation.

In formation flying, the leading aircraft is called the "lead aircraft". The "lead plane" leads the other planes on a predetermined maneuver.

In formation flying, the aircraft that follows the lead aircraft is called the "wingman". The wingman follows the lead, that is, the wingman performs the same maneuver as the lead. In flight simulation, since the lead and wingman are not driven by real pilots, they cannot conduct a unified negotiation before maneuvering. Therefore, the "wingman" needs to predict and evaluate the behavior of the lead plane to achieve the purpose of formation flight, that is, the wingman It is necessary to predict the next maneuver of the lead plane and carry out corresponding control according to the predicted results. When the lead plane's next maneuver is correctly predicted, it enters the "data control" state, otherwise it enters the automatic control state (also known as the "program control" state).

In order to make the effect of formation flight simulation both authentic and perfect, it is necessary to "control" through data control methods. First, save the driving behaviors of various maneuvers performed by the real pilot (that is, the control amount of two sticks and one rudder); then, let the wingman keep predicting the maneuver of the lead plane, and when the prediction is "correct", call the pre-stored the control quantities stored in the file; the aircraft then performs a flight simulation under the control of the control quantities stored in the data file.

Whether the wingman predicts "correctly" depends on (1) the number of correct predictions NumCorrPredict is greater than or equal to the specified threshold minCorrectness; (2) the value of the prediction cost J is less than or equal to the specified threshold minCost.

The formation flight simulation based on "data control" makes the aircraft fly in two states, one is the automatic control state, and the other is the data control state. Among them, the automatic control state is used for routine control of the aircraft, and its applicable scope includes: non-maneuvering flight, "adjustment" control after wrong prediction, and mechanical follow-up control; the data control state is used for formation stunts of the aircraft Flying, its scope of application is the anthropomorphic control after the prediction is correct. The automatic control state is controlled by the control program, also known as "program control"; the data control state is controlled by the pre-stored control volume file.

In flight simulation, formation flight not only requires perfect formation, but also requires formation to be authentic and anthropomorphic. That is to say, in flight simulation, the smaller the error, the better the formation formation is not (including interval difference, distance difference and height difference), but the more realistic the effect of formation flight and the more like the effect of real-life driving, the better. Due to the particularity of formation flight simulation, the wingman is required to predict the flight behavior of the lead plane, and when the prediction is "correct", the data file with the pre-stored control quantity is called for data control.

Based on the above discussion, the formation flight simulation based on the "data control" method can be expressed by the following "solution" process:

for each<x,t>∈D{

S: <x, t> → <y, t>;

Output Y;

}

Among them, D is a data file, that is, a text file that saves control quantities; <x, t> is an element in D, and x is a control quantity vector ("throttle stick control quantity", "elevator control quantity", "rudder control quantity" ", "aileron control value"), t is time, that is to say, a series of control values are stored in D; S is the controlled system; <y, t> is the control value The output value of S under the action of <x, t>; Y is the set (vector) of the output value <y, t>.

When X, Y, and t are the input vector, output vector and time of the controlled system S respectively, the solution process of S is a "multi-step solution process", that is,

If(X←<x ₁ , t ₁ >, <x ₂ , t ₂ >…<x _i , t _i >…)then

Y←<y ₁ ，t ₁ >，<y ₂ ，t ₂ >…<y _i ，t _i >…

Among them, <x ₁ , t ₁ >, <x ₂ , t ₂ >…< _xi , t _i >… are the values of the input vector X (with time t); <y ₁ , t ₁ >, <y ₂ , t ₂ >…<y _i , t _i >… are the values of the output vector Y over time t.

For each set of x ₁ , x ₂ ... _xi ..., we can find a set of corresponding y ₁ , y ₂ , ...y _i ... and save them in a data file for future use in similar situations.

The formation flight simulation based on "data control" is different from the formation flight simulation based on the pure automatic control method. Because the formation flight simulation based on automatic control pursues the stability, speed and accuracy of the control. The goal of this pursuit may be contrary to the real situation. That is to say, the formation flight simulation effect displayed by the automatic control method cannot be achieved by real pilots. All in all, the formation flight simulation based on automatic control lacks "anthropomorphism".

We organically combine the "data control" method with the automatic control method, that is, under normal circumstances, the aircraft flies under the action of the automatic control program. The method of "data control" is used to "control" the aircraft.

From the perspective of solving dynamic equations, the difference between the "data control" method and the automatic control method is that: the "data control" method periodically executes "<x,t>∈D", and D is the data file which saves A series of control values are given; the automatic control method does not implement "<x, t>∈D", its <x, t> is periodically calculated by the formula (control law).

Therefore, the formation flight simulation method based on "data control" proposed by the present invention is a joint control method, which supervises and controls conventional formation flight missions through automatic control methods, and performs predetermined tasks through "data control" methods. Special formation maneuvers, such as aerobatic formation flight simulations.

Method steps of the present invention are as follows:

(1) Let the lead plane fly freely;

(2) The wingman is driven by the pilot;

(3) Record the driving behavior of the wingman pilot (save the control value of the pilot's aircraft);

(4) Let the lead aircraft perform the scheduled maneuver and repeat the steps (1)-(3) above;

(5) In the CGFs simulation environment, start the lead flight routine;

(6) Let the wingman carry out automatic team flight, that is, let the wingman carry out formation flight simulation;

(7) The "prediction and evaluation" module of the wingman performs calculations continuously, and when it is predicted that the lead plane is about to perform a scheduled special formation maneuver, it performs "data control", otherwise it performs automatic control;

(8) When the wingman makes a mistake in judging the future maneuver of the lead plane, use the automatic control method to correct the current state;

FIG. 3 is a scheme diagram of a data control method.

In the above steps, the simulation environment is used as an example to design the method. For example, the pilot driving the wingman means that the pilot controls the wingman on the simulator. The above method steps can also be used for real manipulation of non-simulated environments.

Description of drawings

The principle, method flow and related concepts of the present invention are described below with reference to the accompanying drawings, wherein:

Figure 1 is a schematic diagram of the data control method used in formation flight simulation

Figure 2 is a schematic diagram of "separation", "distance" and "height difference" in formation flight

Figure 3 is a scheme diagram of the data control method

Detailed ways

The "a data control method for formation flight" described in the present invention will be further described below in conjunction with the accompanying drawings.

"A Data Control Method for Formation Flight" starts from the requirement of real behavior model in CGFs simulation engineering, adopts the method of saving the pilot's driving behavior in the data file and using it to control the aircraft formation flight, and carries out the "data Control" method design. The method includes an automatic control part and a data control part, wherein the former is used for conventional control of the aircraft; the latter is used for the control of the formation aerobatic flight of the aircraft. This method involves "acquisition of pilot driving behavior data", "format of data file", "preservation of data file", "reading and control of data file", "prediction and evaluation of lead aircraft behavior", "automatic control and Switching of data control", etc. The specific implementation is as follows:

First, collect the pilot's driving behavior.

The collection of pilot driving behavior refers to the use of certain equipment and algorithms to record the pilot's control of the aircraft. Specifically, in this method, we record the throttle stick control value, the elevator control value, the rudder control value, and the aileron control value for the pilot to fly the aircraft. A set of data is recorded at each time point (sampling time point). The values of these control quantities are saved in the data file in the order of "throttle stick control quantity", "elevator control quantity", "rudder control quantity" and "aileron control quantity". The time period for collecting these data is the entire process of the aircraft doing a certain predetermined maneuver. The specific method steps for collecting the pilot's driving behavior are:

(a) Make the lead aircraft fly according to the intended maneuver;

(b) The pilot flies in formation with a wingman;

(c) Record pilot behavior throughout the procedure and identify it with the intended maneuver.

The novelty of our method is that the lead plane can use software to simulate the flight, and the wingman can also be driven by the pilot on the simulator, and the recording process of the pilot's behavior is completed by the software.

Second, carry out the free flight of the lead plane.

The free flight of the lead plane means that the lead plane flies according to the mission. Flight missions generally include level flight, turning, tactical maneuvers, aerobatics, etc. In a flight mission, all maneuvers are notified in advance to the aircraft participating in the formation flight, but the sequence of maneuvers and the exact time to enter the maneuver are uncertain, and these are controlled (decided) by the lead aircraft in real time. Therefore, the wingman needs to predict the behavior of the lead plane.

Next, the wingman predicts the behavior of the lead plane.

The basis for the wingman to predict the lead plane is:

(a) Calculate the number of correct predictions NumCorrPredict, and determine whether it is greater than or equal to minCorrectness;

(b) Calculate the value of the predicted cost J (1/n)*∑ _i=1 ⁿ [W _i -Ave(W)] ² , and judge whether it is less than or equal to minCost;

Forecasting is done by the "Forecasting and Estimation" module.

When the wingman's prediction of the leader's behavior is correct, it is "controlled" by the "control sequence" stored in the "data file" and marked with maneuvering; otherwise, the flight of the wingman is controlled by the automatic control program.

That is, the wingman flies under the "control" of the two control methods.

Finally, modify the "data control file" according to the effect of formation flight.

On the surface, "data control" is a playback of simulation effects, but in fact it is the playback of the pilot's driving control values. These control values are used to control the aircraft to achieve the predetermined flight effect, thereby indirectly achieving the playback of simulation effects. Therefore, the formation flight simulation based on "data control" needs iterative correction. The goal of the revision is to make formation flight closer to the driving effect of real pilots and meet the simulation requirements.

It should be noted that, in addition to supporting CGFs simulation projects, the present invention also supports various other occasions where "simulation playback" is used. The "data control" method mentioned in the present invention is a "playback" based on the driving behavior of real pilots, which is more realistic and practical than the general "simulation playback" technology. Through the mutual switch between the automatic control part and the data control part in the "data control" method, the continuity and authenticity of the formation flight effect are guaranteed, thereby improving the practicability of the simulation project.

Claims (6)

1. A data control method for formation flight, characterized in that: the "two rods and one rudder" control quantity of the pilot driving the aircraft is stored in a data file, and these control values are read from the data file under the same conditions Similar controls are implemented based on this method; the implementation system based on this method consists of four parts: pilot behavior collection [1], behavior prediction and evaluation [2], behavior data control [3], and control state switching [4]: Pilot behavior collection[1]: refers to letting real pilots drive the aircraft to follow the lead aircraft to fly in formation, and collect the "two rods and one rudder" control information of the pilot flying the aircraft in the process; Behavior prediction and evaluation[2]: It refers to predicting the maneuvers of the lead aircraft, and making statistics and evaluation of historical behavior; Behavioral data control [3]: refers to the control of the simulation model of the aircraft by using the behavioral data files of the pilot driving the aircraft; Switching of the control state [4]: refers to switching between the state controlled by the automatic control method and the state controlled by the "data control" method. 2. A kind of data control method that is used for formation flight according to claim 1, is characterized in that, "two sticks and one rudder" refers to throttle stick, control stick and rudder; "two sticks and one rudder" control amount comprises " throttle control”, “elevator control”, “aileron control” and “rudder control”. 3. A data control method for formation flight according to claim 1, characterized in that: when performing pilot behavior collection [1], first, let the lead plane fly according to a predetermined "maneuver", wherein "maneuver "Refers to the general term for a series of coherent flight actions that can complete a certain function; then, let the pilot drive the wingman to follow the lead plane to fly in formation; "Control volume information. 4. A data control method for formation flight according to claim 1, characterized in that: Behavior prediction and evaluation [2] refers to predicting the maneuvers of the lead aircraft and performing statistics on historical prediction behaviors Among them, the flight of the lead plane is carried out according to the mission, and a mission includes takeoff, climb, level flight, tactical maneuver, return, landing, etc.; the prediction of the lead plane’s maneuver is limited to the upcoming tactical maneuver of the lead plane (including stunts) to make predictions; the prediction involves two indicators of "number of correct predictions" and "prediction cost", among which: "Number of correct predictions" refers to the number of times the wingman correctly predicted the follow-up actions of the lead plane based on the recent situation of the lead plane; "Predicted cost" refers to a specific value calculated according to the predicted cost formula, and the predicted cost formula is a function of sampling times, x coordinate, y coordinate, z coordinate, speed, pitch angle, roll angle and yaw angle. 5. A kind of data control method that is used for formation flight according to claim 1, is characterized in that: the control [3] of behavior data refers to utilizing the behavior data file of the pilot to drive the aircraft to control the simulation model of the aircraft; The behavior data of flying an aircraft is obtained by "Pilot Behavior Collection" [1] and saved as a data file; when controlling behavior data, it is necessary to first perform "Behavior Prediction and Evaluation" [2], and then read from the data file Take the flight control amount and implement the control on the aircraft model. 6. A kind of data control method for formation flight according to claim 1, characterized in that: "switching of control state" refers to the state controlled by the automatic control method and the state controlled by the "data control" method "Switching of control state" is the guarantee for the continuous and effective flight of the aircraft. When the aircraft is in normal flight, the automatic control method is used for control, and when the aircraft performs special formation maneuvering, the "data control" method is used for control; The timing of switching between automatic control and "data control" is whether the result of "behavior prediction and evaluation" [2] meets the preset conditions.

Images