Geometric tracking control of a quadrotor UAV on SE(3)
Summary (2 min read)
Introduction
- A quadrotor unmanned aerial vehicle (UAV) consists of two pairs of counter-rotating rotors and propellers, located at the vertices of a square frame.
- Due to its simple mechanical structure, it has been envisaged for various applications such as surveillance or mobile sensor networks as well as for educational purposes.
- Little attention has been paid to constructing nonlinear control systems for them, particularly to designing nonlinear tracking controllers.
- By characterizing geometric properties of nonlinear manifolds intrinsically, geometric control techniques provide unique insights to control theory that cannot be obtained from dynamic models represented using local coordinates [15].
- The authors construct a tracking controller to follow prescribed trajectories for the center of mass and heading direction.
II. QUADROTOR DYNAMICS MODEL
- Consider a quadrotor vehicle model illustrated in Fig. 1.
- The first and the second axes of the body fixed frame, ~b1,~b2, lie in the plane defined by the centers of the four rotors, as illustrated in Fig.
- The authors develop a controller to follow a prescribed trajectory of the location of the center of mass, xd(t), and the desired direction of the first body-fixed axis, ~b1d(t).
- As a result, the authors are able to achieve almost global exponential attractiveness to the zero equilibrium of tracking errors.
A. Tracking Errors
- The authors define the tracking errors for x, v,R,Ω as follows.
- From this, the attitude tracking error eR is chosen to be eR = 1 2 (RTdR−RTRd)∨. (10) The tangent vectors Ṙ ∈ TRSO(3) and Ṙd ∈ TRdSO(3) cannot be directly compared since they lie in different tangent spaces.
- The authors choose the tracking error for the angular velocity as follows: eΩ = Ω−RTRdΩd. (11).
C. Exponential Asymptotic Stability
- The authors first show exponential stability of the attitude dynamics in the sublevel set L2 = {Rd, R ∈ SO(3) |Ψ(R,Rd) < 2}, and based on this result, they show exponential stability of the complete dynamics in the smaller sublevel set L1 = {Rd, R ∈ SO(3) |Ψ(R,Rd) < 1} Proposition 1: (Exponential Stability of Attitude Dynamics).
- (17), (18) represent a region of attraction for the attitude dynamics.
- More precisely, the position tracking performance is affected by the difference between ~b3 = Re3 and ~b3d = Rde3.
- In the proceeding stability analysis, it turns out that for the stability of the complete translational and rotational dynamics, the attitude error function Ψ should be less than 1, which states that the initial attitude error should be less than 90◦.
- Proposition 2: (Exponential Stability of the Complete Dynamics).
D. Almost Global Exponential Attractiveness
- Proposition 2 requires that the initial attitude error is less than 90◦ to achieve exponential stability of the complete dynamics.
- Suppose that this is not satisfied, i.e. 1 ≤ Ψ(R(0), Rd(0)) < 2. From Proposition 1, the authors are guaranteed that the attitude error function Ψ exponentially decreases, and therefore, it enters the region of attraction of Proposition 2 in a finite time.
- Therefore, by combining the results of Proposition 1 and 2, the authors can show almost global exponential attractiveness.
- This should be distinguished from the stronger notion of exponential stability, in which the constant α(δ) in the above bound is replaced by α(δ) ‖z(0)‖.
- (Almost Global Exponential Attractiveness of the Complete Dynamics) Consider a control system designed according to Proposition 2, also known as Proposition 3.
E. Properties and Extensions
- One of the unique properties of the presented controller is that it is directly developed on SE(3) using rotation matrices.
- It also avoids the ambiguities that arise when using quaternions to represent the attitude dynamics.
- Without these considerations, a quaternion-based controller can exhibit an unwinding phenomenon, where the controller unnecessarily rotates the attitude through large angles [15].
- The attitude error function defined in (8) has the following critical points: the identity matrix, and rotation matrices that can be written as exp(πv̂) for any v ∈ S2.
- The presented controller can be modified accordingly.
IV. NUMERICAL EXAMPLE
- The parameters of the quadrotor UAV are chosen according to a quadrotor UAV developed in [2].
- Simulation results are presented in Figures 4 and 5.
- This corresponds to Proposition 3, which implies almost global exponential attractiveness.
- After that instant, the position tracking error and the angular velocity error converge to zero as shown in Figures 5(c) and 5(d).
- The region of attraction of the proposed control system almost covers SO(3), so that the corresponding controlled quadrotor UAV can recover from being initially upside down.
V. CONCLUSION
- The authors presented a global dynamic model for a quadrotor UAV, and they developed a geometric tracking controller directly on the special Euclidean group that is intrinsic and coordinate-free, thereby avoiding the singularities of Euler angles and the ambiguities of quaternions in representing attitude.
- This controller can be extended as follows.
- But, without changing the controller structure, they can be used to follow arbitrary three-dimensional attitude commands.
- The remaining one input degree of freedom can be used to maintain the altitude as much as possible.
Did you find this useful? Give us your feedback
Citations
1,875 citations
Cites background from "Geometric tracking control of a qua..."
...Note that spatial scaling also applies to the problem with multiple keyframes but the property is less useful as the positions of all keyframes must be scaled by the same factor....
[...]
578 citations
Cites methods from "Geometric tracking control of a qua..."
...The nonlinear controller employed to follow differentiable trajectories was developed in [18], and consists of independent calculations for thrust and moments:...
[...]
421 citations
409 citations
Cites background or result from "Geometric tracking control of a qua..."
...Lee et al. (2010) showed that the two controllers result in a nonlinear controller that explicitly track trajectories in SE( 3)....
[...]
...A similar result with rotation matrices is available in Lee et al. (2010)....
[...]
...…SO( 3) and compensating for the nonlinear inertial terms gives us u2 = J ( −kReR − k e ) + × J −J ( ̂RT Rdes des − RT Rdeṡdes ) (5) If we do not consider constraints on the state or the inputs, Equations (4)–(5) achieve asymptotic convergence to specified trajectories in SE( 3) (Lee et al., 2010)....
[...]
...As shown by Lee et al. (2010), in the absence of constraints on inputs, the maximum permissible error on the angular velocity for stable hover is proportional to the inverse of the square root of the largest moment of inertia....
[...]
...If we do not consider constraints on the state or the inputs, Equations (4)–(5) achieve asymptotic convergence to specified trajectories in SE( 3) (Lee et al., 2010)....
[...]
344 citations
Cites background from "Geometric tracking control of a qua..."
...Thus, Φi(t), d dt Φi(t), d2 dt2 Φi(t), d3 dt3 Φi(t) indicate the desired position, velocity, acceleration and jerk at time t, which are the input for the non-linear controller [15] to calculate desired force and momentum for controlling the quadrotor....
[...]
...We assume the robot is able to follow our generated trajectories through a non-linear controller [15]....
[...]
References
1,328 citations
"Geometric tracking control of a qua..." refers background in this paper
...Geometric control is concerned with the development of control systems for dynamic systems evolving on nonlinear manifolds that cannot be globally identified with Euclidean spaces [12], [ 13 ], [14]....
[...]
1,074 citations
"Geometric tracking control of a qua..." refers background in this paper
...There are several university-level projects [1], [2], [3], [4], and commercial products [5], [6], [7] related to the development and application of quadrotor UAVs....
[...]
...Linear control systems such as proportionalderivative controllers or linear quadratic regulators are widely used to enhance the stability properties of an equilibrium [1], [3], [4], [8], [9]....
[...]
1,066 citations
1,010 citations
"Geometric tracking control of a qua..." refers methods in this paper
...Backstepping and sliding mode technique s are applied in [11]....
[...]
982 citations
"Geometric tracking control of a qua..." refers background in this paper
...All of these assumptions are common [19], [4], and the presented control system can readily be extended to include linear rotor dynamics studied in [11]....
[...]
Related Papers (5)
Frequently Asked Questions (20)
Q2. What is the stability of the attitude dynamics?
Since the direction of the total thrust is the third bodyfixed axis, the stability of the translational dynamics depends on the attitude tracking error.
Q3. What is the torque generated by the i-th propeller?
Since it is assumed that the first and the third propellers rotate clockwise, and the second and the fourth propellers rotate counterclockwise, when they are generating a positive thrust fi along −~b3, the torque generated by the i-th propeller can be written as τi = (−1)icτffi for a fixed constant cτf .
Q4. What is the significance of the attitude tracking error?
In short, this control system is designed so that the position tracking error converges to zero when there is no attitude tracking error, and it is properly adjusted for nonzero attitude tracking errors to achieve asymptotic stability of the complete dynamics.
Q5. What is the referred to as as almost global exponential attractiveness?
Since the region of attraction given by (28) for the attitude almost covers SO(3), this is referred to as almost global exponential attractiveness in this paper.
Q6. What is the angular velocity of the third body-fixed frame?
The total thrust f and the desired direction ~b3d of the third body-fixed axis are chosen such that if there is no attitude tracking error, the thrust vector term −fRe3 in the translational dynamics of (3) becomes this tracking controller in R3.
Q7. What is the simplest way to determine the zero equilibrium of the complete system?
For any positive constants kx, kv , the authors choose positive constants c1, c2, kR, kΩ such thatc1 < min { kv(1− α),4mkxkv(1− α) k2v(1 + α) 2 + 4mkx , √ kxm } ,(24)c2 < min { kΩ,4kΩkRλmin(J) 2k2Ωλmax(J) + 4kRλmin(J) 2 ,√ kRλmin(J),√ 22− ψ1 kRλmax(J)} , (25)λmin(W2) > 4‖W12‖2λmin(W1) . (26)Then, the zero equilibrium of the tracking errors of the complete system is exponentially stable.
Q8. What is the thrust of the quadrotor?
The translational dynamics of a quadrotor UAV is controlled by the total thrust −fRe3, where the magnitude of the total thrust f is directly controlled, and the direction of the total thrust −Re3 is along the third body-fixed axis −~b3.
Q9. What is the direction of the tracking error?
As the attitude tracking error increases, the direction of the control input term fRe3 of the translational dynamics deviates from the desired direction Rde3.
Q10. What is the simplest way to determine the angular velocity of the controller?
This requires a continuous selection of the sign of quaternions or a discontinuous control system, which are shown to be sensitive to small measurement noise [23].
Q11. What is the determinant of the rotor dynamics?
For a given translational command xd(t), the authors select the total thrust f , and the desired direction of the third body-fixed axis ~b3d to stabilize the translational dynamics.
Q12. What is the tangent bundle of the nonlinear space?
The tracking errors for the position and the velocity are given by:tex = x− xd, (6) ev = v − vd. (7)The attitude and angular velocity tracking error should be carefully chosen as they evolve on the tangent bundle of the nonlinear space SO(3).
Q13. What is the configuration of the quadrotor vehicle?
The configuration of this quadrotor UAV is defined by the location of the center of mass and the attitude with respect to the inertial frame.
Q14. What is the angular velocity of the center of mass in the body-fixed frame?
According to the definition of the rotation matrix R ∈ SO(3), the direction of the i-th body fixed axis~bi is given by Rei in the inertial frame, where e1 = [1; 0; 0], e2 = [0; 1; 0], e3 = [0; 0; 1] ∈ R3.
Q15. What is the trajectory of the proposed control system?
The region of attraction of the proposed control system almost covers SO(3), so that the corresponding controlled quadrotor UAV can recover from being initially upside down.
Q16. What is the local positive-definite about R?
(8)This is locally positive-definite about R = Rd within the region where the rotation angle between R and Rd is less than 180◦ [14].
Q17. what is the angular velocity of the body-fixed frame?
Define m ∈ R the total mass J ∈ R3×3 the inertia matrix with respect to the body-fixed frame R ∈ SO(3) the rotation matrix from the body-fixed frame to the inertial frame Ω ∈ R3 the angular velocity in the body-fixed frame978-1-4244-7744-9/10/$26.00 ©2010 IEEE 5420x ∈ R3 the location of the center of mass in the inertial frame v ∈ R3 the velocity of the center of mass in the inertial frame d ∈
Q18. What is the angular velocity of the attitude?
the region of attraction for the attitude almost covers SO(3), and the region of attraction for the angular velocity can be increased by choosing a larger controller gain kR in (18).
Q19. what is the hat map of the quadrotor?
The equations of motion of this quadrotor UAV can be written asẋ = v, (2) mv̇ = mge3 − fRe3, (3)Ṙ = RΩ̂, (4)JΩ̇ + Ω× JΩ = M, (5)where the hat map ·̂ : R3 → so(3) is defined by the condition that x̂y = x× y for all x, y ∈ R3.III.
Q20. what is the angular velocity of the center of mass in the body-fixed frame?
R the distance from the center of mass to the center of each rotor in the ~b1,~b2 plane fi ∈ R the thrust generated by the i-th propeller along the −~b3 axis τi ∈ R the torque generated by the i-th propeller about the ~b3 axisf ∈ R the total thrust, i.e., f = ∑4 i=1 fiM ∈ R3 the total moment in the body-fixed frame