You are on page 1of 37

IMPLEMENTATION OF A LONG RANGE LRF

BY AUTOMATIC MULTI-MODE MEASUREMENT SCHEME

Manukid PARNICHKUN and Viraphan SAMADI
School of Advanced Technologies
Asian Institute of Technology
PO. Box 4 Klong Luang, Pathumthani 12120 THAILAND
E-mail: manukid@ait.ac.th Tel: +66-2-524-5229 Fax: +66-2-524-5697

ABSTRACT

A long range Laser Range Finder (LRF) is proposed and developed in this paper. A mathematical model,

which relates the relation between object distance and Position Sensitive Detector (PSD) voltage output, is

investigated. LRF is a single point optical triangulation instrument, which detects distance information

quickly and easily without touching the object being measured. It can be used in real-time control

application. This kind of instrument is available in today’s market, however it is limited to a short range

application only (maximum range of about 400 mm). The non-linearity is inherent in this kind of

instrument, resulting the sensitivity and accuracy change with the distance nonlinearly. Sensitivity and

accuracy becomes worse abruptly when distance to the object being measured is longer than a limited

range. The proposed long range LRF is developed to achieve a longer operation range of about 1 m or

more. To improve the instrument accuracy at long range, automatic multi-mode measurement scheme is

introduced and implemented. The PSD output voltage is selected properly from a corrected mode. Then it

is shifted and scaled depending on the mode of operation before feeding to analog to digital converter

(ADC) automatically. This automatic multi-mode measurement scheme has showed the ability to make

the accuracy finer considerably depending on the number of mode being used.

1
1. INTRODUCTION

Range or distance sensor is one of the most important sensors being used in today’s manufacturing

processes. Many applications in manufacturing control use the information of the distance between two

objects. Such applications are seen in conveyor systems, identification systems, and robotics systems, for

instances. Mobile robot also needs to know the distance between external obstacle and its body to use in

path planing.

Range sensors can be divided into two categories as contact and non-contact range sensors. There is no

physical contact between sensor and target in non-contact range sensor. In this case, laser, light, acoustic

or CCD camera, etc., are used. In sonar ranging system, the distance is measured based on the return time

of an acoustic signal [1]. Infrared ranging system composes of a pair of LED emitter and a photodiode

detector. Range of an object is determined by the intensity of the light from the emitters reflected back to

the detector at the object.

The parameters which characterize range sensor are:

1) measuring range

2) sensitivity

3) accuracy

4) measuring time

5) type of technology used in measuring

6) etc.

The importance of these parameters depends upon the application using the sensor. For instance, a range

sensor fixed on the end effector of a robot arm needs to be very accurate and fast in response in order to be

use in real-time control. A range sensor, using laser technology to measure distance, is selected when it is

not convenient to make contact with the object being used also. Most of the laser range sensors available

now in the market operate at a measuring range from 4 mm to 400 mm. For example, products from

2
KEYENCE (LB-1001, LB-1101 and LB-1201) operate at a distance range from 5 mm to 100 mm and

accuracy from 2 mm to 50mm [2]. Another products from NFM Inc. (5M point sensor) operate at a

distance range from 0.5 mm to 400 mm and accuracy of 0.1 mm [3]. The accuracy is not linear for this

kind of sensor and it decreases as the sensor range increases.

This research will focus on developing a LRF, which can detect long range (about 1 m), and finding a

way to improve the sensor accuracy. The aim is contributed to the field of range sensing so that for a

certain applications and environments, engineers have more and wider options to choose a suitable sensor.

.

3
2. LASER RANGE FINDER

The principle of operation of the LRF is summarized as shown in figure 1, laser beam is pointed on a

target. The diffusely scattering light from the target is converged through a lens and projected on a

Position Sensitive Detector (PSD). The PSD generates electrical DC voltage proportional to the position

of incident light on PSD surface. This signal is processed through a signal conditioning circuit to produce

a voltage whose range is suitable to be fed to an analog to digital converter (ADC). The digital signal from

ADC is then applied to the microprocessor (Z80). Figure 2 illustrates the detail structure of the LRF

sensor. Block diagram of the electronics components is shown in figure 3. Microprocessor, which is

programmed in assembly language, displays the distance calculated from triangular relation between

target distance and incidence light on PSD surface. Furthermore, this distance information can be used in

the control loop in the closed loop control systems in manufacturing processes.

2.1 Principle of Position Sensitive Detector (PSD)

PSD which is a key component of the system has many advantages such as high position resolution, fast

response speed, simple operating circuit, and independency of the focus of the light spot. PSD consists of

three layers: P-layer at one surface, N-layer at the other surface, and I-layer between the P- and N-layers

over a planar silicon substrate. Incident light falling on the PSD is converted photoelectrically and

detected by the two electrodes on P-layer as photocurrent. Figure 6 shows the PSD sectional view.

When a light spot falls on the PSD, an electric charge proportional to the light energy is generated at

the incident position. This electric charge is driven through the resistive P-layer and collected by the

electrodes. Since the resistivity of the P-layer is uniform, the photocurrent collected by an electrode is

inversely proportional to the distance between the incident position and the electrode. It is possible to

obtain the formulas for the photocurrents I1 and I2 collected by the electrodes, where L and I0 respectively

stands for the electrode interdistance and the total photocurrent.

4
When the center point of the PSD is set as the original point, the currents ratio follows the following

equations.

1 x 
I1 =  − A  I 0 (1)
2 L 

1 x 
I 2 =  + A I 0 (2)
2 L 

I1 L − 2 x A
= (3)
I 2 L + 2x A

I 2 − I1 2 x A
= (4)
I1 − I 2 L

The structure of one-dimensional PSD along with its equivalent circuit is shown in figure 5 where P is

current generator, D is ideal diode, Cj is junction capacitance, Rsh is shunt resistance, and RD is positioning

resistance. Since the PSD has the distribution circuits Cj and Rd, its time constant acts as a decisive

element for waveform response. By using a position signal integration circuit, however, the one

dimensional PSD can be used for position detection of a laser beam of 100 ps pulse width.

2.2 Geometrical Analysis of the Sensor

The main objective in this section is to apply optical triangular method to analyze the following things:

1. the optimum position and orientation of PSD relative to the lens

2. the relation between target distance and voltage output from PSD.

So that a long range operation is possible, the optimum position and orientation of PSD relative to the lens

have to be determined. It is obviously seen that as the target is farther, the reflected light through the lens

is weaker. By the relation between object and image in lens equation, it is found that the concentration of

the scattering light is maximum (and hence the light energy) at the image position as shown in figure 6(a).

Behind the lens, shown in figure 6(b), there is a line of the image positions where the scattering light

5
focuses as spots. Positions of these spots change with the change of the target distances. If the PSD is

aligned such that the image positions are always on the PSD surface as the target moves, this

configuration guarantees maximum light intensity on the PSD surface, and hence the possible longest

operating range is achieved here.

The relation between target distance and voltage output from PSD is necessary since microprocessor

will display the distance to the target information by calculation from the PSD voltage.

2.2.1 PSD Optimum Position and Orientation Analysis

Referring to figure 6(a), the relation between the object and its image is obtained from lens equation.

1 1 1
= + (5)
f xi d

where

f : focal length

xi : distance of image from lens along x axis

d : distance of target from lens along x axis.

Equation (5) is solved for xi , the following formula is obtained:

fd
xi = (6)
d− f

The slope of the line of images, m or tan γ , is calculated and found as

l
m = tan γ = (7)
f

where l is the distance between 2 parallel lines; lens axis and laser beam.

The distance of the image along y axis, y i , is represented by straight-line equation

y i = mxi + b . (8)

6
After substitution of m and b into equation (8), we get

l
yi = xi − l . (9)
f

Substitute xi from equation (6) into equation (9), we get

ld
yi = −l. (10)
d− f

xi from equation (6) and y i from equation (10) define the position of the light spot image. Here f and

l are known whereas d is what we are finding and unknown. The straight-line equation in equation (9) is

the most appropriate model to align the PSD active surface. The measurable range depends upon the

position along the line of alignment of the PSD active surface.

2.2.2 Relation between Target Distance and Voltage Output Analysis

Consider figure 7, positions of x s and y s are defined by the maximum measurable range. If the maximum

measurable range of d max is required, x s and y s are calculated from equation (6) and (10) respectively.

fd max
xs = (11)
d max − f

ld max
ys = −l. (12)
d max − f

If the distance along the PSD is defined as s which starts as 0 at the point on PSD of the image of the

maximum measurable range, ( x s , y s ). The relation between s and d then follows equation (13).

 fd fd max   d d max  2 2
s =  −  sec γ =  −  f + l (13)
 d − f d max − f   d − f d max − f 

or

7
sfd max − sf 2 + fd max f 2 + l2
d= (14)
sd max − sf + f f 2 +l2

Consider figure 8, the relation between output voltage from C3683-01 conditioning circuit, vo , and

distance of incident light spot from the center of PSD, a , is represented by

2a
vo = 10 (15)
L

where L is the active length of PSD.

Since a = s − L / 2 , substitute all necessary variables into equation (15), we get

2( s − L / 2) 20 s 20 f 2 + l 2  d d max 
vo = 10 = − 10 =  −  − 10 . (16)
L L L  d − f d max − f 

The distance to the target in the function of the output voltage is written as

vo fLd max − vo f 2 L + 10 fLd max − 10 f 2 L + 20 fd max f 2 +l2
d= (17)
vo Ld max − vo fL + 10 Ld max − 10 fL + 20 f f 2 + l2

or

vo fA + B
d= (18)
vo A + C

where

A = Ld max − fL, B = 10 fLd max − 10 f 2 L + 20 fd max f 2 + l 2 , and C = 10 Ld max − 10 fL + 20 f f 2 + l 2 (19)

The minimum measurable range is obtained by substitution of vo = 10 V into equation (17), and get

d max (20 fL + 20 f f 2 + l 2 ) − 20 f 2 L
d min = . (20)
20d max L − 20 fL + 20 f f 2 + l2

2.3 Sensor Sensitivity Analysis

Sensitivity is one of the most important characteristics of the sensor, which needs deliberate consideration.

8
The sensitivity of LRF, S , is the ratio between output voltage from the sensor, vo , and the distance to the

target d . The sensitivity is calculated from the following equation.

dvo
S= (21)
dd

By substitution of equation (16) into equation (21), we get

d  20 f 2 + l 2  fd fd max   20 f f 2 + l 2
S=   −  − 10 = − . (22)
dd  fL  d − f d max − f   L (d − f ) 2

The sensitivity varies with the inverse square of the difference between the distance to the target and the

focal length. In the other word, the sensitivity is high at short range and it is low at long range.

2.4 Sensor Resolution and Accuracy Analysis

Since the relation between target distance and image position on PSD surface is nonlinear, it leads to the

fact that the accuracy of the sensor is not constant and even nonlinear also. Even if the target moves in

equal increment, the position of the light spot image on the PSD surface does not move in equal

increment.

There are two components influencing the overall accuracy, namely:

1. PSD accuracy

2. ADC accuracy.

It is obviously known that the component with worse accuracy defines the overall sensor accuracy.

2.4.1 PSD Resolution and Accuracy

The resolution of PSD, ∆s PSD , is defined as the minimum movement of the light spot image on the surface

of PSD that still causes a change in output voltage. This value can be found from product data sheet. It is

9
equal to 7µm for S1352 PSD from Hamamatsu, used in our prototype hardware. The accuracy, APSD , is

nonlinear and it is approximated from the following relation.

dd
APSD = ∆s PSD (23)
ds

or

 (2 f 2 d − f 3 − fd 2 ) f 2 + l 2 
APSD = max max ∆s . (24)
 ( sd − sf + f f 2 + l 2 ) 2  PSD
 max 

The worst accuracy, however, is obtained at the maximum detectable range. The difference of distances at

s = 0 or d = d max and s = ∆s PSD = 7 µm is the worst accuracy and is found exactly as

∆s PSD fd max − ∆s PSD f 2 + fd max f 2 + l2
APSD _ worst = d max − (25)
∆s PSD d max − ∆s PSD f + f f 2 +l2

2.4.2 ADC Resolution and Accuracy

The resolution of ADC is defined from the number of possible output states, and it is related to the number

of output bits from the converter. The number of state, N , is equal to the number of combinations that can

be represented by the number of bits:

N = 2n (26)

where n is the number of bits used in ADC.

The resolution of ADC, ∆s ADC , in single-mode operation is thus equal to

∆s ADC = L / N = L / 2 n . (27)

To overcome the problem of bad resolution, multi-mode measurement scheme is introduced here. This

scheme is summarized as follow:

The detectable range of the sensor is firstly divided into groups. The number of groups depends upon

the accuracy needed. The output voltage from PSD of each group is shifted and scaled to span over the
10
input range of ADC, 0-5V for instance, by different shifting voltage and scaling gain. The shifting voltage

and scaling gain are defined by the location of the group in the output voltage map. A switching circuit is

designed in order to select the active mode of operation respect to the output voltage from PSD and the

map. Comparators and transistors are the main components in the switching circuit. Output signal from

the comparator circuit is also used to acknowledge the microprocessor about the selected mode of

operation so that the microprocessor can calculate the correct value of target distance respective to the

mode.

Three modes have been implemented in our designed LRF. Figure 9 briefly illustrates the three-mode

measurement scheme.

The resolution in multi-mode measurement scheme is improved k times of the resolution in the single

mode. When k represents the gain, which is the ratio between the span of ADC input voltage and the span

of the voltage in a group being considered. In the same manner as PSD accuracy, ADC accuracy can be

approximated by the following equation.

dd
AADC = ∆s ADC (28)
ds

or

 (2 f 2 d − f 3 − fd 2 ) f 2 + l 2 
AADC =  max max ∆s . (29)
 ( sd − sf + f f 2 + l 2 ) 2  ADC
 max 

where

∆s ADC = Lg /(2 n ) . (30)

and L g is the length on PSD surface of the group being considered. In our developed LRF, number of bits

of ADC is 8 bits and number of modes is 3 modes.

The worst accuracy of ADC, AADC _ worst , in the single-mode operation is calculated in the same manner

as the worst accuracy of PSD, we get

11
∆s ADC fd max − ∆s ADC f 2 + fd max f 2 + l2
AADC _ worst = d max − (31)
∆s ADC d max − ∆s ADC f + f f 2 +l2

2.5 LRF Programming

The function of the developed program is to calculate the distance from the output voltage of PSD after

passing conditioning circuit and the circuit which realizes multi-mode measurement scheme, and display

the result in mm unit on the LCD panel. There are two inputs to microprocessor; data and control inputs.

Data input is the digital 8-bit information, which represents the analog voltage information in the range of

0-5 V. Control input is the information, which is used to inform the microprocessor about mode of

operation. The microprocessor selects an equation from the memory to calculate distance depending on

the automatically selected mode. The distance can also be found by looking-up table also. The looking-up

table approach is suitable for the system in which exact alignment, positioning, or calibration is hard to

implement. The data in each row of the table is the mid-distance between the shortest and the longest

distances, obtained by experiment, which return the same ADC output. Flowchart in figure 10 shows the

program of LRF.

12
3. RESULTS AND DISCUSSION

3.1 Results on Geometrical Analysis

In the developed LRF, values of all parameters are chosen and implemented as follow.

maximum measurable Range : d max : 1000 mm

focal length : f : 35 mm

distance between lens axis and laser beam : l : 24.5 mm

active length of PSD : L : 34 mm

resolution of PSD : ∆s PSD : 7 µm

number of bits of ADC : n : 8 bits

number of modes of operation : k : 3 modes

By substitution of all the necessary parameters, the following results are obtained.

slope of line of image,

l 24.5
m = tan γ = = (32)
f 35

or alignment angle,

 24.5 
γ = arctan  = 34.99° (33)
 35 

starting position of PSD,

fd max 35 × 1000
xs = = = 36.27 mm from lens center, (34)
d max − f 1000 − 35

ld max 24.5 × 1000
ys = −l = − 24.5 = 0.89 mm from lens center (35)
d max − f 1000 − 35

Thus the PSD must be placed 36.27 mm behind lens along x-axis, 0.89 mm above lens along y-axis, and

34.99° align backward from lens axis.
13
minimum measurable range,

d max (20 fL + 20 f f 2 + l 2 ) − 20 f 2 L
d min = .
20d max L − 20 fL + 20 f f 2 + l2

1000(20 × 35 × 34 + 20 × 35 35 2 + 24.5 2 ) − 20 × 35 2 × 34
= = 77.06 mm . (36)
20 × 1000 × 34 − 20 × 35 × 34 + 20 × 35 35 2 + 24.5 2

relation between distance along the PSD, s , and distance to target, d ,

 d d max  2 2
s =  −  f + l
 d − f d max − f 

 d 1000  42.72d
= −  35 + 24.5 =
2 2
− 44.27 (37)
 d − 35 1000 − 35  d − 35

The graph in figure 11 illustrates the relation in equation (37).

relation between voltage output from PSD, vo , and distance to target, d ,

20 f 2 + l 2  d d max 
vo =  −  − 10
L  d − f d max − f 

20 35 2 + 24.5 2  d 1000  25.13d
=  −  − 10 = − 36.04 (38)
34  d − 35 1000 − 35  d − 35

The graph in figure 12 illustrates the relation in equation (38).

3.2 Results on Sensor Sensitivity and Accuracy Analysis

The sensor sensitivity, S , is the slope of the relation between voltage output from PSD, vo , and distance

to target, d ,

20 f f 2 +l2 20 × 35 35 2 + 24.5 2 879.59
S =− =− =− (39)
L(d − f ) 2 34(d − 35) 2
(d − 35) 2

The magnitude of sensor sensitivity at each distance is illustrated in graph in figure 13.

14
Magnitude of the minimum sensitivity, S min , is obtained at the maximum target distance.

879.59 879.59
S min = − = − = 0.00094455 V/mm = 0.94455 mV/mm (40)
(d − 35) 2
(1000 − 35) 2

PSD accuracy at each location on PSD surface,

 (2 f 2 d − f 3 − fd 2 ) f 2 + l 2 
APSD =  max max ∆s .
 ( sd − sf + f f 2 + l 2 ) 2  PSD
 max 

 (2 × 35 2 × 1000 − 35 3 − 35 × 1000 2 ) 35 2 + 24.5 2 
= 0.007 = − 9747245.17 . (41)
 (1000 s − 35s + 35 35 2 + 24.5 2 ) 2  (965s + 1495.30) 2
 

The graph in figure 14 illustrates the magnitude of PSD accuracy at each location on PSD surface.

the worst PSD accuracy,

∆s PSD fd max − ∆s PSD f 2 + fd max f 2 + l2
APSD _ worst = d max −
∆s PSD d max − ∆s PSD f + f f 2 +l2

0.007 × 35 × 1000 − 0.007 × 35 2 + 35 × 1000 35 2 + 24.5 2
= 1000 − = 4.34mm (42)
0.007 × 1000 − 0.007 × 35 + 35 35 2 + 24.5 2

ADC accuracy in single-mode operation at each location on PSD surface,

 (2 f 2 d − f 3 − fd 2 ) f 2 + l 2 
AADC =  max max ∆s .
 ( sd − sf + f f 2 + l 2 ) 2  ADC
 max 

 (2 × 35 2 × 1000 − 35 3 − 35 × 1000 2 ) 35 2 + 24.5 2  34 
=   = − 184936571.3 . (43)
 (1000 s − 35s + 35 35 2 + 24.5 2 ) 2  2 8  (965s + 1495.30) 2
 

The graph in figure 15 illustrates the magnitude of ADC accuracy at each location on PSD surface.

the worst ADC accuracy in single-mode operation,

∆s ADC fd max − ∆s ADC f 2 + fd max f 2 + l2
AADC _ worst = d max −
∆s ADC d max − ∆s ADC f + f f 2 +l2

15
(34 / 2 8 ) × 35 × 1000 − (34 / 2 8 ) × 35 2 + 35 × 1000 35 2 + 24.5 2
= 1000 − = 76.18mm (44)
(34 / 2 8 ) × 1000 − (34 / 2 8 ) × 35 + 35 35 2 + 24.5 2

Since the accuracy of ADC is much worse than the accuracy of PSD, the ADC accuracy defines the

overall sensor accuracy. However, the ADC accuracy can be improved by applying of multi-mode

measurement scheme.

3.3 Implementation of Three-Mode Measurement Scheme

From the graph in figure 13, it can be concluded that the LRF sensitivity is quite low at long range. In the

other word, a change in the target distance at long range results in a small change in PSD output voltage.

At short range, however, the sensitivity is quite high, a change in the target distance results in a big change

in PSD output voltage. If mode 1 locates at the current lowest sensitivity range and mode 3 locates at the

current highest sensitivity range, the range of each mode is selected by considering from the accuracy

equation. For simplicity, the accuracy at the end point of each range is set equally. The span of mode i,

L g _ i , is thus selected as

( sd max − sf + f f 2 + l 2 )2 ×L
Lg _ i = k
s =iL / k
(45)
∑ ( sd
i =1
max − sf + f f +l )
2 2 2

s = iL / k

thus, the accuracy in mode i will be improved by

k

∑ ( sd
i =1
max − sf + f f 2 + l 2 )2
s =iL / k
(46)
( sd max − sf + f f 2 + l 2 )2
s =iL / k

times comparing with the accuracy in single-mode operation.

In the developed LRF

Mode 1:

16
(965s + 195.30) 2 × 34 123920681.9 × 34
Lg _ 1 = s =34 / 3
= = 2.48 (47)
k
(123920681.9 + 487024577.2 + 1089349828)
∑ (965s + 195.30)
i =1
2
s =34 i / 3

The accuracy in mode 1 is improved 13.71 times and represented by the following equation.

 (2 f 2 d − f 3 − fd 2 ) f 2 + l 2  L g _ 1 
AADC _ 1 = max max  
 ( sd − sf + f f 2 + l 2 ) 2  2 n 
 max 

 (2 × 35 2 × 1000 − 35 3 − 35 × 1000 2 ) 35 2 + 24.5 2  2.48  13489491.08
=  =− (48)
 (1000 s − 35s + 35 35 2 + 24.5 2 ) 2  2 
8
(965s + 1495.30) 2
 

The minimum accuracy in this mode is at s = 0 and equals to 6.03 mm. The voltage range of mode 1 starts

from -10 V to –10 +1.46 = -8.54 V.

Mode 2:

(965s + 195.30) 2 × 34 487024577.2 × 34
Lg _ 2 = s = 2×34 / 3
= = 9.74 (49)
k
(123920681.9 + 487024577.2 + 1089349828)
∑ (965s + 195.30)
i =1
2
s =34 i / 3

The accuracy in mode 2 is improved 3.49 times and represented by the following equation.

 (2 f 2 d − f 3 − fd 2 ) f 2 + l 2  Lg _ 2 
AADC _ 2 = max max  
 ( sd − sf + f f 2 + l 2 ) 2  2 n 
 max 

 (2 × 35 2 × 1000 − 35 3 − 35 × 1000 2 ) 35 2 + 24.5 2  9.74  52978888.35
=  =− (50)
 (1000 s − 35s + 35 35 2 + 24.5 2 ) 2  2 
8
(965s + 1495.30) 2
 

The minimum accuracy in this mode is at s = L g _ 1 = 2.48 and equals to 3.50 mm. The voltage range of

mode 2 starts from –8.54 V to –8.54 +5.73 = -2.81 V.

Mode 3:

(965s + 195.30) 2 × 34 1089349828 × 34
Lg _ 3 = s = 34
= = 21.78 (51)
k
(123920681.9 + 487024577.2 + 1089349828)
∑ (965s + 195.30)
i =1
2
s =34 i / 3

17
The accuracy in mode 3 is improved 1.56 times and represented by the following equation.

 (2 f 2 d − f 3 − fd 2 ) f 2 + l 2  L g _ 3 
AADC _ 3 =  max max  
 ( sd − sf + f f 2 + l 2 ) 2  2 n 
 max 

 (2 × 35 2 × 1000 − 35 3 − 35 × 1000 2 ) 35 2 + 24.5 2  21.78  18468191.8
=  =− (52)
 (1000s − 35s + 35 35 2 + 24.5 2 ) 2  2 8  (965s + 1495.30) 2
 

The minimum accuracy in this mode is at s = L g _ 1 + L g _ 2 = 12.22 and equals to 0.67 mm. The voltage

range of mode 3 starts from –2.81 V to –2.81 +12.81 = 10 V.

The magnitude of the improved ADC accuracy curve of the developed LRF by three-mode

measurement scheme is plotted comparing with the magnitude of the original ADC accuracy curve by

only single-mode operation and the magnitude of PSD accuracy curve in figure 16. From the graph, the

overall accuracy of the sensor in three-mode measurement scheme is much better than the accuracy in

single-mode measurement.

The sensor operates in the corrected mode automatically by using of the designed circuit shown in

figure 17. Comparators compare the PSD output voltage with the set threshold voltages. If the PSD output

voltage is in a range of operation of a mode, logic circuit of comparator outputs turns the switching circuit

of the selected mode on. Shifting and scaling circuits of the selected mode are then activated. The voltage

outputs from the unselected modes are disabled to 0 V. Analog switching circuits are realized by

switching transistors. Analog multiplexer is implemented by simple summing circuit here. The processed

signals from all modes are summed together. The output signal from the summing circuit is then digitized

by ADC and digitally processed by microprocessor respectively. Since microprocessor must be

acknowledged about the mode of operation in order to display the correct distance, the outputs from the

logic circuits are also fed to the I/O port of the microprocessor.

18
3.4 Description of the Developed Prototype LRF

The developed prototype LRF contains mainly two parts, the frame of instrument and the electronic circuit

board. The instrument frame shown in figure 18 consists of base frame (plastic), laser pointer, optical lens,

PSD and calipers. Laser pointer, which generates laser beam, is pointed to the target. Optical lens is fixed

on the base frame and positioned above the laser pointer. Position Sensitive Detector (PSD) is fixed on the

calipers. It is positioned behind the optical lens with the aligned angle and position determined in the

previous context. Using of calipers makes it flexibly in positioning the PSD as needed.

The electronic circuit board, shown in figure 19, consists of three parts; 1. C3683-01 circuit which is

supplied with the selected PSD, conditioning circuit, and three-mode measurement circuit.

Other view of the developed LRF is shown in figures 20 and 21.

19
4. CONCLUSION

The limitations of measurement by LRF in long range were improved very much by the concept of multi-

mode measurement scheme. As we have known, even if distance to a target changed in equal increments,

position of the image of laser spot on PSD would not change in equal increments. In the other word, LRF

is non-linear typed sensor. Generally, accuracy of LRF is quite good at short range, however, the accuracy

becomes worse at longer range. To realize good accuracy of LRF for the whole range of operation,

measurement is divided into multi modes. Number of modes of operation depends upon how accurate we

need from the LRF. Mostly, ADC accuracy which is much worse than PSD accuracy defines the accuracy

of the sensor. At long range, the operating range in this mode is short in order that the distance accuracy

would be improved. Geometrical analysis of LRF is very important. Small error in alignment, positioning,

or calibration of devices might cause a severe mistake in measurement and display. However, looking-up

table technique, in which the data in each row of the table is the mid-distance between the shortest and the

longest distances returning the same ADC output, can be used to remove the difficulties of these settings..

20
REFERENCES

[1] H.R. Everett. Sensors for Mobile Robots. AK Peters, Ltd, 1995.

[2] Keyence Corporation of America. “Measurement Instruments,”

http://wvw.keyence.com/register/le3.htm, 1998.

[3] New Features Measurement Inc. “M3 Point Sensor,” http://www.nfminc.com/html/em5.htm, 1997.

[4] Hamamatsu Photonics K.K. “Hamamatsu Technical Data Sheet,” 1995.

[5] J.F. Figueroa, J.S. Lamancusa. “A Method for Accurate Detection of Time of Arrival: Analysis and

Design of an Ultrasonic Ranging System,” Journal of the Acoustical Society of America, Vol. 91,

No. 1, pp. 486-494, 1992.

[6] D. Langer, C. Thorpe. “Sonar Based Outdoor Vehicle Navigation and Collision Avoidance,”

International Conference on Intelligent Robots and Systems, IROS ’92, 1992.

[7] R.A. Olson, R.L. Gustavson, R.E. McConnell. “Active Infrared Overhead Vehicle Sensor,” IEEE

Transaction on Vehicular Technology, Vol. 43, No. 1, pp. 79-85, 1994.

[8] M. Hansen, P. Anandan, K. Dana, van der Wal, P. Burt. “Real-Time Scene Stabilization and Mosaic

Construction,” 1994 Image Understanding Workshop, pp. 457-465, 1994.

[9] S.J. King, C.F.R. Weiman. “HelpMate Autonomous Mobile Robot Navigation System, SPIE,

Vol. 1388, Mobile Robots V, pp. 190-198, 1990.

[10] R.A. Jarvis. “A Laser Time-of-Flight Range Scanner for Robotic Vision,” IEEE Transaction on

Pattern Analysis and Machine Intelligence, Vol. PAMI-5, No. 5, pp. 505-512, 1983.

[11] L.H. Matthies. “Stereo Vision for Planetary Rovers: Stochastic Modeling to Near-Real-Time

Implementation,” International Journal of Computer Vision, Vol. 8, No. 1, 1992.

[12] A.J. Wavering, J.C. Fiala, K.J. Roberts, R. Lumia. “TRICLOPS: A High-Powered Trinocular Active

Vision System,” IEEE International Conference on Robotics and Automation, pp. 410-417, 1993.

[13] R.P. Wilders. “Direct Recovery of Three-Dimensional Scene Geometry from Binocular Stereo

21
Disparity,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 13, No. 8,

pp. 761-774, 1991.

[16] K.J. Hanna. “Direct Multi-Resolution Estimation of Ego-Motion and Structure from Motion,”

Proceedings of IEEE Workshop on Visual Motion, pp. 156-162, 1991.

[15] E.W. Kent. “Real Time Cooperative Interaction between Structured Light and Reflectance Ranging

for Robot Guidance,” Robotica, Vol. 3, pp. 7-11, 1985.

22
FIGURES

PSD
Laser
Diode
Lens

Laser
Beam

Scattering
Light Target 1

Target 2

Figure 1 Principle of Laser Range Finder

Calipers
PSD

Lens
Active Area
of PSD

to C3683-01
Circuit Laser Diode

Figure 2 Structure of Laser Range Finder
23
from PSD

C3683-01
Multi-Mode ADC
Conditioning Measurement Microprocessor
Scheme Circuit Circuit
Circuit

Figure 3 Block Diagram of Electronics Components of Laser Range Finder

Incident
Output Light Output
I1 I2
Photocurrent
Electrode Electrode
P-Layer

I-Layer

N-Layer

xA

L

Figure 4 PSD Sectional View

24
(1) (1) (2)
RD

P D Cj Rsh
(2)
(3) (Bias)

(3)
Figure 5 Structure and Equivalent Circuit of One-Dimensional PSD

xi

yi
γ l

f d

(a)

(b)

Figure 6 Image Positions behind Lens

25
L
s

ys

xs dmax

Figure 7 Image Position at Maximum Measurable Range

a

L/2 L/2

10 V

-10 V

Figure 8 Relation between Output Voltage from Conditioning Circuit and Distance of Incident Light Spot
from PSD Center

26
Conditioning
From PSD Circuit Low Pass Filter Switching Circuit

Mode 1 Mode 2 Mode 3
Shifting and Shifting and Shifting and
Scaling Circuit Scaling Circuit Scaling Circuit

Analog
Multiplexer ADC Microprocessor

Figure 9 Three-Mode Measurement Scheme

27
Start

Set I/O Ports,
Define Parameters,
Initialize Values

Read Input Data and
Mode of Operation
from Input Port

Yes Calculate Distance to
Mode 1 Target from Equation
in Mode 1

No

Yes Calculate Distance to Display Distance to
Mode 2 Target from Equation Target on LCD
in Mode 2

No

Yes Calculate Distance to
Mode 3
Target from Equation
in Mode 3

No

Figure 10 Flowchart of LRF Programming

28
Figure 11 Relation between Distance to Target (d) and Distance along PSD (s)

29
Figure 12 Relation between Distance to Target (d) and Voltage Output (vo)

30
Figure 13 Relation between Distance to Target (d) and Magnitude of Sensor Sensitivity (|S|)

31
Figure 14 Relation between Distance along PSD (s) and Magnitude of PSD Accuracy (APSD)

32
Figure 15 Relation between Distance along PSD (s) and Magnitude of ADC Accuracy (AADC)

33
Figure 16 Relation between Distance along PSD (s) and Magnitude of Accuracies (A)

34
shifting scaling
circuit circuit
comparator switching
circuit circuit
op-amp
op-amp
npn-transistor vref
vref
not-gate op-amp
comparator op-amp
npn-transistor vref
vin
nand-gate op-amp op-amp
npn-transistor vref
vref
comparator

op-amp
op-amp
vout

summing conversion
circuit circuit

Figure 17 Multi-Mode Measurement Scheme Circuit of Laser Range Finder

35
Figure 18 Frame of Laser Range Finder

Figure 19 Electronic Circuit of Laser Range Finder

36
Figure 20 Laser Range Finder

Figure 21 Other View of Laser Range Finder

37