Patent Application
20190241076
The invention relates to a multicopter, using which wind measurements can be carried out, and a method, system, and storage medium for wind measurement by means of a multicopter. In particular, the invention relates to a multicopter that is a free-flying drone or a free-flying robot driven by multiple propellers.
Multicopters which are configured for wind measurement are known in the prior art. Thus, a multicopter is disclosed in the publication: “Simultaneous Estimation of Aerodynamic and Contact Forces in Flying Robots: Application to Metric Wind Estimations and Collision Detection”, ICRA 2015, Seattle, Wash. USA May 2015, pages: 5290-5296 of the present inventors: Tomic T. and Haddadin S., which, on the basis of an estimation of a force screw τe acting externally on the multicopter, enables the estimation of a wind speed. The disclosure of this article, in particular the chapters: “III Modeling” and “IV Incorporating Aerodynamics into external Wrench Estimation” is hereby explicitly incorporated into the content of the disclosure of this description.
A controller of a multicopter is disclosed in the article “Aerodynamic Power Control for Multirotor Aerial Vehicles” by Bangura M. et al., in: IEEE International Conference on Robotics and Automation (ICRA), 31.05.-07.06.2014 Hong Kong, China, which has robust behavior in relation to wind disturbance and ground effects.
A method is known from U.S. Pat. No. 8,219,267 B2, using which the wind speed can be estimated during the operation of a UAV.
The object of the present invention is to specify a multicopter and a method for operating a multicopter, using which an improved wind measurement is possible.
The invention results from the features of the independent claims. Advantageous refinements and embodiments are the subject matter of the dependent claims. Further features, possible applications, and advantages of the invention result from the following description, and also the explanation of example embodiments of the invention which are illustrated in the figures.
One aspect of the invention relates to a multicopter having: a number N of electric motors MOTn for driving N propellers PROPn, where n=1, 2, . . . N and N≥2, a first interface for providing first parameters P1 including: a 3D position rB=(xb, yb, zb) of the center of gravity B of the multicopter, the time derivatives: {dot over (r)}B and {umlaut over (r)}B, a 3D orientation oM=(αM, βM, γM) of the multicopter and its time derivative {dot over (o)}M, a second interface for providing an aerodynamic power Pa,n currently generated by the respective propellers PROPn, a unit 803, which, on the basis of the first parameters P1 and/or the aerodynamic power Pa,n and/or an estimation of one or more force screws τe externally acting on the multicopter, determines a relative speed vr:=(vr,x, vr,y, vr,z)T of the multicopter in relation to the air, a unit 805, which, on the basis of the determined relative speed vr:=(vr,x, vr,y, vr,z)T and on the basis of the parameter P1, determines the wind speed vw in an inertial system, and a storage unit for storing the wind speeds vw(rB)=(vw,x(rB), vw,y(rB), vw,z(rB)) determined for the locations rB, and/or a transmission unit for the wireless transmission of the wind speed vw(rB) to a receiver.
The determination of the relative speed vr:=(vr,x, vr,y, vr,z)T of the multicopter in relation to the air is advantageously performed by means of methods of neuronal learning and/or autonomous robot learning.
The arrangements of the propeller planes on the multicopter are fundamentally arbitrary in this case. The determination of the relative speed vr:=(vr,x, vr,y, vr,z)T therefore includes combinations of the estimated force screw and the estimated power.
If the propeller planes are arranged essentially parallel and in particular in one plane, the following special case thus results.
A further aspect of the invention relates to a multicopter having: a number N of electric motors MOTn for driving N propellers PROPn, where n=1, 2, . . . , N and N≥2, a first interface for providing first parameters P1 including: a 3D position rB=(xb, yb, zb) of the center of gravity B of the multicopter (or providing items of information from which the position rB=(xb, yb, zb) of the center of gravity B of the multicopter can be determined/estimated), the time derivatives: {dot over (r)}B and {umlaut over (r)}B, a 3D orientation oM=(αM, βM, γM) of the multicopter and its time derivative {dot over (o)}M (the variables {dot over (r)}B, {umlaut over (r)}B, oM=(αM, βM, γM), {dot over (o)}M are advantageously measured (for example, by means of GPS, LIDAR, cameras, etc., one or more inertial measuring units (IMU)) etc., observed, or derived (for example, by visual odometry, sensor data fusion such as Kalman filters etc.), a second interface for providing an aerodynamic power Pa,n currently generated by the respective propellers PROPn (which is advantageously measured or observed), a first unit, which, on the basis of the first parameters P1, on the basis of a provided model M1 for describing dynamics of the multicopter, and an estimation determined on the basis of the model M1 of a force screw τe acting externally on the multicopter, determines horizontal components (vr,x, vr,y) of a relative speed vr:=(vr,x, vr,y, vr,z)T of the multicopter in relation to the air (for this purpose function approximators are advantageously used, which are known from the field of “machine learning”, for example, linear regression, neuronal networks, multilayer perceptrons, support vector machines (SVM) etc.), a second unit which, on the basis of the determined horizontal components (vr,x, vr,y), and on the basis of the aerodynamic power Pa,n, determines the vertical component (vr,z) of the relative speed vr, a unit, which, on the basis of the determined horizontal components (vr,x, vr,y), the vertical component (vr,z), and also on the basis of the parameters P1, determines the wind speed vw, in an inertial system, a storage unit for storing the wind speeds vw(rB)=(vw,x(rB), vw,y(rB), vw,z(rB)) determined for the locations rB and/or a transmission unit for wirelessly transmitting the wind speed vw(rB) to a receiver.
In the present case, the term “multicopter” describes an aircraft which has two or more drive units, in particular motor-driven propellers. The multicopter can also have lift and/or control surfaces. The arrangement of the N propellers PROPn on the multicopter is arbitrary in principle. Each propeller is advantageously described in its own coordinate system. The N propellers PROPn can advantageously be vectored, i.e., each propeller can be described in its own time-variable coordinate system. The method can accordingly be generalized in principle to other arrangements, in which the arrangement of the propellers is not necessarily in one plane. The corresponding required combinations of the estimators is derivable in an obvious manner from the described common case.
In principle, the proposed method can in other words execute a regression from the external force screw τe to the relative speed vr, a regression from the aerodynamic powers Pa,n to the relative speed vr, and/or a regression from the external force screw τe and the aerodynamic powers Pa,n to the relative speed vr.
The relationship: vr=rB−vw advantageously applies in this case, where: vr:=(vr,x, vr,y, vr,z)T:=relative speed of the multicopter in relation to the air, and vw:=(vw,x, vw,y, vw,z)T is the wind speed.
The proposed multicopter thus solely uses the approach disclosed in the article indicated at the outset to determine the horizontal components (vr,x, vr,y) of the relative speed vr, in which, based on a model of the dynamics of the multicopter and an estimation generated thereby of a force screw τe acting externally on the multicopter, an estimation of the relative speed vr is performed.
The determination of the vertical component (vr,z) of the relative speed vr is performed in the present case on the basis of the determination of the aerodynamic power of the respective motor-propeller combinations and also on the basis of the previously determined horizontal components (vr,x, vr,y) of the relative speed vr. The proposed multicopter thus enables a more accurate and robust determination of wind speeds during flight. It can therefore be used, in particular, as a flying wind sensor, which stores the determined wind data on board and/or transmits them to a ground station.
Several relationships and mathematical principles are described hereafter, which are used for the explanation and the implementation of the invention.
The movement equations applicable to the present multicopter can fundamentally be formulated as follows:
m{umlaut over (r)}=mge
3
+Rf+Rf
e, (Eq. 1)
I{dot over (ω)}=(Iω)×ω−mg(rg)×RT e3+
{dot over (R)}=R(ω)× (Eq. 3)
where:
In this case, there are the target force screw τ=[fT,
The estimation of the force screw τe=[feT,
where:
More detailed statements in this regard result, for example, from the article by Tomic T., “Evaluation of acceleration-based disturbance observation for multicopter control” in European Control Conference (ECC), 2014, pages 2937-2944.
The forces acting on a propeller of the multicopter are dependent on the free flow speed v∞ (relative wind speed). The free flow speed v∞ of the n-th propeller in a reference system of the propeller can be expressed as follows:
v
∞
(n)
=R
pb
(n)(Rbi(n)vr+ω×rn) (Eq. 5)
wherein vr={dot over (r)}B−vw is the true airspeed of the multicopter, vw is the wind speed, Rbi(n) is the rotation matrix from the inertial system to the reference system of the multicopter, Rpb(n) is the rotation matrix from the reference system of the multicopter to the reference system of the propeller, ω is the angular speed of the multicopter, and rn is the position of the respective propeller in relation to the reference system of the multicopter. On the basis of preservation of momentum, the following results for the thrust T generated by the n-th propeller
T=2ρAviU (Eq. 6)
where: ρ:=air density, A:=area swept by the propeller, and U=∥vie3+v∞∥ is the total wake flow generated by the propeller. The flow speed vi induced by the propeller can advantageously be determined as follows:
v
i
=v
h
2/√{square root over (vxy2+(vi−vz)2)}. (Eq. 7)
The solution is advantageously performed by means of a Newton-Raphson method in a few steps with known vh and v∞. For the horizontal and the vertical component of the free flow speed, the following applies: vxy=v∞−vz and vz=e3Tv∞. In hovering flight, the following applies for the induced speed: vi=vh=√{square root over (Th/2ρA)}, wherein the hovering flight thrust Th results from Th=ρD4CT
P
a=2ρAviU(vi−vz). (Eq. 8)
The aerodynamic power in forward flight in relation to the hovering flight power is:
P
a
/P
h=(vi−vz)/vh (Eq. 9)
given by Ph=2ρAvh3. Deviations from the ideal behavior can be taken into consideration by introducing a factor FM between 0 and 1. The relationship between aerodynamic power Pa of a drive unit (electric motor and propeller) and the motor power PM generated by an electric motor results in this case as: Pa=PM·FM.
The following dynamic motor model is advantageously used to estimate the aerodynamic power of a drive unit.
τm=(Kq0−Kq1ia)ia (Eq. 10)
I
r
E=τm−Dr (Eq. 11)
where: ia:=motor current,
P
m=τm
Pr=Ir
P
a
=FM((Kq0−Kq1ia)ia−Ir
In summary, this means that the motor current ia is measured or estimated. The variables
V. Estimation of the Wind Speed vw Based on the Determination of an External Force Screw τe
As stated in the article mentioned at the outset, the wind speed can be determined based on the force screw τe acting externally on the multicopter. It is assumed in this case that the force screw τe may be attributed exclusively to aerodynamic τd friction forces: τe=τd, so that the fundamental aerodynamic model M1 merely has to be inverted: τd=d(vr). For simple aerodynamic models, this can be performed via a simple relation or iteration. If one uses, for example, a linear model M1 as a basis, the following thus applies, for example:
d(vr)=Dlvr Σ
where Dl:=linear coefficient matrix. If {circumflex over (f)}e=d(vr) is used, the following results:
where: D:=a coefficient matrix. This simple model furthermore implies that the multicopter has a symmetrical shape.
Alternatively, a learning-based approach can be followed. The above relation can also be modeled by means of a radial basis function (RBF) neuronal network. This has the advantage that the inverse relation is coded directly in the radial basis function. The relation vr=d−1(τe) is advantageously modeled as a normalized RBF network having K base functions:
where: W ∈ 3×K:=matrix having weighting constants of the RBFs for each speed component:
and where: φ=[φ(r1), . . . , φ(rK)]T of the evaluated base functions. The network is advantageously normalized by the factor: φΣ=Σi=1K φ(ri). Gaussian base functions are advantageously used:
where σ:=form parameter, x:=the determined vector, and ci:=the center of the i-th base function.
Proceeding from the equations (Eq. 7), (Eq. 8), and (Eq. 9), the aerodynamics for a propeller can be formulated as a system of nonlinear equations F(vr,z, vr,xy, vi, vh, Pa)=0 and F=[F1, F2, F3]T, where:
F
1
=v
i
4−2vi3vr,z+vi2(vr,xy2+vr,z2)−vh4=0,
F
2
=v
i
U(vi−vr,z)−Pa/(2ρAn)=0, and
F
3
=v
h
2(vi−vr,z)−Pa/(2ρAn)=0. (Eq. 21)
It is presumed that Pa/(2ρA) and vh are known and the vector x=[vr,x, vr,y, vr,z, vi]T is to be determined. The above-mentioned system of nonlinear equations (Eq. 21) is under-determined, since three unknown and only two known variables are present (in this case, the horizontal components vr,x and vr,y are coupled in vr,xy. A plurality of solutions of the equation system (Eq. 21) thus results. To solve this minimization problem, it is proposed that the system of the equations (Eq. 21) be expanded in such a way that a plurality (number K) of “measurements” of Pa is integrated into the equation system.
Overall, K measurements are thus carried out to determine the aerodynamic power Pa,n. The aerodynamic power Pa,n can be produced in this case, for example, K times for a single electric motor and its propeller. Advantageously, the K “measurements” of the aerodynamic power Pa,n are performed for two or more different electric motors and the propellers thereof. Finally, for the K different measurements, a transformation of the equation system (Eq. 21) into a common reference system is performed. This fundamentally enables the estimation of all three components of the relative wind direction and the wind speed vi induced by the propeller by solving the nonlinear quadratic minimization problem (Eq. 21).
It is presumed in this case that during the K “measurements” of the aerodynamic power Pa, the wind speed vw=[vw,x, vw,y, vw,z] remains equal. If the K measurements are carried out in a sufficiently short time, this assumption is thus adequately justified. Sufficiently accurate wind measurements require K determinations of the aerodynamic power, wherein it is sufficient to select K advantageously ≤10, or K advantageously in the range of 3 to 8. The higher K is selected, the greater is the computing effort and accompanying this also the period of time in which measurements are carried out, so that as a period of time becomes longer, the probability of a variation of the wind in the period of time rises.
Advantageously, N measurements or determinations of the aerodynamic power Pa are carried out simultaneously on the K electric motor propeller units. If, for example, K=8 is selected and the multicopter has four drive units, only two successive “measurements” of the aerodynamic power Pa are thus required per drive unit. “Measurements” of the aerodynamic power Pa from a variety of poses of the multicopter at different points in time in a small time window can advantageously be combined. If the flight of the multicopter is not aggressive, i.e., the orientation of the multicopter does not change significantly in the measuring period of time, the free flow speed v∞ in the reference system of the multicopter can thus be estimated sufficiently accurately. Overall, K “measurements” of the aerodynamic power Pa are thus carried out to determine the vertical component [vw,z] of the wind direction vw.
The state vector x to be determined for K measurements is:
x=[vr,x, vr,y, vr,z, vi,1, vi,2, . . . , vi,K]T (Eq. 22)
wherein the expanded equation system has to be solved:
F(vr,x, vr,y, vr,z, vi,1, vh,1, Pa,1, . . . , vi,K, vh,K, Pa,K)=0
F=[F1,1, F2,1, F3,1, . . . , F1,K, F2,K, F3,K]T (Eq. 23)
wherein F1,k, F2,k, F3,k are evaluations of the equation system (Eq. 21) for the k-th “measurement” of the aerodynamic powers. To solve the equation (Eq. 23), a Jacobi matrix is required. The Jacobi matrix for the k-th measurement results as:
where: Jij(k)=∂ Fi,k/∂ xj,k. The expanded Jacobi matrix ∈
3K×K+3 can thus now be constructed. In the example case of three measurements (K=3), the following results: x|N=3=[vr,x, vr,y, vr,z, vi,1, vi,2, vi,3]T and
An expansion to K measurements is easily possible for a person skilled in the art. If measurements from various poses of the multicopter are combined, the wind speeds resulting in each case have to be transformed into a common reference system.
The free flow speed for the n-th propeller is advantageously defined as follows:
wherein the transformed speeds are used to compute equations (Eq. 21) and (Eq. 24). The offset speed v0(n) can be obtained from a pose estimation system as the difference speed between two measurements. The offset speed of the propeller on the basis of an angular speed of the multicopter can advantageously also be used, for example, v0(n)=Rpb(n) ω×rn. The rotation matrix R(n) transforms the relative speed from the common coordinate system into the propeller coordinate system. This formulation advantageously enables the independent determination of all three components of the free flow speed. Furthermore, it also permits the determination of the wind components vw for the case in which the propellers are not arranged in a coplanar configuration on the multicopter.
If the equations “match”, a multidimensional optimization problem is to be solved. The solution is then in the intersection of all nonlinear functions for which the following applies: F=0. In the general case, however, an intersection does not result for all nonlinear functions. In this case, a nonlinear quadratic minimization problem has to be solved using the following objective function:
This is advantageously performed using a Levenberg-Marquard method. If an exact solution exists,
The solution space of the equation (Eq. 26) contains local optima. Based on the fundamental physics, the same measured aerodynamic powers Pa can occur at different wind speeds vw and induced speeds vi. The optimized variables are speeds. It is therefore advisable to use physical considerations as the foundation, to differentiate reasonable solutions from nonsensical solutions. A multicopter has to use power to generate thrust, which means T>0 and Pa>0. The induced speed vi is less than the hovering flight speed vh. The induced speed is advantageously restricted to a range 0<vi<vh.
One advantageous refinement of the multicopter is distinguished in that the second unit is embodied and configured for the purpose of solving the following quadratic minimization problem or one which is transferable thereto or is equivalent in its basic concepts for a number of a total of K measurements of the aerodynamic power Pa,n for f, where k=1, . . . , K and K≥1:
wherein the following applies:
F(vr,x, vr,y, vr,z, vi,1, vh,1, Pa,1, . . . , vi,K, vh,K, Pa,K)=0, (2)
F=[F1,1, F2,1, F3,1, . . . , F1,K, F2,K, F3,K], (3)
F
1,K
=v
i,k
4−2vi,k3 vr,z+vi,k2(vr,x2+vr,y2+vr,z2)−vh,k4=0, (4)
F
2,k
=v
i,k
U
k(vi,k−vr,z)−Pa,k/(2ρAn)=0, (5)
F
3,k
=v
h,k
2(vi,k−vr,z)−Pa,k/(2ρAn)=0, and (6)
U
k=√{square root over (vr,x2+vr,y2+(vi,k−vr,z)2)} (7)
where
Above equations (1) to (7) are described in the respective propeller coordinate system.
One advantageous refinement of the multicopter is distinguished in that the second unit is embodied and configured in such a way that the horizontal components (vr,x, vr,y) of the relative speed vr determined by the first unit are taken into consideration in the nonlinear quadratic minimization problem by the following relationships or relationships transferable thereto:
F
4,k
=v
r,x,k
−v
r,x=0 and (9)
F
5,k
=v
r,y,k
−v
r,y=0. (10)
The equations (9) and (10) are described in the respective propeller coordinate system. Therefore, the speed estimated according to present claim 1 can be projected in the propeller coordinate system, so that these equations apply.
An advantageous refinement of the multicopter is distinguished in that the model M1 is based on the following movement equation or the model M1 can be traced back to the following movement equation:
M{dot over (v)}+C(v)v+gM=JTτ+τe
wherein the following applies: τe=τd(vr) and vr={dot over (r)}B−vw, where
One advantageous refinement of the multicopter is distinguished in that a control system is provided for controlling and regulating the electric motors MOTn, wherein the control system is embodied and configured in such a way that the electric motors MOTn are regulated in such a way that the projection of the determined wind speed vw on the direction of axis of rotation of the propellers PROPn is maximized.
The fundamental optimization problem can also be mathematically formulated. For this purpose, an orientation o* has to be determined which minimizes the condition number of the Jakobi matrix, by o*=argmin K(J), wherein K is the condition number.
Trajectory planning can advantageously be used for this purpose in particular and/or the control of the multicopter can be adapted to the optimal wind estimation. The solution can also be generated by gradient methods, potential-based methods, or other equivalent mathematical solutions.
The accuracy of the wind measurement is enhanced and the robustness of the measurement is improved by this measure.
A 3D acceleration sensor and a gyroscope and also a position determination system are advantageously provided in the multicopter and are connected to the first interface for determining the first parameters P1. The position determination system is advantageously a satellite navigation system and/or an optical navigation system.
One advantageous refinement of the multicopter is distinguished in that for the determination of the aerodynamic power Pa,n per electric motor MOTn, a motor current sensor, a motor rotational speed sensor, or a motor rotational speed estimator are provided, and also a unit which determines the aerodynamic power Pa,n on the basis of a predetermined model M2 for describing the propeller dynamic range, the detected motor currents, and motor rotational speeds. At least one corresponding model M2 was already explained above.
A further aspect of the invention relates to methods for measuring a wind speed vw using a multicopter, wherein the multicopter has a number N of electric motors MOTn for driving N propellers PROPn, where n=1, 2, . . . , N and N≥2, including the following steps: providing first parameters P1 including: a 3D position rB=(xb, yb, zb) of the center of gravity B of the multicopter, the time derivatives: {dot over (r)}B and {umlaut over (r)}B, a 3D orientation oM=(αM, βM, γM) of the multicopter and its time derivative {dot over (o)}M, providing an aerodynamic power Pa,n currently generated by the respective propellers PROPn on the basis of the first parameters P1 and/or the aerodynamic power Pa,n and/or an estimation of one or more force screws τe acting externally on the multicopter, determining a relative speed vr:=(vr,x, vr,y, vr,z)T of the multicopter in relation to the air on the basis of the determined relative speed vr:=(vr,x, vr,y, vr,z)T and on the basis of the parameters P1, determining the wind speed vw in an inertial system, and storing the wind speeds vw(rB)=(vw,x(rB), vw,y(rB), vw,z(rB)) determined for the locations rB, and/or transmitting the wind speed vw(rB) to a receiver.
A further aspect of the present invention relates to a method for measuring a wind speed vw using a multicopter, wherein the multicopter has a number N of electric motors MOTn for driving N propellers PROPn, where n=1, 2, . . . , N and N≥2 including the following steps.
In one step, a provision of first time-dependent parameters P1 is performed, including a 3D position rB=(xb, yb, zb) of the center of gravity B of the multicopter (or providing pieces of information from which the position rB=(xb, yb, zb) of the center of gravity B of the multicopter can be determined/estimated), the time derivatives: {dot over (r)}B and {umlaut over (r)}B, a 3D orientation oM=(αM, βM, γM) of the multicopter and the time derivative {dot over (o)}M (the variables {dot over (r)}B, {umlaut over (r)}B, oM=(αM, βM, γM), {dot over (o)}M are advantageously measured (for example, by means of GPS, LIDAR, cameras, etc., one or more inertial measuring units (IMU)) etc., observed or derived (for example, by visual odometry, sensor data fusion such as Kalman filters etc.). In a further step, a provision of an aerodynamic power Pa,n generated by the respective propellers PROPn is performed (which is advantageously measured or observed). In a further step, on the basis of the first parameters P1 and/or a provided model M1 for describing dynamics of the multicopter and/or an estimation determined on the basis of the model M1 of one or more force screws τe acting externally on the multicopter, a determination is performed of horizontal components (vr,x, vr,y) of a relative speed vr of the multicopter in relation to the air (for this purpose functional approximators are advantageously used, which are known from the field of “machine learning”, for example, linear regression, neuronal networks, multilayer perceptrons, deep neuronal networks, convolution networks, recurrent networks, for example, LSTM networks, support vector machines (SVM) etc.). In a further step, on the basis of the determined horizontal components (vr,x, vr,y) and the aerodynamic power Pa,n, a determination is performed of the vertical component (vr,z) of the relative speed vr. In a further step, on the basis of the determined relative speed vr=(vr,x, vr,y, vr,z) and on the basis of the parameters P1, a determination is performed of the wind speed vw in an inertial system. Finally, storage is performed of the wind speed determined for rB: vw(rB)=(vw,x(rB), vw,y(rB), vw,z(rB)) and/or transmission of the wind speed: vw(rB)=(vw,x(rB), vw,y(rB), vw,z(rB)) to a receiver.
The horizontal components (vr,x vr,y) are thus determined by measuring the force screw acting externally on the multicopter using the equations (Eq. 16) and (Eq. 17). Based on the determined horizontal components (vr,x vr,y) and a determined aerodynamic power Pa, the vertical component (vr,z) is determined.
Finally, the three-dimensional wind speed can be determined, stored, transmitted, and possibly subsequently output by way of the relationship vr={dot over (r)}−vw.
One advantageous refinement of the proposed method is distinguished in that the following nonlinear quadratic minimization problem or one transferable thereto is solved for a number of a total of K measurements of the aerodynamic power Pa,n for f, where k=1, 2, . . . , K and K≥1:
wherein the following applies:
F(vr,x, vr,y, vr,z, vi,1, vh,1, Pa,1, . . . , vi,K, vh,K, Pa,K)=0, (2)
F=[F1,1, F2,1, F3,1, . . . , F1,K, F2,K, F3,K], (3)
F
1,K
=v
i,k
4−2vi,k3 vr,z+vi,k2(vr,x2+vr,y2+vr,z2)−vh,k4=0, (4)
F
2,k
=v
i,k
U
k(vi,k−vr,z)−Pa,k/(2ρAn)=0, (5)
F
3,k
=v
h,k
2(vi,k−vr,z)−Pa,k/(2ρAn)=0, and (6)
U
k=√{square root over (vr,x2+vr,y2+(vi,k−vr,z)2)} (7)
where
Above equations (1) to (7) are described in the respective propeller coordinate system.
One advantageous refinement of the proposed method is distinguished in that the determined horizontal components (vr,x, vr,y) are taken into consideration in the nonlinear quadratic minimization problem by the following relationships or relationships transferable thereto:
F
4,k
=v
r,x,k
−v
r,x=0 and (9)
F
5,k
=v
r,y,k
−v
r,y=0. (10)
Equations (9) and (10) are described in the respective propeller coordinate system. Therefore, the speed estimated according to present claim 1 can be projected into the propeller coordinate system so that these equations apply.
One advantageous refinement of the proposed method is distinguished in that the model M1 is based on the following movement equation or the model M1 can be traced back to the following movement equation:
M{dot over (v)}+C(v)v+gM=JTτ+τe
wherein the following applies: τe=τd(vr) and vr={dot over (r)}B−vw, where
One advantageous refinement of the proposed method is distinguished in that a control system is provided for controlling and regulating the electric motors MOTn, wherein the electric motors MOTn are regulated in such a way that the projection of the wind speed vw on the direction of the axis of rotation of the propellers is maximized.
One advantageous refinement of the proposed method is distinguished in that for the determination of the aerodynamic power Pa,n per electric motor MOTn, a motor current and a motor rotational speed are determined, and, on the basis of a predetermined model M2 for describing the propeller dynamic range, the detected motor currents, and motor rotational speeds, the aerodynamic power Pa,n is determined.
One advantageous refinement of the proposed method is distinguished in that the determination of the horizontal components (vr,x, vr,y) is performed by means of a model of the aerodynamic force screw, which describes the dependence vr(τe).
One advantageous refinement of the proposed method is distinguished in that the nonlinear quadratic minimization problem is solved by means of a nonlinear optimization method (for example, Levenberg-Marquardt).
A further aspect of the invention relates to a computer system having a data processing device, wherein the data processing device is designed in such a way that a method as described above is executed on the data processing device.
A further aspect of the invention relates to a digital storage medium having electronically readable control signals, wherein the control signals can interact with a programmable computer system so that a method as described above is executed.
A further aspect of the invention relates to a computer program product having program code stored on a machine-readable carrier for carrying out the method as described above with the program code executed on a data processing device.
A further aspect of the invention relates to a computer program having program code for carrying out the method as described above when the program runs on a data processing device. For this purpose, the data processing device can be designed as an arbitrary computer system known from the prior art.
A further aspect of the invention relates to a system including two or more multicopters as described above, wherein the multicopters are each also embodied and configured for bilateral data exchange with one another in a data network and the multicopters transmit the respective determined wind speed: vw(rB)=(vw,x(rB), vw,y(rB), vw,z(rB)) to the respective other multicopter. The data exchange preferably takes place via radio and/or optical data transmission.
The multicopters are advantageously embodied as agents or softbots for wireless communication in the data network. The term “agent” is used in the present case in the following meaning: “computer system which is located in a specific environment and is capable of carrying out independent actions in this environment to achieve its (predetermined) goals”.
One advantageous refinement of the proposed method is distinguished in that the control system for controlling and regulating the electric motors MOTn of the respective multicopter is embodied and configured in such a way that determined wind speeds transmitted from the other multicopters: vw(rB)=(vw,x(rB), vw,y(rB), vw,z(rB)) are taken into consideration in the regulation and movement planning.
One advantageous refinement of the proposed method is distinguished in that the multicopters transmit the positions rB of the center of gravity B and/or the time derivatives: {dot over (r)}B and/or {umlaut over (r)}B and/or the 3D orientation oM and/or its time derivative {dot over (o)}M of the respective multicopter and/or the force screws τe acting externally on the respective multicopter and/or the aerodynamic power Pa,n to the respective other multicopters.
One advantageous refinement of the proposed system is distinguished in that one or more multicopters and/or a control center, which is configured for the data exchange with the multicopters in the data network, is/are configured and embodied to solve an optimization problem according to claim 2. This enables in particular the proposed system to be used as a distributed wind estimator, in which a corresponding equation system is prepared, which takes into consideration all or some of the multicopters (agents), and an optimization problem is solved, for example, as in present claim 2. The learning of the regression via the example above-mentioned methods can also be applied to the multicopter scenario.
Further advantages, features, and details result from the following description in which—possibly with reference to the drawings—at least one example embodiment is described in detail. Identical, similar, and/or functionally-identical parts are provided with identical reference signs.
In the figures:
The multicopter furthermore includes a second unit 104, which, on the basis of the determined horizontal components (vr,x, vr,y) and on the basis of the aerodynamic power Pa,n, determines the vertical component (vr,z) of the relative speed vr, a third unit 105, which, on the basis of the determined horizontal components (vr,x, vr,y), the vertical component (vr,z), and on the basis of the parameters P1, determines the wind speed vw in an inertial system, a storage unit 106 for storing the wind speeds vw(rB)=(vw,x(rB), vw,y(rB), vw,z(rB)) determined for the locations rB and a transmission unit 107 for the wireless transmission of the wind speed vw(rB) to a receiver (not shown).
Although the invention was illustrated and explained in greater detail by the preferred example embodiments, the invention is not thus restricted by the disclosed examples and other variations can be derived therefrom by a person skilled in the art without leaving the scope of protection of the invention. It is therefore clear that a variety of possible variations exists. It is also clear that embodiments mentioned by way of example actually only represent examples which are not to be interpreted in any way as a limitation of, for example, the scope of protection, the possible applications, or the configurations of the invention. Rather, the above description and the description of the figures make a person skilled in the art capable of implementing the example embodiments in concrete form, wherein a person skilled in the art, aware of the disclosed concept of the invention, can perform manifold modifications, for example, with respect to the function or the arrangement of individual elements mentioned in an example embodiment, without leaving the scope of protection, which is defined by the claims and the legal equivalents thereof, for example, more extensive explanations in the description.
101 first interface
102 second interface
103 first unit
104 second unit
105 third unit
106 storage unit
107 transmission unit
201-206 method steps
| Number | Date | Country | Kind |
|---|---|---|---|
| 10 2016 119 152.3 | Oct 2016 | DE | national |
The present application is the U.S. National Phase of PCT/EP2017/001190, filed on Oct. 9, 2017, which claims priority to German Patent Application No. 10 2016 119 152.3, filed on Oct. 7, 2016, the entire contents of which are incorporated herein by reference.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2017/001190 | 10/9/2017 | WO | 00 |
