A Flight Control System for

bustlingdivisionElectronics - Devices

Nov 15, 2013 (3 years and 11 months ago)

89 views

A Flight Control System for
Autonomous Helicopter

People


Group Members:



Jacky, SHEN Jie



Frank, WANG Tao



Marl, MA Mo


Supervisors:



Prof. QIU Li



Prof. LI Zexiang

Presentation Flow


Introduction


Controlling a Model Helicopter


Attitude Estimation


Hardware and Software


Results and Further Work

Introduction

What is an

Autonomous


Helicopter??

Introduction

What is Autonomous Helicopter?


An Autonomous Helicopter is a
helicopter who is fully or semi
-

controlled by on
-
board intelligence and
computing power.

Pictures are from CMU: http://www
-
2.cs.cmu.edu/afs/cs/project/chopper


What could an

Autonomous Helicopter do?


Fly to a designated area on a
prescribed path while avoiding
obstacles.


Search and locate object of interest in
the designated area.


Visually lock on to and track or, if
necessary, pursue the objects.


Send back images to a ground station
while tracking the objects.

What could an

Autonomous Helicopter do?

Pictures are from CMU: http://www
-
2.cs.cmu.edu/afs/cs/project/chopper/www/goals.html

Objectives and Goal


Step 1: On
-
board electronic system




development


Step 2: Data collection from





human controlled flights


Step 3: Algorithm simulations in PC


Step 4: On
-
board real
-
time algorithm



implementation and testing


Goal: Achieving a hover flight

Controlling a Helicopter

-

Dynamic Model

From MIT: http://gewurtz.lids.mit.edu/index.htm

Controlling a Helicopter

-

Dynamic Model


u, v, w, the velocity in x, y, z axis


p, q, r, the angle velocity in 3 axis


Θ
, pitch,
Φ
, yaw

From MIT: http://gewurtz.lids.mit.edu/index.htm

Controller Result


Successfully achieved a hover flight


Flied forward, backward and sideward


Flied on a prefixed path

Flight Controller Demonstration

A small demonstration

of autonomous

helicopter controller

Attitude Estimation

Information needed by flight
controller

The parameter we need to estimate:


Body orientation: pitch, roll, yaw

Body linear velocity vector: Vb (u, v, w)

Body angular velocity vector: Wb (p, q, r)

NED position : x, y, z




The complementary property of attitude
estimation by gyro and gravity vector

Angular
velocity
propagation

Gravity vector
tracking

Short term
accuracy

High

Low

Long term
accuracy

Diverging

Stable

The complementary property of body
velocity estimation by GPS and body
acceleration integration

GPS velocity

Acceleration
integration

Advantage


Good accuracy


Error not
accumulating


Low delay(<0.1s)


High bandwidth

(up to 100 Hz)

Disadvantage



High delay(>0.4s)


Low bandwidth

(5 Hz)


Long term
diverging

What is Kalman fitler?


The Kalman filter is a multiple
-
input,
multiple
-
output digital filter that can
optimally estimate, in real time, the states
of a system based on its noisy outputs.




The Kalman filter estimates a process by
using a form of feedback control: the filter
estimates the process state at some time
and then obtains feedback in the form of
(noisy) measurements.


The Kalman filter

x: actual state vector

z: measurement vector

w: process variance

v: measurement variance

u: control input


The Kalman filter



Gyro
noise

variance

Can be
calculated
from gyro
reading

orientation state:

Pitch, roll, yaw



1
1
1






k
k
k
k
w
Bu
Ax
x
Pitch, Roll, Yaw

Calculated from

Accelerometer

ASSUMMING the helicopter do
NOT has any body linear
acceleration.

Variance introduced
by the resulting
error of the
“acceleration free”
assumption


k
k
k
v
Hx
z


The helicopter has several strong
vibration sources

Main rotor at
29 Hz

Structural vibration at 10 Hz

Low
-
pass filtering


Fortunately vibration noises and
helicopter dynamics are not in the
same frequency, we can low
-
pass the
data to eliminate the noise.


Hardware dumper :

cutoff frequency 7
-
9Hz

FIR filter:

cutoff frequency 5Hz

Filter result


The result of our filter is satisfying


For example, The remaining noise is


+
-
0.1 m/s^2 in x axis acc sensor and
around +
-
1 degree/s in y axis gyro
reading. The noise will be further
eliminated in the Kalman filter and
integration operation.

Sensor Offset Effect

GPS Antenna

IMU Sensor

C.G.

IMU Offset Compensation

The IMU offset vector is



The accelerometer reading follows:



The largest error is introduced by the
term

T
z
y
x
IMU
R
R
R
R
]
,
,
[

IMU
R



IMU
IMU
cg
m
R
R
a
a










IMU Offset Compensation

The IMU offset compensation is



































g
R
R
a
a
t
w
t
v
t
u
body
local
IMU
measure
linear
0
0


GPS antenna offset
compensation

The GPS offset vector is


The GPS offset compensation equation is





T
GZ
GY
GX
GPS
r
r
r
R

GPS
local
body
antenna
rotorhead
R
R
z
y
x
z
y
x























GPS
offset
with
body
local
g
c
R
r
q
p
w
v
u
R
w
v
u

































_
.
Sensor offset compensation
effects


Kalman filter result

(attitude and velocity estimation)


Measure attitude from acc gps
and compass


Background:


In strap down inertia navigation filter, the
attitude information should be continuously
measured from the accelerometer, GPS, and
magnetic sensor.


In static situation the only acceleration
accelerometer sensed is the gravitational
force, the pitch and roll in the Euler angles
can be measured by the following method:

































cos
cos
sin
cos
cos
sin
sin
sin
sin
cos
sin
cos
sin
cos
cos
cos
sin
sin
sin
cos
sin
sin
sin
cos
sin
cos
sin
cos
cos







































z
y
x
R
R
R
R
'
0
0
g
g
n
gravitatio


'
'
cos
cos
sin
cos
sin
z
y
x
n
gravitatio
ter
accelerome
a
a
a
g
g
g
g
R
a












)
/
(
tan
)
/
(
sin
1
1
z
y
x
a
a
g
a











However, when in dynamic environment the
accelerometer sensed not only the static
gravitational force but also linear
acceleration which can be obtained from
derivative of GPS ground velocity reading.


Because the first and the second part of are
no longer zero so the first two column of will
make the and no longer easy to solve, thus
a good method should be developed to solve
this problem.

Introduction

of the

Hardware System

Hardware System


Electrical System






-

GPS






-

IMU






-

Compass



Mechanical Damper

Overall Electrical System

GPS



Main Feature of GPS


5 Hz Position Velocity and
Time


(PVT) output



Robust Signal Tracking


Satellite Based
Augmentation System

IMU

Main Feature of IMU


96 Hz Sampling Rate


MEMS Technology


Digital Outputs


+/
-

2g Acceleration
Measurement Range


User
-
configurable FIR Filters


Compass

Main Feature of Compass


1
°

Heading Accuracy,
0.1
°

Resolution


15Hz Response Time


UART/SPI Interface


Mechanical Dampers

Main Feature of Dampers


7
-
9 Hz Cutoff Frequency


11 Hz in horizontal plane


13 Hz in the vertical
direction

Communication

Between Devices


Overall Block Diagram

SPI Communication

ARM & Microprocessors

SPI Communication

SD Card & Microprocessor

UART

Microprocessor & Servo Motor

Result

Achievements

&

Further

Development

Achievements


PD Controller successfully
implemented


Attitude estimation


Hover Flight


Estimation Result

Further Development


Maneuver Flight Possibility



Vision Tracking


Q & A