MotionControlWMRsTT Slides
-
Upload
austinvishal -
Category
Documents
-
view
214 -
download
0
Transcript of MotionControlWMRsTT Slides
![Page 1: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/1.jpg)
Autonomous and Mobile RoboticsProf. Giuseppe Oriolo
Motion Control of WMRs:Trajectory Tracking
![Page 2: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/2.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 2
• feedback control is needed because the application of nominal (open-loop) control inputs would lead to unacceptable performance in practice
• kinematic models are used to design feedback laws because (1) dynamic terms can be canceled via feedback (2) wheel actuators are equipped with low-level PID loops that accept velocities as reference
• we consider a unicycle in the following
motion control
![Page 3: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/3.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 3
motion control problems
trajectory tracking(predictable transients)
posture regulation(no prior planning)
![Page 4: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/4.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 4
• the desired (reference) trajectory (xd(t), yd(t)) must be admissible: there must exist vd and !d such that
trajectory tracking
• thanks to flatness, given (xd(t), yd(t)) we can compute
![Page 5: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/5.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 5
• rather than using directly the state error qd — q , use its rotated version defined as
• we find
(e1, e2) is ep (previous figure) in a frame rotated by µ
![Page 6: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/6.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 6
• now use the input transformation
• we obtain
which is clearly invertible
lineartime-varying
nonlineartime-varying nonlinear
![Page 7: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/7.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 7
• linearize the error dynamics around the reference trajectory (e ¼ 0, hence sin e3 ¼ e3)
via approximate linearization
• define the linear feedback
• we obtain
![Page 8: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/8.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 8
• letting
• caveat: this does not guarantee asymptotic stability, unless vd and !d are constant (as on circles and lines); even in this case, asymptotic stability of the unicycle is not global (indirect Lyapunov method)
with a > 0, ³ 2 (0,1), the characteristic polynomial of A(t) becomes time-invariant
real negative
eigenvalue
pair of complexeigenvalues withnegative real part
![Page 9: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/9.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 9
• the actual velocity inputs v, ! are obtained plugging the feedbacks u1, u2 in the input transformation
• note: k2 ! 1 as vd ! 0, hence this controller can only be used with persistent cartesian trajectories (stops are not allowed)
• note:(v, !) ! (vd, !d) as e ! 0 (pure feedforward)
• global stability is guaranteed by a nonlinear version
if k1, k3 bounded, positive, with bounded derivatives
![Page 10: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/10.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 10
via input/output linearization
• choose the cartesian coordinates of point B as output
• idea: find an output whose dynamics can be made linear via feedback, i.e., with an input transformation
![Page 11: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/11.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 11
• if b 6= 0, we may set
obtaining
• differentiating wrt time
determinant = b
![Page 12: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/12.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 12
• µ is not controlled with this scheme, which is based on output error feedback (compare with the previous)
• the desired trajectory for B can be arbitrary; in particular, square corners may be included
• achieve global exponential convergence of y1, y2 to the desired trajectory letting
with k1, k2 > 0
![Page 13: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/13.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 13
simulations
n error
tracking a circle via approximate linearization
![Page 14: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/14.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 14
simulations
error
tracking an 8-figure via nonlinear feedback
![Page 15: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/15.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 15
simulations
tracking a square via i/o linearization
b = 0.75 ) the unicycle rounds the corners
![Page 16: MotionControlWMRsTT Slides](https://reader038.fdocuments.us/reader038/viewer/2022110109/577cce1f1a28ab9e788d6189/html5/thumbnails/16.jpg)
Oriolo: Autonomous and Mobile Robotics - Motion Control of WMRs: Trajectory Tracking 16
simulations
tracking a square via i/o linearization
b = 0.2 ) accurate tracking but velocities increase