Autonomus outdoor off-road robot

rockclodbuster · August 5, 2009, 4:16am

Hi everybody
I have start to built and programme a autonomus outdoor off-road robot based on a clod buster axle and chassie.

I use:
2 Integy 12v lathe motor for power
1 Hitec HS5645MG servo for front steering
1 Hitec HS645MG servo for rear steering
1 Novak Super Duty XR speed control
1 arduino duemilanove as brain
1 Pololu micro serial servo controler
3 Sharp GP2D12 IR range sensor
1 Duratrax 3800mAh 7.2V batterie for motor

I will add:
1 3 axe accelerometer for tilt sensor
1 Phiget voltage sensor
1 comunication device
2 magnetic encoder
2 ultrasonic range sensor
1 GPS sensor
and a remote controle and display

I have start the initial programming of my robot. I have buy three Sharp GP2D12 IR range sensor and Summerour Sharp IR Turret mounted in the front of the chassie for indoor speed and brake test.

For now, I only use the center sensor for simplier programming. they run a about half speed (1.5m/sec) and they stop at 10cm of the wall then backward arount 20cm and go forward again for 10cm, almost evry time. I have try it outdoor but it dont see the concret as well as painted wall. so after a hard hit, I planning ordering utrasonic sensor for longuer range and less sunray perturbation.

Next step is making it follow wall, creating a smooter speed control and creating my magnetic encoder.

moor coming soon

rockclodbuster · August 13, 2009, 1:10am

few bad quality picture of my robot
with body on:
http://i615.photobucket.com/albums/tt240/rockclodbuster/autoclod1.jpg
http://i615.photobucket.com/albums/tt240/rockclodbuster/autoclod2.jpg
http://i615.photobucket.com/albums/tt240/rockclodbuster/autoclod3.jpg
without body:
http://i615.photobucket.com/albums/tt240/rockclodbuster/autoclod4.jpg

rockclodbuster · August 13, 2009, 1:31am

front Sharp IR sensor, I wait two ultrasonic range finder to put in place.
http://i615.photobucket.com/albums/tt240/rockclodbuster/autoclod5.jpg

the front axle (exacly the same in the rear end) from front, you can see the shock, the batterie and the front steering servo:
http://i615.photobucket.com/albums/tt240/rockclodbuster/autoclod6.jpg

from bottom, you can see the motor and the suspention linkage:
http://i615.photobucket.com/albums/tt240/rockclodbuster/autoclod7.jpg

and finaly, the speed control:
http://i615.photobucket.com/albums/tt240/rockclodbuster/autoclod8.jpg

But now I have a probleme, My pololu micro servo controler is freaking. Normaly, the yellow led is on and the green led is flashing really quikly (almost steady on) a few second after power up. Now, the yellow led trun on, then it turn off and the green turn on but the controler continu working correctly. But here the probleme, many time the green turn off, the yellow turn on, the red flash one time per seconde and the controler stop working (motor stop, servo go to any position). some time it don’t do this for many minute, other time it swich for green to yellow/red so fast the led doesn’t actualy turn on completely before they turn off. What can cause this??? oh yes, one more thing, the Tx led on the arduino don’t flash.

here my code[code]int sensor0 = 0;
int sensor1 = 1;
int sensor2 = 2;
int x = 0;
int y = 0;
int z = 0;

void setup()
{
Serial.begin(4800);
delay(10);
Serial.print(128,BYTE); //start
Serial.print(1,BYTE); //device ID
Serial.print(0,BYTE); //command; set parameters
Serial.print(0,BYTE); //servo num: servo 0
Serial.print(69,BYTE); // servo0 on, fwd, range 90
delay(10);
Serial.print(128,BYTE); //start
Serial.print(1,BYTE); //device ID
Serial.print(0,BYTE); //command; set parameters
Serial.print(1,BYTE); //servo num: servo 1
Serial.print(79,BYTE); // servo1 on, fwd, range 90
delay(10);
Serial.print(128,BYTE); //start
Serial.print(1,BYTE); //device ID
Serial.print(0,BYTE); //command; set parameters
Serial.print(1,BYTE); //servo num: servo 2
Serial.print(111,BYTE); // servo2 on, rev, range 90
delay(10);
}

void loop()
{
x = ((600-analogRead(sensor1))/5.3);
y = ((((analogRead(sensor0)*0)+(analogRead(sensor1)*1000)+(analogRead(sensor2)*2000))/(analogRead(sensor0)+analogRead(sensor1)+analogRead(sensor2)))/4);
Serial.print(128,BYTE); //start
Serial.print(1,BYTE); //device ID
Serial.print(2,BYTE); //command; set position
Serial.print(0,BYTE); //servo num: servo 0
Serial.print(x,BYTE); // servo0 position
delay(10);
Serial.print(128,BYTE); //start
Serial.print(1,BYTE); //device ID
Serial.print(2,BYTE); //command; set position
Serial.print(1,BYTE); //servo num: servo 1
Serial.print(y,BYTE); // servo1 position
delay(10);
}

[/code]

rockclodbuster · August 13, 2009, 6:07pm

thank for the help, I think is the Y position, it will easily pass 127. also, thank for the servo initialing, the copy/past metode save some time but you can do error.

rockclodbuster · August 14, 2009, 2:56am

I have add a servo for turning the sensor turrel. Now the sensor will “look” in the direction of the move like the human eyes. much easier to program, I think.

http://i615.photobucket.com/albums/tt240/rockclodbuster/autoclod9.jpg

http://i615.photobucket.com/albums/tt240/rockclodbuster/autoclod10.jpg

now, I try to make it work:)

rockclodbuster · August 29, 2009, 1:46am

some news for you, my truck run by itself but lack of sensor range for most of obstacles. I will order some new ultrasonic range sensor and re-test.

I have program the bot for going forward most of time. It turn the steering for navigate around the house. I have try to make it remembre the last 5 seconde of move, so it can use it to get out of trouble (in front of a wall, in a corner, too incline longitudinaly (±45 degre) or lateraly (±30 degre)), but it appeir they calculate too fast and they don’t reverse.

I wish it can run a full speed (fast walking speed for a human) on a race track that use black drain pipe as “wall”. If it can to that, I will put the electronic in my race truck for faster run (40Mph!! yes 40Mph). I imagine the face of the other driver went they will see me putting the remote on the floor in the middle of the race.

I don’t know if the arduino can compute fast enought for keeping the truck in the track. If it don’t work, I will put a arm with a camera on the truck and make a empty pepsi can collector.

rockclodbuster · September 25, 2009, 2:53am

I have made several tests during the last three week. I have learn a lot:

]IR sensor don’t like direct sunray
I will use two ultrasound sensor in place./:m]
]80 cm is short distance for a fast robot
I will use two ultrasound sensor that have 6 meter of range./:m]
]making robot running at full speed all time is not effective
I will use a PID control for the speed./:m]
]three sensor are not enought
I will use two ultrasound sensor with scanning/mapping program./:m]
]40 minute off run time are not enought
two batterie of 3800mAh (one hour and half of run time) will be better./:m]
]a robot without mission is useless
I don’t have found one for mine:confused:/:m]

I have to create a mapping program using two ultrasound rangefinder mounted on two servomotor. the sensor will sweep left to right in 16 step and the distance will be store for each step for each sensor, so i have to store 64 time 4 byte per scan. I need to know what type of memorie a should use for storing all the data of scannig, the map (basic map, “I can or I can’t go here” type), the encoder distance form start and the accelerometer x,y,z axis. In short, a data base that my robot could use to navigate, it will have a hour and half of run time so this will be a lot of data. Can I use a SD card as memorie?

I also need to know what is the wide of the beam of the Devantech SRF02 Low cost Ultrasonic Range Finder.

rockclodbuster · October 8, 2009, 5:08am

Thanks Gardangel, I will think about using my robot as surveillance drone.

Now, I want too use a PICAXE Datalogger Memory Expander PCB Kit, for logging the data of all sensor, but I don’t know witch chip it use. Witch one is better, micro SD or PICAXE Datalogger?

jbrunet · August 10, 2009, 5:43pm

Sounds very interesting, can’t wait to see some pictures!

Keep it up.

jbrunet · August 13, 2009, 3:00pm

In your initialization routine I see something worth verifying, you are initializing servo 1 twice:

void setup()
{
  Serial.begin(4800);
  delay(10);
  Serial.print(128,BYTE);  //start
  Serial.print(1,BYTE);    //device ID
  Serial.print(0,BYTE);    //command; set parameters
  Serial.print(0,BYTE);    //servo num: servo 0
  Serial.print(69,BYTE);   // servo0 on, fwd, range 90
  delay(10);
  Serial.print(128,BYTE);  //start
  Serial.print(1,BYTE);    //device ID
  Serial.print(0,BYTE);    //command; set parameters
  Serial.print(1,BYTE);    //servo num: servo 1
  Serial.print(79,BYTE);   // servo1 on, fwd, range 90
  delay(10);
  Serial.print(128,BYTE);  //start
  Serial.print(1,BYTE);    //device ID
  Serial.print(0,BYTE);    //command; set parameters
  **Serial.print(1,BYTE);    //servo num: servo 2**
  Serial.print(111,BYTE);  // servo2 on, rev, range 90
  delay(10);
}

Also, make sure that the x and y values stay within the 0-127 range. If you are having problems in Pololu Mode, try Mini SSC II Mode.

richard-28 · October 8, 2009, 10:25pm

The best way to approach this Rockclodbuster would be to document your data storage requirements. How many bytes of data over what time span. Once you know the operational timeframe for your robot’s “missions”, you can easily determine how much storage will be required. Often times a simple EEPROM is sufficient, sometimes larger media is required.

So if you need to record temperature (1 Byte) every second and you will be operating for 1 hour, you will need 1 x 60 x 60 = 3600 bytes of storage. As you can see, increasing the sample frequency can dramatically influence the data storage requirements. Some data you don’t need to sample as often. Play around with some numbers and tally them up

gardangel · September 3, 2009, 8:48pm

Hey, your project is really cool. After it’s finished, maybe you can do the same but with an airplane ! )

Keep going and keep us informed.

Jeff

gardangel · October 2, 2009, 10:56pm

Hey, it’s good to hear about your project. Your errors help others not to do the same. About the IR, it’s true sunlight is messing with it. I found that an optical mouse with sunlight on the surface below will not work…

I thought about your project and mine actually… About the energy problem, maybe it could be a good solution if your robot is able to switch battery. One remains charging itself since the other is in use. Also, the solar energy systems could be a solution to improve it.

About the mission, you already have one : a surveillance drone. Unfortunatly, if you let the drone alone, it could be stole… But, inside your house, it could patrols for intruders. Also, it can be a tele op if you want to be in your house all the time, across the internet.

The idea about using this kind of technology is good because it’s near the perfect structure to run outdoor. If you can put some gears like tanks on the wheels, it will rocks !

Keep going and keep us informed.

Jeff

gardangel · October 9, 2009, 3:20am

hmmm logging data from all sensors ? It could be a large amount of datas. What about using wireless ? Send the data to a computer. A large amount of datas could be hard to process in a low processor. In my project, all the contextual parts are separated and communicate using wireless. The ‘mobile’ parts sends information to a notebook using wireless. With this solution, I can separate the computer from the mobile part which is safer to protect the notebook from falling. Also, you have less electronics parts to plug. And less energy will be required since the computer will be plug directly in the house. But there could be some delays between the two parts… You should process the navigation system with less delay.

Good luck,

Jeff