(Version 3.04, Feb
14, 2002)
(revisions by John
Hansen)
Department of Computer Science
Utrecht University
P.O. Box 80.089,
3508 TB Utrecht
the Netherlands
The Lego MindStorms, CyberMaster, and Spybotics robots are
wonderful new toys from which a wide variety of robots can be constructed, that
can be programmed to do all sorts of complicated tasks. Unfortunately, the
software that comes with the robots is, although visually attractive, rather
limited in functionality. Hence, it can only be used for simple tasks. To
unleash the full power of the robots, you need a different programming
environment. NQC is a programming language, written by Dave Baum, which was
especially designed for the Lego robots. If you have never written a program
before, don't worry. NQC is really easy to use and this tutorial will tell you
all about it. Actually, programming the robots in NQC is a lot easier than
programming a normal computer, so this is a chance to become a programmer in an
easy way.
To make writing programs even easier, there is the Bricx
Command Center. This utility helps you to write your programs, to send them to
the robot, and to start and stop the robot. Bricx Command Center works almost
like a
text processor, but with some extras. This tutorial will use Bricx Command
Center (version 3.3 or higher) as programming environment. You can download it
for free from the web at the address
Bricx Command Center runs on Windows PC’s (95, 98, ME, NT,
2K, XP). The language NQC can also be used on other platforms. You can download
it from the web at address
Most of this tutorial also applies to the other platforms
(assuming you use NQC version 2.0 or higher), except that you loose some of the
tools and the color-coding.
In this tutorial I assume that you have the MindStorms (RCX)
robot. Most of the contents also applies to the CyberMaster, Scout, and Spybot robots
although some of the functionality is not available for those robots. Also the
names of e.g. the motors are different so you will have to change the examples
a little bit to make them work.
I would like to thank Dave Baum for developing NQC. Also
many thanks to Kevin Saddi for writing a first version of the first part of
this tutorial.
Preface____________________________________________________________________ 2
Acknowledgements__________________________________________________________________ 2
Contents___________________________________________________________________ 3
I. Writing your first program___________________________________________________ 5
Building a robot_____________________________________________________________________ 5
Starting Bricx Command Center________________________________________________________ 5
Writing the program__________________________________________________________________ 6
Running the program_________________________________________________________________ 7
Errors in your program________________________________________________________________ 7
Changing the speed__________________________________________________________________ 8
Summary_________________________________________________________________________ 8
II. A more interesting program_________________________________________________ 9
Making turns_______________________________________________________________________ 9
Repeating commands________________________________________________________________ 9
Adding comment___________________________________________________________________ 10
Summary________________________________________________________________________ 11
III. Using variables_________________________________________________________ 12
Moving in a spiral__________________________________________________________________ 12
Random numbers__________________________________________________________________ 13
Summary________________________________________________________________________ 13
IV. Control structures_______________________________________________________ 14
The if statement___________________________________________________________________ 14
The do statement__________________________________________________________________ 15
Summary________________________________________________________________________ 15
V. Sensors________________________________________________________________ 16
Waiting for a sensor_________________________________________________________________ 16
Acting on a touch sensor_____________________________________________________________ 16
Light sensors______________________________________________________________________ 17
Summary________________________________________________________________________ 18
VI. Tasks and subroutines____________________________________________________ 19
Tasks___________________________________________________________________________ 19
Subroutines_______________________________________________________________________ 20
Inline functions____________________________________________________________________ 20
Defining macro’s___________________________________________________________________ 21
Summary________________________________________________________________________ 22
VII. Making music_________________________________________________________ 23
Built-in sounds____________________________________________________________________ 23
Playing music_____________________________________________________________________ 23
Summary________________________________________________________________________ 24
VIII. More about motors_____________________________________________________ 25
Stopping gently____________________________________________________________________ 25
Advanced commands_______________________________________________________________ 25
Varying motor speed________________________________________________________________ 26
Summary________________________________________________________________________ 26
IX. More about sensors______________________________________________________ 27
Sensor mode and type_______________________________________________________________ 27
The rotation sensor_________________________________________________________________ 28
Putting multiple sensors on one input____________________________________________________ 28
Making a proximity sensor___________________________________________________________ 29
Summary________________________________________________________________________ 30
X. Parallel tasks____________________________________________________________ 31
A wrong program___________________________________________________________________ 31
Stopping and restarting tasks__________________________________________________________ 31
Using semaphores__________________________________________________________________ 32
Summary________________________________________________________________________ 33
XI. Communication between robots____________________________________________ 34
Giving orders______________________________________________________________________ 34
Electing a leader___________________________________________________________________ 35
Cautions_________________________________________________________________________ 35
Summary________________________________________________________________________ 36
XII. More commands_______________________________________________________ 37
Timers___________________________________________________________________________ 37
The display_______________________________________________________________________ 37
Datalogging_______________________________________________________________________ 38
XIII. Final remarks_________________________________________________________ 39
In this chapter I will show you how to write an extremely
simple program. We are going to program a robot to move forwards for 4 seconds,
then backwards for another 4 seconds, and then stop. Not very spectacular but
it will introduce you to the basic idea of programming. And it will show you
how easy this is. But before we can write a program, we first need a robot.
The robot we will use throughout this tutorial is a simple
version of the top-secret robot that is described on page 39-46 of your
constructopedia. We will only use the basis chassis. Remove the whole front
with the two arms and the touch sensors. Also, connect the motors slightly
different such that the wires are connected to the RCX at the outside. This is
important for your robot to drive in the correct direction. Your robot should
look like this:
Also make sure that the infra-red port is correctly
connected to your computer and that it is set to long range. (You might want to
check with the RIS software that the robot is functioning well.)
We write our programs using Bricx Command Center. Start it
by double clicking on the icon BricxCC. (I assume you already installed Bricx
Command Center. If not, download it from the web site (see the preface), and install
it in any directory you like.) The program will ask you where to locate the
robot. Switch the robot on and press OK.
The program will (most likely) automatically find the robot. Now the user
interface appears as shown below (without a window).
The interface looks like a standard text editor, with the
usual menu’s, and buttons to open and save files, print files, edit files, etc.
But there are also some special menus for compiling and downloading programs to
the robot and for getting information from the robot. You can ignore these for
the moment.
We are going to write a new program. So press the New File button to create a new, empty
window.
Now type in the following program:
task main()
{
OnFwd(OUT_A);
OnFwd(OUT_C);
Wait(400);
OnRev(OUT_A+OUT_C);
Wait(400);
Off(OUT_A+OUT_C);
}
It might look a bit complicated at first, so let us analyze
it. Programs in NQC consist of tasks. Our program has just one task, named main.
Each program needs to have a task called main which is the one that will
be executed by the robot. You will learn more about tasks in Chapter VI.
A task consists of a number of commands, also called statements. There are
brackets around the statements such that it is clear that they all belong to
this task. Each statement ends with a semicolon. In this way it is clear where
a statement ends and where the next statement begins. So a task looks in
general as follows:
task main()
{
statement1;
statement2;
…
}
Our program has six statements. Let us look at them one at
the time:
OnFwd(OUT_A);
This statement tells the robot to start output A, that is,
the motor connected to the output labeled A on the RCX, to move forwards. It
will move with maximal speed, unless you first set the speed. We will see later
how to do this.
OnFwd(OUT_C);
Same statement but now we start motor C. After these two statements,
both motors are running, and the robot moves forwards.
Wait(400);
Now it is time to wait for a while. This statement tells us
to wait for 4 seconds. The argument, that is, the number between the
parentheses, gives the number of “ticks”. Each tick is 1/100 of a second. So
you can very precisely tell the program how long to wait. So for 4 seconds, the
program does do nothing and the robot continues to move forwards.
OnRev(OUT_A+OUT_C);
The robot has now moved far enough so we tell it to move in
reverse direction, that is, backwards. Note that we can set both motors at once
using OUT_A+OUT_C as argument. We could also have combined
the first two statements this way.
Wait(400);
Again we wait for 4 seconds.
Off(OUT_A+OUT_C);
And finally we switch both motors off.
That is the whole program. It moves both motors forwards for
4 seconds, then backwards for 4 seconds, and finally switches them off.
You probably noticed the colors when typing in the program.
They appear automatically. The colors
and styles used by the editor when it performs syntax highlighting are
customizable.
Once you have written a program, it needs to be compiled
(that is, changed into code that the robot can understand and execute) and send
to the robot using the infra red link (called “downloading” the program). There
is a button that does both at once (see the figure above). Press this button
and, assuming you made no errors when typing in the program, it will correctly
compile and be downloaded. (If there are errors in your program you will be
notified; see below.)
Now you can run your program. To this end press the green
run button on your robot or, more easily, press the run button on your window
(see the figure above). Does the robot do what you expected? If not, the wires
are probably connected wrong.
When typing in programs there is a reasonable chance that
you make some errors. The compiler notices the errors and reports them to you
at the bottom of the window, like in the following figure:
It automatically selects the first error (we mistyped the
name of the motor). When there are more errors, you can click on the error
messages to go to them. Note that often errors at the beginning of the program
cause other errors at other places. So better only correct the first few errors
and then compile the program again. Also note that the syntax highlighting
helps a lot in avoiding errors. For example, on the last line we typed Of rather than Off. Because this is
an unknown command it is not highlighted.
There are also errors that are not found by the compiler. If
we had typed OUT_B
this would have gone unnoticed because that motor exists (even though we do not
use it in the robot). If your robot exhibits unexpected behavior, there is most
likely something wrong in your program.
As you noticed, the robot moved rather fast. Default the
robot moves as fast as it can. To change the speed you can use the command SetPower(). The power is a number
between 0 and 7. 7 is the fastest, 0 the slowest (but the robot will still
move). Here is a new version of our program in which the robot moves slow:
task main()
{
SetPower(OUT_A+OUT_C,2);
OnFwd(OUT_A+OUT_C);
Wait(400);
OnRev(OUT_A+OUT_C);
Wait(400);
Off(OUT_A+OUT_C);
}
In this chapter you wrote your first program in NQC, using Bricx
Command Center. You should now know how to type in a program, how to download
it to the robot and how to let the robot execute the program. Bricx Command
Center can do many more things. To find out about them, read the documentation
that comes with it. This tutorial will primarily deal with the language NQC and
only mention features of Bricx Command Center when you really need them.
You also learned some important aspects of the language NQC.
First of all, you learned that each program has one task named main that is always executed by
the robot. Also you learned the four most important motor commands: OnFwd(), OnRev(), SetPower() and Off().
Finally, you learned about the Wait()
statement.
Our first program was not very spectacular. So let us try to
make it more interesting. We will do this in a number of steps, introducing
some important features of our programming language NQC.
You can make your robot turn by stopping or reversing the
direction of one of the two motors. Here is an example. Type it in, download it
to your robot and let it run. It should drive a bit and then make a 90-degree
right turn.
task main()
{
OnFwd(OUT_A+OUT_C);
Wait(100);
OnRev(OUT_C);
Wait(85);
Off(OUT_A+OUT_C);
}
You might have to try some slightly different numbers than
85 in the second Wait() command to make a precise
90-degree turn. This depends on the type of surface on which the robot runs.
Rather than changing this in the program it is easier to use a name for this
number. In NQC you can define constant values as shown in the following
program.
#define MOVE_TIME 100
#define TURN_TIME 85
task main()
{
OnFwd(OUT_A+OUT_C);
Wait(MOVE_TIME);
OnRev(OUT_C);
Wait(TURN_TIME);
Off(OUT_A+OUT_C);
}
The first two lines define two constants. These can now be
used throughout the program. Defining constants is good for two reasons: it
makes the program more readable, and it is easier to change the values. Note
that Bricx Command Center gives the define statements its own color. As we will
see in Chapter VI,
you can also define things other than constants.
Let us now try to write a program that makes the robot drive
in a square. Going in a square means: driving forwards, turning 90 degrees,
driving forwards again, turning 90 degrees, etc. We could repeat the above
piece of code four times but this can be done a lot easier with the repeat
statement.
#define MOVE_TIME 100
#define TURN_TIME 85
task main()
{
repeat(4)
{
OnFwd(OUT_A+OUT_C);
Wait(MOVE_TIME);
OnRev(OUT_C);
Wait(TURN_TIME);
}
Off(OUT_A+OUT_C);
}
The number behind the repeat statement, between
parentheses, indicates how often something must be repeated. The statements
that must be repeated are put between brackets, just like the statements in a
task. Note that, in the above program, we also indent the statements. This is
not necessary, but it makes the program more readable.
As a final example, let us make the robot drive 10 times in
a square. Here is the program:
#define MOVE_TIME 100
#define TURN_TIME 85
task main()
{
repeat(10)
{
repeat(4)
{
OnFwd(OUT_A+OUT_C);
Wait(MOVE_TIME);
OnRev(OUT_C);
Wait(TURN_TIME);
}
}
Off(OUT_A+OUT_C);
}
There is now one repeat statement inside the other. We call
this a “nested” repeat statement. You can nest repeat statements as much as you
like. Take a careful look at the brackets and the indentation used in the
program. The task starts at the first bracket and ends at the last. The first
repeat statement starts at the second bracket and ends at the fifth. The
second, nested repeat statement starts at the third bracket and ends at the
fourth. As you see the brackets always come in pairs, and the piece between the
brackets we indent.
To make your program even more readable, it is good to add
some comment to it. Whenever you put //
on a line, the rest of that line is ignored and can be used for comments. A
long comment can be put between /*
and */. Comments are syntax
highlighted in the Bricx Command Center. The full program could look as
follows:
/* 10 SQUARES
by Mark Overmars
This program make the robot run 10 squares
*/
#define MOVE_TIME 100 // Time for a straight move
#define TURN_TIME 85 // Time for turning 90 degrees
task main()
{
repeat(10) // Make 10
squares
{
repeat(4)
{
OnFwd(OUT_A+OUT_C);
Wait(MOVE_TIME);
OnRev(OUT_C);
Wait(TURN_TIME);
}
}
Off(OUT_A+OUT_C); // Now turn the motors off
}
In this chapter you learned the use of the repeat
statement and the use of comment. Also you saw the function of nested brackets
and the use of indentation. With all you know so far you can make the robot
move along all sorts of paths. It is a good exercise to try and write some
variations of the programs in this chapter before continuing with the next
chapter.
Variables form a very important aspect of every programming
language. Variables are memory locations in which we can store a value. We can
use that value at different places and we can change it. Let me describe the
use of variables using an example.
Assume we want to adapt the above program in such a way that
the robot drives in a spiral. This can be achieved by making the time we sleep
larger for each next straight movement. That is, we want to increase the value
of MOVE_TIME each
time. But how can we do this? MOVE_TIME is a
constant and constants cannot be changed. We need a variable instead. Variables
can easily be defined in NQC. You can have 32 of these, and you can give each
of them a separate name. Here is the spiral program.
#define TURN_TIME 85
int
move_time; // define a
variable
task main()
{
move_time = 20; // set the initial value
repeat(50)
{
OnFwd(OUT_A+OUT_C);
Wait(move_time); // use the variable for sleeping
OnRev(OUT_C);
Wait(TURN_TIME);
move_time += 5; // increase the variable
}
Off(OUT_A+OUT_C);
}
The interesting lines are indicated with the comments. First
we define a variable by typing the keyword int followed by a name we
choose. (Normally we use lower-case letters for variable names and uppercase
letters for constants, but this is not necessary.) The name must start with a
letter but can contain digits and the underscore sign. No other symbols are
allowed. (The same applied to constants, task names, etc.) The strange word int stands for integer. Only integer numbers can be stored in
it. In the second interesting line we assign the value 20 to the variable. From
this moment on, whenever you use the variable, it stands for 20. Now follows
the repeat loop in which we use the variable to indicate the time to sleep and,
at the end of the loop we increase the value of the variable with 5. So the
first time the robot sleeps 20 ticks, the second time 25, the third time 30,
etc.
Besides adding values to a
variable we can also multiply a variable with a number using *=, subtract using -= and divide using /=. (Note that for division the result is rounded to the
nearest integer.) You can also add one variable to the other, and write down
more complicated expressions. Here are some examples:
int aaa;
int bbb, ccc;
task main()
{
aaa = 10;
bbb
= 20 * 5;
ccc = bbb;
ccc /= aaa;
ccc -= 5;
aaa = 10 * (ccc + 3);
// aaa
is now equal to 80
}
Note
on the first two lines that we can define multiple variables in one line. We
could also have combined all three of them in one line.
In all the above programs we defined exactly what the robot
was supposed to do. But things get a lot more interesting when the robot is
going to do things that we don’t know. We want some randomness in the motions.
In NQC you can create random numbers. The following program uses this to let
the robot drive around in a random way. It constantly drives forwards for a
random amount of time and then makes a random turn.
int move_time, turn_time;
task main()
{
while(true)
{
move_time = Random(60);
turn_time = Random(40);
OnFwd(OUT_A+OUT_C);
Wait(move_time);
OnRev(OUT_A);
Wait(turn_time);
}
}
The program defines two variables, and then assigns random
numbers to them. Random(60) means a random number between
0 and 60 (it can also be 0 or 60). Each time the numbers will be different.
(Note that we could avoid the use of the variables by writing e.g. Wait(Random(60)).)
You also see a new type of loop here. Rather that using the
repeat statement we wrote while(true). The while statement
repeats the statements below it as long as the condition between the
parentheses is true. The special word true is always true, so the statements
between the brackets are repeated forever, just as we want. You will learn more
about the while statement in Chapter IV.
In this chapter you learned about the use of variables.
Variables are very useful but, due to restrictions of the robots, they are a
bit limited. You can define only 32 of them and they can store only integers.
But for many robot tasks this is good enough.
You also learned how to create random numbers, such that you
can give the robot unpredictable behavior. Finally we saw the use of the while
statement to make an infinite loop that goes on forever.
In the previous chapters we saw the repeat and while
statements. These statements control the way the other statements in the
program are executed. They are called “control structures”. In this chapter we
will see some other control structures.
Sometimes you want that a particular part of your program is
only executed in certain situations. In this case the if statement is used. Let
me give an example. We will again change the program we have been working with
so far, but with a new twist. We want the robot to drive along a straight line
and then either make a left or a right turn. To do this we need random numbers
again. We pick a random number between 0 and 1, that is, it is either 0 or 1.
If the number is 0 we make a right turn; otherwise we make a left turn. Here is
the program:
#define MOVE_TIME 100
#define TURN_TIME 85
task main()
{
while(true)
{
OnFwd(OUT_A+OUT_C);
Wait(MOVE_TIME);
if (Random(1) == 0)
{
OnRev(OUT_C);
}
else
{
OnRev(OUT_A);
}
Wait(TURN_TIME);
}
}
The if statement looks a bit like the while statement. If
the condition between the parentheses is true the part between the brackets is
executed. Otherwise, the part between the brackets after the word else
is executed. Let us look a bit better at the condition we use. It reads Random(1) == 0. This means that Random(1) must be equal to 0 to make
the condition true. You might wonder why we use == rather than =. The reason is
to distinguish it from the statement that put a value in a variable. You can
compare values in different ways. Here are the most important ones:
== equal to
< smaller than
<= smaller than or equal to
> larger than
>= larger than or equal to
!= not equal to
You can combine conditions use &&,
which means “and”, or ||,
which means “or”. Here are some examples of conditions:
true always
true
false never
true
ttt != 3 true when ttt is not
equal to 3
(ttt >= 5) && (ttt <= 10) true when ttt lies between 5 and 10
(aaa == 10) || (bbb == 10) true if either aaa or bbb (or both) are equal
to 10
Note that the if statement has two parts. The part
immediately after the condition, which is executed when the condition is true,
and the part after the else, which is executed when the condition is false. The
keyword else and the part after it are optional. So you can leave them away if
there is nothing to do when the condition is false.
There is another control structure, the do statement. It has
the following form:
do
{
statements;
}
while (condition);
The statements between the brackets after the do part are
executed as long as the condition is true. The condition has the same form as
in the if statement described above. Here is an example of a program. The robot
runs around randomly for 20 seconds and then stops.
int move_time, turn_time,
total_time;
task main()
{
total_time = 0;
do
{
move_time = Random(100);
turn_time = Random(100);
OnFwd(OUT_A+OUT_C);
Wait(move_time);
OnRev(OUT_C);
Wait(turn_time);
total_time += move_time; total_time +=
turn_time;
}
while
(total_time < 2000);
Off(OUT_A+OUT_C);
}
Note in this example that we placed two statements on one
line. This is allowed. You can place as many statements on a line as you like
(as long as there are semicolons in between). But for readability of the
program this is often not a good idea.
Note also that the do statement behaves almost the same as
the while statement. But in the while statement the condition is tested before
executing the statements, while in the do statement the condition is tested at
the end. For the while statement, the statements might never be executed, but
for the do statement they are executed at least once.
In this chapter we have seen two new control structures: the
if statement and the do statement. Together with the repeat statement and the
while statement they are the statements that control the way in which the
program is executed. It is very important that you understand what they do. So
better try some more examples yourself before continuing.
We also saw that we can place multiple statements on a line.
One of the nice aspects of the Lego robots is that you can
connect sensors to them and that you can make the robot react to the sensors.
Before I can show how to do this we must change the robot a bit by adding a
sensor. To this end, build the sensor construction shown in figure 4 on page 28
of the constructopedia. You might want to make it slightly wider, such that
your robot looks as follows:
Connect the sensor to input 1 on the RCX.
Let us start with a very simple program in which the robot
drives forwards until it hits something. Here it is:
task main()
{
SetSensor(SENSOR_1,SENSOR_TOUCH);
OnFwd(OUT_A+OUT_C);
until
(SENSOR_1 == 1);
Off(OUT_A+OUT_C);
}
There are two important lines here. The first line of the
program tells the robot what type of sensor we use. SENSOR_1 is the number of the
input to which we connected the sensor. The other two sensor inputs are called SENSOR_2 and SENSOR_3. SENSOR_TOUCH indicates that this is a touch sensor. For the light
sensor we would use SENSOR_LIGHT. After we specified the
type of the sensor, the program switches on both motors and the robot starts
moving forwards. The next statement is a very useful construction. It waits
until the condition between the brackets is true. This condition says that the
value of the sensor SENSOR_1 must be 1, which means
that the sensor is pressed. As long as the sensor is not pressed, the value is
0. So this statement waits until the sensor is pressed. Then we switch off the
motors and the task is finished.
Let us now try to make the robot avoid obstacles. Whenever
the robot hits an object, we let it move back a bit, make a turn, and then
continue. Here is the program:
task main()
{
SetSensor(SENSOR_1,SENSOR_TOUCH);
OnFwd(OUT_A+OUT_C);
while
(true)
{
if
(SENSOR_1 == 1)
{
OnRev(OUT_A+OUT_C); Wait(30);
OnFwd(OUT_A); Wait(30);
OnFwd(OUT_A+OUT_C);
}
}
}
As in the previous example, we first indicate the type of
the sensor. Next the robot starts moving forwards. In the infinite while loop
we constantly test whether the sensor is touched and, if so, move back for 1/3
of a second, turn right for 1/3 of a second, and then continue forwards again.
Besides touch sensors, you also get a light sensor with your
MindStorms system. The light sensor measures the amount of light in a
particular direction. The light sensor also emits light. In this way it is
possible to point the light sensor in a particular direction and make a
distinction between the intensity of the object in that direction. This is in
particular useful when trying to make a robot follow a line on the floor. This
is what we are going to do in the next example. We first need to attach the
light sensor to the robot such that it is in the middle of the robot, at the
front, and points downwards. Connect it to input 2. For example, make a
construction as follows:
We also need the race track that comes with the RIS kit
(This big piece of paper with the black track on it.) The idea now is that the
robot makes sure that the light sensor stays above the track. Whenever the
intensity of the light goes up, the light sensor is off the track and we need
to adapt the direction. Here is a very simple program for this that only works
if we travel around the track in clockwise direction.
#define THRESHOLD 40
task main()
{
SetSensor(SENSOR_2,SENSOR_LIGHT);
OnFwd(OUT_A+OUT_C);
while (true)
{
if (SENSOR_2 > THRESHOLD)
{
OnRev(OUT_C);
until (SENSOR_2 <= THRESHOLD);
OnFwd(OUT_A+OUT_C);
}
}
}
The program first indicates that sensor 2 is a light sensor.
Next it sets the robot to move forwards and goes into an infinite loop.
Whenever the light value is bigger than 40 (we use a constant here such that
this can be adapted easily, because it depends a lot on the surrounding light)
we reverse one motor and wait till we are on the track again.
As you will see when you execute the program, the motion is
not very smooth. Try adding a Wait(10) command before the until command to make the robot move better. Note that the program
does not work for moving counter-clockwise. To enable motion along arbitrary
path a much more complicated program is required.
In this chapter you have seen how to work with touch sensors
and light sensors. We also saw the until
command that is useful when using sensors.
I recommend you to write a number of programs yourself at
his stage. You have all the ingredients to give your robots pretty complicated
behavior now. For example, try to put two touch sensors on your robot, one on
the left front and the other on the right front, and make the robot move away
from the obstacles it hits. Also, try to make a robot that stays within an area
indicated by a thick black border line on the floor.
Up to now all our programs consisted of just one task. But
NQC programs can have multiple tasks. It is also possible to put pieces of code
in so-called subroutines that you can use at different places in your program.
Using tasks and subroutines makes your programs easier to understand and more
compact. In this chapter we will look at the various possibilities.
An NQC program consists of at most 10 tasks. Each task has a
name. One task must have the name main,
and this task will be executed. The other tasks will only be executed when a
running tasks tells them to be executed using a start command. From this moment
on both tasks are running simultaneously (so the first task continues running).
A running task can also stop another running task by using the stop command.
Later this task can be restarted again, but it will start from the beginning;
not from the place where it was stopped.
Let me demonstrate the use of tasks. Put your touch sensor
again on your robot. We want to make a program in which the robot drives around
in squares, like before. But when it hits an obstacle it should react to it. It
is difficult to do this in one task, because the robot must do two things at
the same moment: drive around (that is, switching on and off motors at the
right moments) and watch for sensors. So it is better to use two tasks for
this, one task that drives the squares; the other that reacts to the sensors.
Here is the program.
task main()
{
SetSensor(SENSOR_1,SENSOR_TOUCH);
start
check_sensors;
start
move_square;
}
task move_square()
{
while
(true)
{
OnFwd(OUT_A+OUT_C); Wait(100);
OnRev(OUT_C); Wait(85);
}
}
task check_sensors()
{
while
(true)
{
if
(SENSOR_1 == 1)
{
stop
move_square;
OnRev(OUT_A+OUT_C); Wait(50);
OnFwd(OUT_A); Wait(85);
start
move_square;
}
}
}
The main task just sets the sensor type and then starts both
other tasks. After this, task main is finished. Task move_square moves the robot forever in squares. Task check_sensors checks whether the
touch sensor is pushed. If so it takes the following actions: First of all it
stops task move_square. This
is very important. check_sensors
now takes control over the motions of the robot. Next it moves the robot back a
bit and makes it turn. Then it can start move_square
again to let the robot again drive in squares.
It is very important to remember that tasks that you start
are running at the same moment. This can lead to unexpected results. Chapter X
explains these problems in detail and gives solutions for them.
Sometimes you need the same piece of code at multiple places
in your program. In this case you can put the piece of code in a subroutine and
give it a name. Now you can execute this piece of code by simply calling its
name from within a task. NQC (or actually the RCX) allows for at most 8
subroutines. Let us look at an example.
sub turn_around()
{
OnRev(OUT_C); Wait(340);
OnFwd(OUT_A+OUT_C);
}
task main()
{
OnFwd(OUT_A+OUT_C);
Wait(100);
turn_around();
Wait(200);
turn_around();
Wait(100);
turn_around();
Off(OUT_A+OUT_C);
}
In this program we have defined a subroutine that makes the
robot rotate around its center. The main task calls the subroutine three times.
Note that we call the subroutine by writing down its name with parentheses behind
it. So it looks the same as many of the commands we have seen. Only there are
no parameters, so there is nothing between the parentheses.
Some warnings are in place here. Subroutines are a bit
weird. For example, subroutines cannot be called from other subroutines.
Subroutines can be called from different tasks but this is not encouraged. It
very easily leads to problems because the same subroutine might actually be run
twice at the same moment by different tasks. This tends to give unwanted effects.
Also, when calling a subroutine from different tasks, due to a limitation in
the RCX firmware, you cannot use complicated expressions anymore. So, unless
you know precisely what you are doing, don’t
call a subroutine from different tasks!
As indicated above, subroutines cause certain problems. The
nice part is that they are stored only once in the RCX. This saves memory and,
because the RCX does not have so much free memory, this is useful. But when
subroutines are short, better use inline functions instead. These are not
stored separately but copied at each place they are used. This costs more
memory but problems like the ones with using complicated expressions, are no
longer present. Also there is no limit on the number of inline functions.
Defining and calling inline functions goes exactly the same
way as with subroutines. Only use the keyword void rather than sub.
(The word void is used because this
same word appears in other languages like C.) So the above example, using
inline functions, looks as follows:
void turn_around()
{
OnRev(OUT_C); Wait(340);
OnFwd(OUT_A+OUT_C);
}
task main()
{
OnFwd(OUT_A+OUT_C);
Wait(100);
turn_around();
Wait(200);
turn_around();
Wait(100);
turn_around();
Off(OUT_A+OUT_C);
}
Inline functions have another advantage over subroutines.
They can have arguments. Arguments can be used to pass a value for certain
variables to an inline function. For example, assume, in the above example, we
can make the time to turn an argument of the function, as in the following
examples:
void turn_around(int turntime)
{
OnRev(OUT_C); Wait(turntime);
OnFwd(OUT_A+OUT_C);
}
task main()
{
OnFwd(OUT_A+OUT_C);
Wait(100);
turn_around(200);
Wait(200);
turn_around(50);
Wait(100);
turn_around(300);
Off(OUT_A+OUT_C);
}
Note that in the parenthesis behind the name of the inline
function we specify the argument(s) of the function. In this case we indicate
that the argument is an integer (there are some other choices) and that its
name is turntime. When there are more arguments, you must separate them with
commas.
There is yet another way to give small pieces of code a
name. You can define macros in NQC (not to be confused with the macros in Bricx
Command Center). We have seen before that we can define constants, using
#define, by giving them a name. But actually we can define any piece of code.
Here is the same program again but now using a macro for turning around.
#define turn_around OnRev(OUT_C);Wait(340);OnFwd(OUT_A+OUT_C);
task main()
{
OnFwd(OUT_A+OUT_C);
Wait(100);
turn_around;
Wait(200);
turn_around;
Wait(100);
turn_around;
Off(OUT_A+OUT_C);
}
After the #define statement the word turn_around stands for
the text behind it. Now wherever you type turn_around, this is replaced by this
text. Note that the text should be on one line. (Actually there are ways of
putting a #define statement on multiple lines, but this is not recommended.)
Define statements are actually a lot more powerful. They can
also have arguments. For example, we can put the time to turn as an argument in
the statement. Here is an example in which we define four macro’s; one to move
forwards, one to move backwards, one to turn left and one to turn right. Each
has two arguments: the speed and the time.
#define turn_right(s,t) SetPower(OUT_A+OUT_C,s);OnFwd(OUT_A);OnRev(OUT_C);Wait(t);
#define turn_left(s,t) SetPower(OUT_A+OUT_C,s);OnRev(OUT_A);OnFwd(OUT_C);Wait(t);
#define forwards(s,t) SetPower(OUT_A+OUT_C,s);OnFwd(OUT_A+OUT_C);Wait(t);
#define backwards(s,t) SetPower(OUT_A+OUT_C,s);OnRev(OUT_A+OUT_C);Wait(t);
task main()
{
forwards(3,200);
turn_left(7,85);
forwards(7,100);
backwards(7,200);
forwards(7,100);
turn_right(7,85);
forwards(3,200);
Off(OUT_A+OUT_C);
}
It is very useful to define such macros. It makes your code
more compact and readable. Also, you can more easily change your code when you
e.g. change the connections to the motors.
In this chapter you saw the use of tasks, subroutines,
inline functions, and macros. They have different uses. Tasks normally run at
the same moment and take care of different things that have to be done at the
same moment. Subroutines are useful when larger pieces of code must be used at
different places in the same task. Inline functions are useful when pieces of
code must be used a many different places in different tasks, but they use more
memory. Finally macros are very useful for small pieces of code that must be
used a different places. They can also have parameters, making them even more
useful.
Now that you have worked through the chapters up to here,
you have all the knowledge you need to make your robot do complicated things.
The other chapters in this tutorial teach you about other things that are only
important in certain applications.
The RCX has a built-in speaker that can make sounds and even
play simple pieces of music. This is in particular useful when you want to make
the RCX tell you that something is happening. But it can also be funny to have
the robot make music while it runs around.
There are six built-in sounds in the RCX, numbered from 0 to
5. They sound as follows:
0 Key click
1 Beep beep
2 Decreasing frequency sweep
3 Increasing frequency sweep
4 ‘Buhhh’ Error sound
5 Fast increasing sweep
You can play them using
the commands PlaySound(). Here is a small program that plays all of them.
task main()
{
PlaySound(0);
Wait(100);
PlaySound(1);
Wait(100);
PlaySound(2);
Wait(100);
PlaySound(3);
Wait(100);
PlaySound(4);
Wait(100);
PlaySound(5);
Wait(100);
}
You might wonder why
there are these wait commands. The reason is that the command that plays the
sound does not wait for it to finish. It immediately executes the next command.
The RCX has a little buffer in which it can store some sounds but after a while
this buffer get full and sounds get lost. This is not so serious for sounds but
it is very important for music, as we will see below.
Note that the argument
to PlaySound() must be
a constant. You cannot put a variable here!
For more interesting
music, NQC has the command PlayTone(). It has two arguments. The first is the frequency, and the second the
duration (in ticks of 1/100h of a second, like in the wait command). Here is a
table of useful frequencies:
Sound 1 2
3 4 5 6 7
8
G# 52 104 208 415 831 1661 3322
G 49 98 196 392 784 1568 3136
F# 46 92 185 370 740 1480 2960
F 44 87 175 349 698 1397 2794
E 41 82 165 330 659 1319 2637
D# 39 78 156 311 622 1245 2489
D 37 73 147 294 587 1175 2349
C# 35 69 139 277 554 1109 2217
C 33 65 131 262 523 1047 2093 4186
B 31 62 123 247 494 988 1976 3951
A# 29 58 117 233 466 932 1865 3729
A 28 55 110 220 440 880 1760 3520
As we noted above for
sounds, also here the RCX does not wait for the note to finish. So if you use a
lot in a row better add (slightly longer) wait commands in between. Here is an
example:
task main()
{
PlayTone(262,40); Wait(50);
PlayTone(294,40); Wait(50);
PlayTone(330,40); Wait(50);
PlayTone(294,40); Wait(50);
PlayTone(262,160);
Wait(200);
}
You can create pieces of music very easily using the Brick
Piano that is part of the Bricx Command Center.
If you want to have the RCX play music while driving around,
better use a separate task for it. Here you have an example of a rather stupid
program where the RCX drives back and forth, constantly making music.
task music()
{
while
(true)
{
PlayTone(262,40); Wait(50);
PlayTone(294,40); Wait(50);
PlayTone(330,40); Wait(50);
PlayTone(294,40); Wait(50);
}
}
task main()
{
start
music;
while(true)
{
OnFwd(OUT_A+OUT_C); Wait(300);
OnRev(OUT_A+OUT_C); Wait(300);
}
}
In this chapter you learned how to let the RCX make sounds
and music. Also you saw how to use a separate task for music.
There are a number of additional motor commands that you can
use to control the motors more precisely. In this chapter we discuss them.
When you use the Off()
command, the motor stops immediately, using the brake. In NQC it is also
possible to stop the motors in a more gentle way, not using the brake. For this
you use the Float() command. Sometimes this is
better for your robot task. Here is an example. First the robot stops using the
brakes; next without using the brakes. Note the difference. (Actually the
difference is very small for this particular robot. But it makes a big
difference for some other robots.)
task main()
{
OnFwd(OUT_A+OUT_C);
Wait(200);
Off(OUT_A+OUT_C);
Wait(100);
OnFwd(OUT_A+OUT_C);
Wait(200);
Float(OUT_A+OUT_C);
}
The command OnFwd()
actually does two things: it switches the motor on and it sets the direction to
forwards. The command OnRev() also does two things: it
switches the motor on and sets the direction to reverse. NQC also has commands
to do these two things separately. If you only want to change one of the two
things, it is more efficient to use these separate commands; it uses less
memory in the RCX, it is faster, and it can result in smoother motions. The two
separate commands are SetDirection() that sets the direction (OUT_FWD, OUT_REV or OUT_TOGGLE which
flips the current direction) and SetOutput() that sets the mode (OUT_ON, OUT_OFF or OUT_FLOAT). Here is a
simple program that makes the robot drive forwards, backwards and forwards
again.
task main()
{
SetPower(OUT_A+OUT_C,7);
SetDirection(OUT_A+OUT_C,OUT_FWD);
SetOutput(OUT_A+OUT_C,OUT_ON);
Wait(200);
SetDirection(OUT_A+OUT_C,OUT_REV);
Wait(200);
SetDirection(OUT_A+OUT_C,OUT_TOGGLE);
Wait(200);
SetOutput(OUT_A+OUT_C,OUT_FLOAT);
}
Note that, at the start of every program, all motors are set
in forward direction and the speed is set to 7. So in the above example, the
first two commands are not necessary.
There are a number of other motor commands, which are
shortcuts for combinations of the commands above. Here is a complete list:
On(‘motors’) Switches
the motors on
Off(‘motors’) Switches
the motors off
Float(‘motors’) Switches
the motors of smoothly
Fwd(‘motors’) Switches
the motors forward (but does not make them drive)
Rev(‘motors’) Switches
the motors backwards (but does not make them drive)
Toggle(‘motors’) Toggles
the direction of the motors (forward to backwards and back)
OnFwd(‘motors’) Switches
the motors forward and turns them on
OnRev(‘motors’) Switches
the motors backwards and turns them on
OnFor(‘motors’,’ticks’) Switches the motors on
for ticks time
SetOutput(‘motors’,’mode’) Sets the output mode (OUT_ON, OUT_OFF or OUT_FLOAT)
SetDirection(‘motors’,’dir’) Sets the output direction (OUT_FWD, OUT_REV or OUT_TOGGLE)
SetPower(‘motors’,’power’) Sets the output power (0-9)
As you probably noticed, changing the speed of the motors
does not have much effect. The reason is that you are mainly changing the
torque, not the speed. You will only see an effect when the motor has a heavy
load. And even then, the difference between 2 and 7 is very small. If you want
to have better effects the trick is to turn the motors on and off in rapid succession.
Here is a simple program that does this. It has one task, called run_motor that
drives the motors. It constantly checks the variable speed to see what the
current speed is. Positive is forwards, negative backwards. It sets the motors
in the right direction and then waits for some time, depending on speed, before
switching the motors off again. The main task simply sets speeds and waits.
int speed, __speed;
task run_motor()
{
while
(true)
{
__speed = speed;
if
(__speed > 0) {OnFwd(OUT_A+OUT_C);}
if
(__speed < 0) {OnRev(OUT_A+OUT_C); __speed = -__speed;}
Wait(__speed);
Off(OUT_A+OUT_C);
}
}
task main()
{
speed = 0;
start
run_motor;
speed = 1;
Wait(200);
speed = -10; Wait(200);
speed = 5;
Wait(200);
speed = -2;
Wait(200);
stop
run_motor;
Off(OUT_A+OUT_C);
}
This program can be made much more powerful, allowing for
rotations, and also possibly incorporating a waiting time after the Off() command. Experiment yourself.
In this chapter you learned about the extra motor commands
that are available: Float() that stops the motor gently, SetDirection() that sets the direction (OUT_FWD, OUT_REV or OUT_TOGGLE which flips the current direction) and SetOutput()
that sets the mode (OUT_ON,
OUT_OFF or OUT_FLOAT). You saw
the complete list of motor commands available. You also learned a trick to
control the motor speed in a better way.
In Chapter V
we discussed the basic aspects of using sensors. But there is a lot more you
can do with sensors. In this chapter we will discuss the difference between
sensor mode and sensor type, we will see how to use the rotation sensor (a type
of sensor that is not provided with the RIS but can be bought separately and is
very useful), and we will see some tricks to use more than three sensors and to
make a proximity sensor.
The SetSensor() command that we saw before
does actually two things: it sets the type of the sensor, and it sets the mode
in which the sensor operates. By setting the mode and type of the a sensor
separately, you can control the behavior of the sensor more precisely, which is
useful for particular applications.
The type of the sensor is set with the command SetSensorType(). There are four different types: SENSOR_TYPE_TOUCH,
which is the touch sensor, SENSOR_TYPE_LIGHT, which is the light sensor, SENSOR_TYPE_TEMPERATURE,
which is the temperature sensor (this type of sensor is not part of the RIS but
can be bought separately), and SENSOR_TYPE_ROTATION, which is the rotation sensor (also not
part of the RIS but available separately). Setting the type sensor is in
particular important to indicate whether the sensor needs power (like e.g. for
the light of the light sensor). I know of no uses for setting a sensor to a
different type than it actually is.
The mode of the sensor is set with the command SetSensorMode(). There are eight different modes. The most
important one is SENSOR_MODE_RAW.
In this mode, the value you get when checking the sensor is a number between 0
and 1023. It is the raw value produced by the sensor. What it means depends on
the actual sensor. For example, for a touch sensor, when the sensor is not
pushed the value is close to 1023. When it is fully pushed, it is close to 50.
When it is pushed partially the value ranges between 50 and 1000. So if you set
a touch sensor to raw mode you can actually find out whether it is touched
partially. When the sensor is a light sensor, the value ranges from about 300
(very light) to 800 (very dark). This gives a much more precise value than
using the SetSensor() command.
The second sensor mode is SENSOR_MODE_BOOL. In this mode the value is 0 or 1. When the
raw value is above about 550 the value is 0, otherwise it is 1. SENSOR_MODE_BOOL is
the default mode for a touch sensor. The modes SENSOR_MODE_CELSIUS and SENSOR_MODE_FAHRENHEIT are useful with temperature sensors
only and give the temperature in the indicated way. SENSOR_MODE_PERCENT turns the raw value
into a value between 0 and 100. Every raw value of 400 or lower is mapped to
100 percent. If the raw value gets higher, the percentage slowly goes down to
0. SENSOR_MODE_PERCENT
is the default mode for a light sensor. SENSOR_MODE_ROTATION seems to be useful only for the
rotation sensor (see below).
There are two other interesting modes: SENSOR_MODE_EDGE and SENSOR_MODE_PULSE.
They count transitions, that is changes from a low to a high raw value or
opposite. For example, when you touch a touch sensor this causes a transition
from high to low raw value. When you release it you get a transition the other
direction. When you set the sensor mode to SENSOR_MODE_PULSE, only transitions from low to high are
counted. So each touch and release of the touch sensor counts for one. When you
set the sensor mode to SENSOR_MODE_EDGE,
both transitions are counted. So each touch and release of the touch sensor
counts for two. So you can use this to count how often a touch sensor is
pushed. Or you can use it in combination with a light sensor to count how often
a (strong) lamp is switched on and off. Of course, when you are counting
things, you should be able to set the counter back to 0. For this you use the
command ClearSensor(). It clears the counter for
the indicated sensor(s).
Let us look at an example. The following program uses a
touch sensor to steer the robot. Connect the touch sensor with a long wire to
input one. If touch the sensor quickly twice the robot moves forwards. It you
touch it once it stops moving.
task main()
{
SetSensorType(SENSOR_1,SENSOR_TYPE_TOUCH);
SetSensorMode(SENSOR_1,SENSOR_MODE_PULSE);
while(true)
{
ClearSensor(SENSOR_1);
until
(SENSOR_1 >0);
Wait(100);
if
(SENSOR_1 == 1) {Off(OUT_A+OUT_C);}
if
(SENSOR_1 == 2) {OnFwd(OUT_A+OUT_C);}
}
}
Note that we first set the type of the sensor and then the
mode. It seems that this is essential because changing the type also effects
the mode.
The rotation sensor is a very useful type of sensor that is
unfortunately not part of the standard RIS. It can though be bought separately
from Lego. The rotation sensor contains a hole through which you can put an
axle. The rotation sensor measures the amount the axle is rotated. One full
rotation of the axle is 16 steps (or –16 if you rotate it the other way).
Rotation sensors are very useful to make the robot make precisely controlled
movements. You can make an axle move the exact amount you want. If you need
finer control than 16 step, you can always use gears to connect it to an axle
that moves faster, and use that one for counting steps.
One standard application is to have two rotation sensors
connected to the two wheels of the robot that you control with the two motors.
For a straight movement you want both wheels to turn equally fast. Unfortunately,
the motors normally don’t run at exactly the same speed. Using the rotation
sensors you can see that one wheel turns faster. You can then temporarily stop
that motor (best using Float()) until both sensors give the
same value again. The following program does this. It simply lets the robot
drive in a straight line. To use it, change your robot by connecting the two
rotation sensors to the two wheels. Connect the sensors to input 1 and 3.
task main()
{
SetSensor(SENSOR_1,SENSOR_ROTATION);
ClearSensor(SENSOR_1);
SetSensor(SENSOR_3,SENSOR_ROTATION);
ClearSensor(SENSOR_3);
while
(true)
{
if
(SENSOR_1 < SENSOR_3)
{OnFwd(OUT_A); Float(OUT_C);}
else
if (SENSOR_1
> SENSOR_3)
{OnFwd(OUT_C); Float(OUT_A);}
else
{OnFwd(OUT_A+OUT_C);}
}
}
The program first indicates that both sensors are rotation
sensors, and resets the values to zero. Next it start an infinite loop. In the
loop we check whether the two sensor readings are equal. If they are the robot
simply moves forwards. If one is larger, the correct motor is stopped until
both readings are again equal.
Clearly this is only a very simple program. You can extend
this to make the robot drive exact distances, or to let it make very precise
turns.
The RCX has only three inputs so you can connect only three
sensors to it. When you want to make more complicated robots (and you bought
some extra sensors) this might not be enough for you. Fortunately, with some
tricks, you can connect two (or even more) sensors to one input.
The easiest is to connect two touch sensors to one input. If
one of them (or both) is touched, the value is 1, otherwise it is 0. You cannot
distinguish the two but sometimes this is not necessary. For example, when you
put one touch sensor at the front and one at the back of the robot, you know
which one is touched based on the direction the robot is driving in. But you
can also set the mode of the input to raw (see above). Now you can get a lot
more information. If you are lucky, the value when the sensor is pressed is not
the same for both sensors. If this is the case you can actually distinguish
between the two sensors. And when both are pressed you get a much lower value
(around 30) so you can also detect this.
You can also connect a touch sensor and a light sensor to
one input. Set the type to light (otherwise the light sensor won’t work). Set
the mode to raw. In this case, when the touch sensor is pushed you get a raw
value below 100. If it is not pushed you get the value of the light sensor
which is never below 100. The following program uses this idea. The robot must
be equipped with a light sensor pointing down, and a bumper at the front
connected to a touch sensor. Connect both of them to input 1. The robot will
drive around randomly within a light area. When the light sensor sees a dark
line (raw value > 750) it goes back a bit. When the touch sensor touches
something (raw value below 100) it does the same. Here is the program:
int ttt,tt2;
task moverandom()
{
while
(true)
{
ttt = Random(50)
+ 40;
tt2 = Random(1);
if
(tt2 > 0)
{ OnRev(OUT_A); OnFwd(OUT_C); Wait(ttt); }
else
{ OnRev(OUT_C); OnFwd(OUT_A);Wait(ttt); }
ttt = Random(150)
+ 50;
OnFwd(OUT_A+OUT_C);Wait(ttt);
}
}
task main()
{
start
moverandom;
SetSensorType(SENSOR_1,SENSOR_TYPE_LIGHT);
SetSensorMode(SENSOR_1,SENSOR_MODE_RAW);
while
(true)
{
if
((SENSOR_1 < 100) || (SENSOR_1 > 750))
{
stop
moverandom;
OnRev(OUT_A+OUT_C);Wait(30);
start moverandom;
}
}
}
I hope the program is clear. There are two tasks. Task moverandom makes the robot move
around in a random way. The main task first starts moverandom,
sets the sensor and then waits for something to happen. If the sensor reading
gets too low (touching) or too high (out of the white area) it stops the random
moves, backs up a little, and start the random moves again.
It is also possible to connect two light sensors to the same
input. The raw value is in some way related to the combined amount of light
received by the two sensors. But this is rather unclear and seems hard to use.
Connecting other sensors with rotation or temperature sensors seems not to be
useful.
Using touch sensors, your robot can react when it hits
something. But it would be a lot nicer when the robot could react just before
it hits something. It should know that it is near to some obstacle.
Unfortunately there are no sensors for this available. There is though a trick
we can use for this. The robot has an infra-red port with which it can
communicate with the computer, or with other robots. (We will see more about
the communication between robots in Chapter XI.)
It turns out that the light sensor that comes with the robot is very sensitive
to infra-red light. We can build a proximity sensor based on this. The idea is
that one tasks sends out infra-red messages. Another task measures fluctuations
in the light intensity that is reflected from objects. The higher the
fluctuation, the closer we are to an object.
To use this idea, place the light sensor above the infra-red
port on the robot, pointing forwards. In this way it only measures reflected
infra-red light. Connect it to input 2. We use raw mode for the light sensor to
see the fluctuations as good as possible. Here is a simple program that lets
the robot run forwards until it gets near to an object and then makes a 90
degree turn to the right.
int
lastlevel; // To store
the previous level
task send_signal()
{
while(true)
{SendMessage(0); Wait(10);}
}
task
check_signal()
{
while(true)
{
lastlevel = SENSOR_2;
if(SENSOR_2 > lastlevel + 200)
{OnRev(OUT_C); Wait(85); OnFwd(OUT_A+OUT_C);}
}
}
task main()
{
SetSensorType(SENSOR_2, SENSOR_TYPE_LIGHT);
SetSensorMode(SENSOR_2, SENSOR_MODE_RAW);
OnFwd(OUT_A+OUT_C);
start send_signal;
start check_signal;
}
The task send_signal
send out 10 IR signals every seconds, using the command SendMessage(0). The task check_signal repeatedly
saves the value of the light sensor. Then it checks whether it (slightly later)
has become at least 200 higher, indicating a large fluctuation. If so, it lets
the robot make a 90-degree turn to the right. The value of 200 is rather arbitrary.
If you make it smaller, the robot turns further away from obstacles. If you
make it larger, it gets closer to them. But this also depends on the type of
material and the amount of light available in the room. You should experiment
or use some more clever mechanism for learning the correct value.
A disadvantage of the technique is that it only works in one direction.
You probably still need touch sensors at the sides to avoid collisions there.
But the technique is very useful for robots that must drive around in mazes.
Another disadvantage is that you cannot communicate from the computer to the
robot because it will interfere with the infra-red commands send out by the
robot. (Also the remote control on your television might not work.)
In this chapter we have seen a number of additional issues
about sensors. We saw how to separately set the type and mode of a sensor and
how this could be used to get additions information. We learned how to use the
rotation sensor. And we saw how multiple sensors can be connected to one input
of the RCX. Finally, we saw a trick to use the infra red connection of the
robot in combination with a light sensor, to create a proximity sensor. All
these tricks are extremely useful when constructing more complicated robots.
Sensors always play a crucial role there.
As has been indicated before, tasks in NQC are executed
simultaneously, or in parallel as people usually say. This is extremely useful.
In enables you to watch sensors in one task while another task moves the robot
around, and yet another task plays some music. But parallel tasks can also
cause problems. One task can interfere with another.
Consider the following program. Here one task drives the
robot around in squares (like we did so often before) and the second task
checks for the touch sensor. When the sensor is touched, it moves a bit
backwards, and makes a 90-degree turn.
task main()
{
SetSensor(SENSOR_1,SENSOR_TOUCH);
start
check_sensors;
while
(true)
{
OnFwd(OUT_A+OUT_C); Wait(100);
OnRev(OUT_C); Wait(85);
}
}
task check_sensors()
{
while
(true)
{
if
(SENSOR_1 == 1)
{
OnRev(OUT_A+OUT_C);
Wait(50);
OnFwd(OUT_A);
Wait(85);
OnFwd(OUT_C);
}
}
}
This probably looks like a perfectly valid program. But if
you execute it you will most likely find some unexpected behavior. Try the
following: Make the robot touch something while it is turning. It will start
going back, but immediately moves forwards again, hitting the obstacle. The
reason for this is that the tasks may interfere. The following is happening.
The robot is turning right, that is, the first task is in its second sleep
statement. Now the robot hits the sensor. It start going backwards, but at that
very moment, the main task is ready with sleeping and moves the robot forwards
again; into the obstacle. The second task is sleeping at this moment so it
won’t notice the collision. This is clearly not the behavior we would like to
see. The problem is that, while the second task is sleeping we did not realize
that the first task was still running, and that its actions interfere with the
actions of the second task.
One way of solving this problem is to make sure that at any
moment only one task is driving the robot. This was the approach we took in
Chapter VI.
Let me repeat the program here.
task main()
{
SetSensor(SENSOR_1,SENSOR_TOUCH);
start
check_sensors;
start
move_square;
}
task move_square()
{
while
(true)
{
OnFwd(OUT_A+OUT_C); Wait(100);
OnRev(OUT_C); Wait(85);
}
}
task check_sensors()
{
while
(true)
{
if
(SENSOR_1 == 1)
{
stop
move_square;
OnRev(OUT_A+OUT_C); Wait(50);
OnFwd(OUT_A); Wait(85);
start move_square;
}
}
}
The crux is that the check_sensors
task only moves the robot after stopping the move_square
task. So this task cannot interfere with the moving away from the obstacle.
Once the backup procedure is finished, it starts move_square
again.
Even though this is a good solution for the above problem,
there is a problem. When we restart move_square,
it starts again at the beginning. This is fine for our small task, but often
this is not the required behavior. We would prefer to stop the task where it is
and continue it later from that point. Unfortunately this cannot be done
easily.
A standard technique to solve this problem is to use a
variable to indicate which task is in control of the motors. The other tasks are
not allowed to drive the motors until the first task indicates, using the
variable, that it is ready. Such a variable is often called a semaphore. Let sem be such a semaphore. We
assume that a value of 0 indicates that no task is steering the motors. Now,
whenever a task wants to do something with the motors it executes the following
commands:
until (sem == 0);
sem = 1;
// Do
something with the motors
sem = 0;
So we first wait till nobody needs the motors. Then we claim
the control by setting sem to
1. Now we can control the motors. When we are done we set sem back to 0. Here you find the
program above, implemented using a semaphore. When the touch sensor touches
something, the semaphore is set and the backup procedure is performed. During
this procedure the task move_square
must wait. At the moment the back-up is ready, the semaphore is set to 0 and move_square can continue.
int sem;
task main()
{
sem = 0;
start
move_square;
SetSensor(SENSOR_1,SENSOR_TOUCH);
while
(true)
{
if
(SENSOR_1 == 1)
{
until
(sem == 0); sem = 1;
OnRev(OUT_A+OUT_C); Wait(50);
OnFwd(OUT_A); Wait(85);
sem = 0;
}
}
}
task move_square()
{
while
(true)
{
until
(sem == 0); sem = 1;
OnFwd(OUT_A+OUT_C);
sem = 0;
Wait(100);
until
(sem == 0); sem = 1;
OnRev(OUT_C);
sem = 0;
Wait(85);
}
}
You could argue that it is not necessary in move_square to set the semaphore
to 1 and back to 0. Still this is useful. The reason is that the OnFwd() command is in fact two
commands (see Chapter VIII).
You don’t want this command sequence to be interrupted by the other task.
Semaphores are very useful and, when you are writing
complicated programs with parallel tasks, they are almost always required.
(There is still a slight chance they might fail. Try to figure out why.)
In this chapter we studied some of the problems that can
occur when you use different tasks. Always be very careful for side effects.
Much unexpected behavior is due to this. We saw two different ways of solving
such problems. The first solution stops and restarts tasks to make sure that
only one critical task is running at every moment. The second approach uses
semaphores to control the execution of tasks. This guarantees that at every
moment only the critical part of one task is executed.
If you own more than one RCX this chapter is for you. The
robots can communicate with each other through the infra-red port. Using this
you can have multiple robots collaborate (or fight with each other). Also you
can build one big robot using two RCXs, such that you can have six motors and
six sensors (or even more using the tricks in Chapter IX).
Communication between robots works, globally speaking, as
follows. A robot can use the command SendMessage()
to send a value (0-255) over the infra-red port. All other robots receive this
message and store it. The program in a robot can ask for the value of the last
message received using Message(). Based on this value the
program can make the robot perform certain actions.
Often, when you have two or
more robots, one is the leader. We call him the master. The other robots are slaves.
The master robot sends orders to the slaves and the slaves execute these.
Sometimes the slaves might send information back to the master, for example the
value of a sensor. So you need to write two programs, one for the master and
one for the slave(s). From now on we assume that we have just one slave. Let us
start with a very simple example. Here the slave can perform three different
orders: move forwards, move backwards, and stop. Its program consists of a
simple loop. In this loop it sets the value of the current message to 0 using the
ClearMessage() command. Next it waits until
the message becomes unequal to 0. Based on the value of the message it executes
one of the three orders. Here is the program.
task main() // SLAVE
{
while
(true)
{
ClearMessage();
until
(Message() != 0);
if
(Message() == 1) {OnFwd(OUT_A+OUT_C);}
if
(Message() == 2) {OnRev(OUT_A+OUT_C);}
if
(Message() == 3) {Off(OUT_A+OUT_C);}
}
}
The master has an even simpler program. It simply send the
messages corresponding to orders and then waits a bit. In the program below it
orders the slave to move forwards, then, after two seconds, backwards, and
then, again after two seconds, to stop.
task main() // MASTER
{
SendMessage(1);
Wait(200);
SendMessage(2);
Wait(200);
SendMessage(3);
}
After you have written these two program, you need to
download them to the robots. Each program must go to one of the robots. Make
sure you switch the other one off in the meantime (see also the cautions
below). Now switch on both robots and start the programs: first the one in the
slave and then the one in the master.
If you have multiple slaves, you have to download the slave
program to each of them in turn (not simultaneously; see below). Now all slaves
will perform exactly the same actions.
To let the robots communicate with each other we defined,
what is called, a protocol: We decided that a 1 means to move forwards, a 2 to
move backwards, and a 3 to stop. It is very important to carefully define such
protocols, in particular when you are dealing with lots of communications. For
example, when there are more slaves, you could define a protocol in which two
numbers are sent (with a small sleep in between): the first number is the
number of the slave, and the second is the actual order. The slave than first
check the number and only perform the action if it is his number. (This
requires that each slave has its own number, which can be achieved by letting
each slave have a slightly different program in which e.g. one constant is
different.)
As we saw above, when dealing with multiple robots, each
robot must have its own program. It would be much easier if we could download
just one program to all robots. But then the question is: who is the master?
The answer is easy: let the robots decide themselves. Let them elect a leader
which the others will follow. But how do we do this? The idea is rather simple.
We let each robot wait a random amount of time and then send a message. The one
that sends a message first is the leader. This scheme might fail if two robots
wait exactly the same amount of time but this is rather unlikely. (You can
build more complicated schemes that detect this and try a second election in
such a case.) Here is the program that does it:
task main()
{
ClearMessage();
Wait(200); // make sure all robots are on
Wait(Random(400));
// wait
between 0 and 4 seconds
if
(Message() > 0) // somebody else was first
{
start
slave;
}
else
{
SendMessage(1); // I am the master now
Wait(400); // make sure everybody else knows
start
master;
}
}
task master()
{
SendMessage(1);
Wait(200);
SendMessage(2);
Wait(200);
SendMessage(3);
}
task slave()
{
while
(true)
{
ClearMessage();
until
(Message() != 0);
if
(Message() == 1) {OnFwd(OUT_A+OUT_C);}
if
(Message() == 2) {OnRev(OUT_A+OUT_C);}
if
(Message() == 3) {Off(OUT_A+OUT_C);}
}
}
Download this program to all robots (one by one, not at the
same moment; see below). Start the robots at about the same moment and see what
happens. One of them should take command and the other(s) should follow the
orders. In rare occasions, none of them becomes the leader. As indicated above,
this requires more careful protocols to solve.
You have to be a bit careful
when dealing with multiple robots. There are two problems: If two robots (or a
robot and the computer) send information at the same time this might be lost.
The second problem is that, when the computer sends a program to multiple
robots at the same time, this causes problems.
Let us start with the second
problem. When you download a program to the robot, the robot tells the computer
whether it correctly receives (parts of) the program. The computer reacts on
that by sending new pieces or by resending parts. When two robots are on, both
will start telling the computer whether they correctly receive the program. The
computer does not understand this (it does not know that there are two
robots!). As a result, things go wrong and the program gets corrupted. The
robots won’t do the right things. Always
make sure that, while you are downloading programs, only one robot is on!
The other problem is that
only one robot can send a message at any moment. If two messages are being send
at roughly the same moment, they might get lost. Also, a robot cannot send and
receive messages at the same moment. This is no problem when only one robot
sends messages (there is only one master) but otherwise it might be a serious
problem. For example, you can imagine writing a program in which a slave sends
a message when it bumps into something, such that the master can take action.
But if the master sends an order at the same moment, the message will get lost.
To solve this, it is important to define your communication protocol such that,
in case a communication fails, this is corrected. For example, when the master
sends a command, it should get an answer from the slave. If is does not get an
answer soon enough, it resends the command. This would result in a piece of
code that looks like this:
do
{
SendMessage(1);
ClearMessage();
Wait(10);
}
while (Message() != 255);
Here 255 is used for the acknowledgement.
Sometimes, when you are dealing with multiple robots, you
might want that only a robot that is very close by receives the signal. This
can be achieved by adding the command SetTxPower(TX_POWER_LO) to the program of the master. In
this case the IR signal send is very low and only a robot close by and facing
the master will “hear” it. This is in particular useful when building one
bigger robot out of two RCXs. Use SetTxPower(TX_POWER_HI) to set the robot again in long
range transmission mode.
In this chapter we studied
some of the basic aspects of communication between robots. Communication uses
the commands to send, clear, and check messages. We saw that is is important to
define a protocol for how the communication works. Such protocols play a
crucial role in any form of communication between computers. We also saw that
there are a number of restrictions in the communication between robots which
makes it even more important to define good protocols.
NQC has a number of additional commands. In this chapter we
will discuss three types: the use of timers, commands to control the display,
and the use of the datalog feature of the RCX.
The RCX has four built-in timers. These timers tick in
increments of 1/10 of a second. The timers are numbered from 0 to 3. You can
reset the value of a timer with the command ClearTimer() and get the current value of the timer
with Timer(). Here is an example of
the use of a timer. The following program lets the robot drive sort of random
for 20 seconds.
task main()
{
ClearTimer(0);
do
{
OnFwd(OUT_A+OUT_C);
Wait(Random(100));
OnRev(OUT_C);
Wait(Random(100));
}
while
(Timer(0)<200);
Off(OUT_A+OUT_C);
}
You might want to compare this program with the one given in
Chapter IV
that did exactly the same task. The one with timers is definitely simpler.
Timers are very useful as a replacement for a Wait() command. You can sleep for a
particular amount of time by resetting a timer and then waiting till it reaches
a particular value. But you can also react on other events (e.g. from sensors)
while waiting. The following simple program is an example of this. It lets the
robot drive until either 10 seconds are past, or the touch sensor touches
something.
task main()
{
SetSensor(SENSOR_1,SENSOR_TOUCH);
ClearTimer(3);
OnFwd(OUT_A+OUT_C);
until
((SENSOR_1 == 1) || (Timer(3)
>100));
Off(OUT_A+OUT_C);
}
Don’t forget that timers work in ticks of 1/10 of a second,
while e.g. the wait command uses ticks of 1/100 of a second.
It is possible to control the
display of the RCX in two different ways. First of all, you can indicate what
to display: the system clock, one of the sensors, or one the motors. This is
equivalent to using the black view button on the RCX. To set the display type,
use the command SelectDisplay(). The following program shows
all seven possibilities, one after the other.
task main()
{
SelectDisplay(DISPLAY_SENSOR_1); Wait(100); // Input 1
SelectDisplay(DISPLAY_SENSOR_2); Wait(100); // Input 2
SelectDisplay(DISPLAY_SENSOR_3); Wait(100); // Input 3
SelectDisplay(DISPLAY_OUT_A); Wait(100); // Output A
SelectDisplay(DISPLAY_OUT_B); Wait(100); // Output B
SelectDisplay(DISPLAY_OUT_C); Wait(100); // Output C
SelectDisplay(DISPLAY_WATCH); Wait(100); // System clock
}
Note that you should not use SelectDisplay(SENSOR_1).
The second way you can control
the display is by controlling the value of the system clock. You can use this
to display e.g. diagnostic information. For this use the command SetWatch(). Here is a tiny program that
uses this:
task main()
{
SetWatch(1,1);
Wait(100);
SetWatch(2,4);
Wait(100);
SetWatch(3,9);
Wait(100);
SetWatch(4,16);
Wait(100);
SetWatch(5,25);
Wait(100);
}
Note that the arguments to SetWatch() must be constants.
The RCX can store values of
variables, sensor readings, and timers, in a piece of memory called the
datalog. The values in the datalog cannot be used inside the RCX, but they can
be read by your computer. This is useful to e.g. check what is going on in your
robot. Bricx Command Center has a special window in which you can view the
current contents of the datalog.
Using the datalog consists of
three steps: First, the NQC program must define the size of the datalog, using
the command CreateDatalog(). This also clears the current
contents of the datalog. Next, values can be written in the datalog using the command AddToDatalog(). The values will be written one
after the other. (If you look at the display of the RCX you will see that one
after the other, four parts of a disk appear. When the disk is complete, the
datalog is full.) If the end of the datalog is reached, nothing happens. New
values are no longer stored. The third step is to upload the datalog to the PC.
For this, choose in Bricx Command Center the command Datalog in the Tools
menu. Next press the button labelled Upload
Datalog, and all the values appear. You can watch them or save them to a
file to do something else with them. People have used this feature to
e.g. make a scanner with the RCX.
Here is a simple example of a
robot with a light sensor. The robot drives for 10 seconds, and five times a
second the value of the light sensor is written into the datalog.
task main()
{
SetSensor(SENSOR_2,SENSOR_LIGHT);
OnFwd(OUT_A+OUT_C);
CreateDatalog(50);
repeat
(50)
{
AddToDatalog(SENSOR_2);
Wait(20);
}
Off(OUT_A+OUT_C);
}
XIII.
If you have worked your way through this tutorial you can
now consider yourself an expert in NQC. If you have not done this up to now, it
is time to start experimenting yourself. With creativity in design and
programming you can make Lego robots do the most wonderful things.
This tutorial did not cover all aspects of the Bricx Command
Center. You are recommended to read the documentation at some stage. Also NQC
is still in development. Future version might incorporate additional
functionality. Many programming concepts were not treated in this tutorial. In
particular, we did not consider learning behavior of robots or other aspects of
artificial intelligence.
It is also possible to steer a Lego robot directly from a
PC. This requires you to write a program in a language like Visual Basic, Java
or Delphi. It is also possible to let such a program work together with an NQC
program running in the RCX itself. Such a combination is very powerful. If you
are interested in this way of programming your robot, best start with
downloading the spirit technical reference from the Lego MindStorms web site.
The web is a perfect source for additional information. Some
other important starting points are on LUGNET, the LEGOรข Users Group Network
(unofficial):
A lot of information can also be found in the
newsgroup lugnet.robotics and lugnet.robotics.rcx.nqc at news.lugnet.com
