RU2487052C1 - Method of generating drone stabilisation system control signal - Google Patents

Abstract

FIELD: aircraft engineering.

SUBSTANCE: proposed method consists in generation of control signal input into stabilisation unit by its calculation on differential game basis. Said control signal is calculated proceeding from analysis of location of minimax region drone attainability domain constructed with due allowance for all possible disturbances satisfying preset constraints and preset normal overload magnitude.

EFFECT: control signal with compensated constrained disturbances with unknown statistical properties.

5 dwg

Description

The invention relates to the field of rocket science, and in particular to methods for controlling an UAV under external disturbances with unknown statistical properties, leading to a change in the trajectory of movement. It can be used in the instrumentation of the UAV.

It is known that the dynamic properties of UAVs without an onboard stabilization system do not provide the required control accuracy (Design of anti-aircraft guided missiles / Edited by I.S. Golubev and V.G. Svetlov. - Second ed., Revised and enlarged. - M .: Publ. MAI, 2001). This is explained by the insufficient moment of aerodynamic damping and the large dependence of the realized overload on the static stability of the UAV. As a result, the UAV reaction to the input control command under the action of external disturbances (random gusts of wind, a blast wave from undermining the warheads of other UAVs) is a long-lasting undamped oscillatory process with great overshoot, which leads to the oscillatory form of the UAV flight path, large misses (for example , for SAM), and in some cases - to the destruction of the UAV on the trajectory.

Thus, the control of high-precision UAVs is possible only with the help of on-board stabilization systems, including disturbance compensation units.

A known method of generating control signals for a symmetrical rocket in autopilot using two lateral control channels in the planes of the wings of the rocket, in each of which the current values of lateral accelerations and angular velocities of rotation of the rocket relative to the transverse axes are measured, and error signals proportional to the measured parameters are transmitted to the rudders and a control channel around the longitudinal axis, in which the roll angle of the rocket is measured and a signal proportional to it is transmitted to control the ailerons (Project Anti-aircraft guided missiles / Under the editorship of I.S. Golubev and V.G. Svetlov. - Second ed., revised and enlarged. - M .: Ed. MAI, 2001). This control method is implemented in the stabilization circuit of normal overload by generating a signal in the control generation unit proportional to the required value of normal overload, and calculating the stabilization error based on the signals received from the differentiating gyroscope and the linear acceleration sensor. Based on the magnitude of the stabilization error, a signal is calculated that is input to the steering gear.

Stabilization systems of this type provide a fairly good parry wind regular and random disturbances. But, nevertheless, further expansion of their capabilities is required to compensate for the action of any perturbations that satisfy the given constraints, and an increase in speed.

The closest analogue (prototype) of the claimed invention adopted a method of generating a control signal in autopilot (patent No. 2293686 "Autopilot for anti-aircraft guided missile stabilized roll", 2005). This method extends the functionality of the well-known autopilot and consists in the following: control is formed using two transverse control channels identical in structure in the planes of the wings of the rocket, in each of which the current values of lateral accelerations and angular velocities of rotation of the rocket relative to the transverse axes are measured and transferred to the rudders error signals proportional to the measured parameters, while the proportionality coefficients are adjusted inversely with the value korostnogo head, and a control channel about the longitudinal axis, wherein the measured roll angle of the missile and a proportional signal is transmitted to control the ailerons.

The main disadvantage of this method is that the control in the stabilization circuit is formed without taking into account the future UAV reaction to external disturbances, and is formed only on the basis of the measured current acceleration and angular velocity of rotation, which leads to an increase in stabilization error and, as a result (for SAM) to increase miss on hover.

The claimed invention has the task to develop a method for generating a control signal for the UAV stabilization system, taking into account the effects of external disturbances with unknown statistical properties.

The solution of this problem is achieved by introducing into the stabilization circuit a normal overload of the control signal generated in the stabilization signal generating unit based on calculations using information about the current state of the UAV phase vector, which provides a given value of normal overload and compensates for the operating disturbances.

The task of compensating for external limited disturbances in the proposed application is solved by predicting the guaranteed error of the stabilization system, which will not increase during the movement under the action of any disturbances that satisfy the given restrictions.

The inventive method of forming control of the normal overload of the UAV is illustrated by the drawings.

Figure 1 shows the structural diagram of the stabilization circuit of the normal overload of the UAV.

Figure 2 shows the relative position of the minimax reachable UAV region (region G).

Figure 3 shows graphs of transients in the stabilization system of normal overload.

Figures 4 and 5 show examples in the form of graphs of changes in the normal overload of missiles.

The problem is solved as follows.

An additional stabilization signal generating unit (BFSS) 2 is introduced into the well-known stabilization loop for UAV normal overload 1 (Fig. 1) with a differentiating gyroscope and a linear acceleration sensor, in which a control signal is generated that compensates for the effect of disturbances. The stabilization system under consideration (Fig. 1) includes a UAV 1 as an object of regulation, a steering gear (RP) 3 and two negative feedbacks: the angular velocity ω _z measured by a differentiating gyroscope (DG) 4 and the normal overload N _y measured by the sensor linear accelerations (DLU) 5. The input for the stabilization system is the control command u generated in the stabilization signal generating unit (BFSS) 2. The inputs for the BFSS are the required value of the normal overload N _y required generated in the control forming unit (BFU) 6, and the vector of the current state of the UAV x. A disturbance ξ acts on the UAV. In the scheme (1) denotes: k _y, k _DW, k _DLU - gains, δ _B - Bookmarks rudder angle, ε - the signal input of the steering actuator. Positions 7 and 8 indicate the subtraction blocks.

The following approach is used to evaluate the guaranteed stabilization error.

The calculation of the control signal u, which compensates for the action of disturbances, is based on the solution of the following antagonistic differential game, in which two players are involved, pursuing opposite interests. The first player chooses the control of the UAV, and the second player selects the external disturbance.

Let the first player acting in the interests of the UAV search for a minimum of a criterion defined as a functional:

$$J=\sqrt{{\left(N_y(T)-N_{y_{tReb}}\right)}^2+{\dot{N}}_y{(T)}^2}$$

the interests of the second player (disturbance) are opposite.

- скорость изменения нормальной перегрузки. Момент окончания переходного процесса Т не задан и определяется в процессе стабилизации.Here N _{y req} is the required value of the normal UAV overload, $${\dot{N}}_y$$

- rate of change of normal overload. The end of the transition process T is not set and is determined in the stabilization process.

It is required to find a UAV control that provides a minimum of functionality at time T under the assumption that the interests of the second player are opposite. In this differential game, positional strategies are the best ways to select player controls.

To solve this problem, we use an approach based on solving auxiliary problems of minimax program control.

Player controls will be selected at discrete points in time

t ₀ = 0, t ₀ + Δt, t ₀ + 2Δt, ...,

and so on, where Δt is the control selection step.

To select a control in the position {t _* , z (t _* )} (here z is the state vector of the system), an auxiliary minimax optimal control problem is compiled, which is posed as the initial conflict problem, but it is solved from the position {t _* , z (t _* )} and player controls are defined only as functions of time on the time interval [t _* , T).

и в качестве оптимального управления первого игрока в позиции {t_*,z(t_*)} принимаетсяAs a result of solving the auxiliary minimax problem, program control is determined $${\tilde{u}}_{PR}(t)$$

and as the optimal control of the first player in the position {t _* , z (t _* )} is taken

. $$\tilde{u}\left(t_{*},z\left(t_{*}\right)\right)={\tilde{u}}_{PR}\left(t\right){|}_{t=t_{*}}$$

.

With the selected control, a transient occurs over time Δt. At the same time, a certain perturbation ξ is realized, which is not measured during the operation of the stabilization system. In the position {t _* + Δt, z (t _* + Δt)} the auxiliary task is again drawn up, etc.

(Фиг.2). Особенность вспомогательной задачи состоит в том, что минимаксная (область G на фиг.2) строится с учетом противодействия второго игрока.The solution of the auxiliary problem is based on the calculation for a number of future hypothetical moments of the end of the transition process of the minimax reachability region (G) of the stabilization system in the plane of two coordinates: normal overload N _y and rate of change $${\dot{N}}_y$$

(Figure 2). A feature of the auxiliary problem is that the minimax (region G in FIG. 2) is constructed taking into account the opposition of the second player.

The algorithm for calculating the control u at the position {t _* , z (t _* )} using the developed method consists of the following operations.

1. The initial hypothetical moment of the end of the transient process T _* = t _* + ΔT, where ΔT is the change step T _{*, is set} .

строится минимаксная ОД игроков G для гипотетического момента окончания переходного процесса T_* (фиг.2).2. In the coordinate plane $$O{N}_y{\dot{N}}_y$$

the minimax OD of the players G is constructed for the hypothetical moment of the end of the transition process T _* (Fig. 2).

(точка А). Здесь заданная точка соответствует требуемому значению перегрузки, подаваемому на вход БФСС.3. The point of general OD closest to the given point is determined $$\left(N_{y_{tReb}}0\right)$$

(point A). Here, the set point corresponds to the required overload value supplied to the BFSS input.

4. A hypothetical miss ε _* (t _* , T _* ) is determined.

, обеспечивающая перемещение из позиции {t_*,z(t_*)} в точку А в момент T_* и управление и $$\tilde{u}\left(t_{*},T_{*}\right)=\tilde{u}(t_{*})$$

.5. The control program of the first player is determined $$\tilde{u}\left(t\right)$$

providing movement from the position {t _* , z (t _* )} to point A at time T _* and control and $$\tilde{u}\left(t_{*},T_{*}\right)=\tilde{u}(t_{*})$$

.

) итак далее.6. The hypothetical moment of the end of the transition process increases by ΔT and again determined by ε _* (t _* , T _* + ΔТ), $$\tilde{u}\left(t_{*},T_{*}+\Delta T\right)$$

) and so on.

в позиции {t_*,z(t_*)} находится из условия7. The optimal hypothetical moment of the end of the transition process $${\tilde{T}}_{*}$$

in the position {t _* , z (t _* )} is found from the condition

, $$\underset{t*\le T*\le \tau }{\min }{\epsilon }_{*}\left(t_{*},T_{*}\right)$$

,

where τ is the maximum transition time.

в позиции {t_*,z(t_*)} принимается $$\tilde{u}\left(t_{*},z\left(t_{*}\right)\right)=\tilde{u}\left(t_{*},{\tilde{T}}_{*}\right)$$

, то есть управление, вычисленное по взаимному расположению ОД игроков и заданной точки в оптимальный гипотетический момент окончания переходного процесса $${\tilde{T}}_{*}$$

.8. As an optimal control $$\tilde{u}\left(t_{*},z\left(t_{*}\right)\right)$$

at the position {t _* , z (t _* )} is accepted $$\tilde{u}\left(t_{*},z\left(t_{*}\right)\right)=\tilde{u}\left(t_{*},{\tilde{T}}_{*}\right)$$

, that is, the control calculated by the relative position of the OD of the players and the given point at the optimal hypothetical moment of the end of the transition process $${\tilde{T}}_{*}$$

.

The results of applying the proposed method of forming control in the stabilization loop of the UAV normal overload to compensate for the effects of external disturbances is illustrated by the following examples.

(кривая А). Требуемое значение нормальной перегрузки, подаваемое с БФУ, равнялось $$N_{y_{треб}}=5$$

.Figure 3 shows the graphs of transients in the stabilization system of normal overload under the action of a perturbation of the form ξ = ξ _max sin (ωt + φ), one of which is obtained by forming the control by the proposed method (curve B), the second is obtained by the action of a constant control signal $$u\left(t\right)=N_{y_{tReb}}=const$$

(curve A). The required value of the normal overload supplied from the BFU was $$N_{y_{tReb}}=5$$

.

и при действии одинаковых возмущений показали, что быстродействие системы стабилизации перегрузки БПЛА, управление в которой формируется предложенным способом выше, чем у известной системы стабилизации. Установившаяся ошибка управления в контуре стабилизации перегрузки БПЛА, управление которым формируется предложенным способом, всегда меньше, чем в прототипе при действии аналогичного возмущения.The results of computer simulation when applying a constant input control signal $$u\left(t\right)=N_{y_{tReb}}=const$$

and under the action of the same disturbances, they showed that the speed of the UAV overload stabilization system, in which the control is formed by the proposed method, is higher than that of the known stabilization system. The steady-state control error in the stabilization circuit of the UAV overload, the control of which is formed by the proposed method, is always less than in the prototype under the action of a similar disturbance.

Figure 4 and figure 5 shows examples in the form of graphs of changes in the normal overload of missiles in the process of guidance by the method of proportional navigation to a maneuvering target under the action of a disturbance using the proposed stabilization method.

These graphs show:

- the required value of the overload, calculated in the control formation unit (BFC) (curve A);

- a graph of the change in congestion in the process of pointing missiles to the target when forming control in the stabilization contour of the missiles in the proposed way (curve B).

Goal management was selected by law $$\alpha \left(t\right)=\{\begin{array}{cc}{\alpha }_{max},& 0\le t<\mathrm{one}c;\\ -{\alpha }_{max},& t\ge \mathrm{one}c.\end{array}$$

Here, the target control is the angle of attack α of the _target , where α _max is the maximum allowable angle of attack.

A SAM was affected by a perturbation of the form ξ = ξ _max sin (ωt + φ).

As can be seen from the graphs, the missile defense stabilization system, the control of which is formed by the proposed method, surpasses both the accuracy and speed of the prototype stabilization system, as a result, the accuracy of the missile guidance is increased. So, for example, the miss obtained in the formation of control of the missile stabilization system by the proposed method is 4.25 m, and when using the prototype stabilization system - 9.41 m, i.e. pointing accuracy increased by more than 2 times.

Thus, the invention made it possible to obtain a technical result, namely, to increase the speed and accuracy of control of the stabilization system of the UAV normal overload, which led to an increase in control accuracy (for missiles - to increase the accuracy of guidance). In the present invention, to calculate the control signal that compensates for the action of external disturbances, information on the nature of the disturbance is not required, because control is calculated on the basis of the analysis of the relative position of the minimax reachability range of the UAV, which is built taking into account all kinds of disturbances that satisfy the given restrictions and the given value of the normal overload.

Claims (1)

The method of generating a control signal for the stabilization system of the normal overload of an unmanned aerial vehicle, including measuring the motion parameters of an unmanned aerial vehicle, calculating a control signal and inputting it into the stabilization loop of a normal overload, characterized in that for calculating a control signal based on the calculation of reachability areas in the coordinate plane "normal overload - the rate of change of the normal overload ”for a number of future discrete transition end times process of choosing optimal hypothetical time transient closure assayed position in the coordinate region of reachability "normal overload - rate of change of normal overload" plane at this time, and based on this analysis, calculating a control signal.

Images