Laser Deflection System: Disturbance Correction

1
Laser Deflection SystemDisturbance Correction

• Final Presentation
• Team 5
• April 23, 2003
• By Tyler Ferman
• Matt DiLeo
• Jack Damerji

2
Laser Disturbance Correction

• Goals movie - movingpantilt.mpeg
• Correct for a measurable input disturbance.
• Redirect laser to target according to measured
  disturbance of input trajectory.
• Objectives
• Develop accurate controller in order to keep a
  laser communication link.
• Develop system to measure input trajectory
  disturbance.

3
Original Specifications

• Input Laser Pen
• Range of motion 53o
• Location 5 x 6 x 0
• Assume user input of 0.1 sec to travel across
  mirror
• Controller
• 5 mirror mounted on center of each axis
• Range of motion 35o
• Settling time 0.1s
• Overshoot lt 1
• Output Point on screen
• 36 away

4
Original design Constraints

• First pan-tilt modified to hold a laser pen.
• cheap and accurate
• Narrows input to 2 DOF
• Second pan-tilt modified to carry a mirror.
• Challenges
• Accurately calculating input
• Positioning
• Calculation of desired mirror angles
• Developing fast and accurate controller

5
Project construction and functional tests

• Construction movie - showcase.mpeg
• Input Pan-Tilt
• Controller Pan-Tilt
• Mounting both system on one plate
• Friction measurements
• TiltAverage Viscous Friction .002 Coulomb
  Friction 0.18
• PanAverage Viscous Friction .0005 Coulomb
  Friction 0.08

6
Controller Design

• Linearizing System
• Finding a PID compensator
• Simulating the compensator on nonlinear system

7
Controller Design

• Pan
• Tilt

8
Step Response Pan Side
9
Step Response Tilt Side
10
Actual Performance Results 1
Hit rate 100 Avg pan err 0.0029 Avg tilt err
0.0029
movie - target.mpeg
11
Actual Performance Results 2
Hit rate 99.6 Avg pan err   0.0045 Avg tilt
err 0.49
12
Actual Performance Results 3
Hit rate 85.3 Avg pan err  0.0044 Avg tilt
err 0.0057
13
Actual Performance Results 4
Hit rate 65.8 Avg pan err   0.0091 Avg tilt
err 0.0052
movie - crazyfreq.mpeg
14
Sinusoidal Response Pan side
15
Sinusoidal Response Tilt side
16
Comparison of Performance
Specification Actual Performance
I/P Range of motion 53o 0.925 rad Pan 54o 0.94 rad Tilt 40o 0.70 rad
Controller range of motion 35o 0.61 rad Pan 26.4o 0.46 rad Tilt 40.8o 0.71 rad
Speed 6 rad/s 3rad/s
Error 1 cm 1 in
Settling time lt 0.1s Pan,- 0.28s, 0.36s Tilt,- 0.58s, 0.52s
Percent Overshoot lt 1 Pan,- 0, 1.5 Tilt,- 0, 0
Steady State Error lt 0.1o Pan,- -0.46, 0.006 Tilt,- 0.34, 0.57
17
System Improvement

• Max Disturbance without controller
• 13in on average from each side
• VS
• Max Disturbance with controller
• 1in from each side
• Movie closeup.mpeg

18
Success and challenges

• Success movie - mirrorview.mpeg
• Robust Controller
• Accurate calculation for desired angles using
  math model
• 1300 improvement of disturbance rejection
• Quick interaction between input pan-tilt and
  controller pan-tilt

19
Success and challenges

• Challenges
• Discrepancy between system model simulation and
  physical system
• Initialization of input and mirror angles
• Quantization Effects
• Steady-state error
• Oscillation due to derivative control
• Design controller for random input (different
  speeds/frequencies)

20
Recommendations

• Adaptive controller to allow control for random
  input
• Calibration system
• Use Kalman filter to reduce quantization effects
• Recalculate mass matrix, inertia matrix and
  friction calculation

21
Questions