Ask Your Question
0

turtlebot orientation and moving towards target

asked 2014-02-03 04:48:27 -0500

nadal11 gravatar image

updated 2014-05-14 04:08:06 -0500

Hi everybody,

I am new to ROS and currently I am using turtlebot for my experiment. My aim is to rotate the turtlebot to face my target coordinate and move towards it. However, the turtlebot is not behaving as I expected. The turtlebot is keep rotating or keep moving forward. Here is my source code and kindly give me some advices.

for example destination is (1,-0.5)

curr_loc = msg->pose.pose;

double x=msg->pose.pose.position.x;
double y=msg->pose.pose.position.y;
double z=msg->pose.pose.orientation.z;
double w=msg->pose.pose.orientation.w;

curr_head=atan2(2*(x*y+w*z), w*w + x*x - y*y - z*z);

DestX=1; DestY=-0.5;

left_X=DestX-x;
left_Y=DestY-y;

left_dist=sqrt((left_X * left_X) + (left_Y*left_Y));

if(left_dist <= 0.05)
{
   reached=true;
}

if(!reached)
{
   heading=atan2(left_locY,left_locX);
}

n_head=curr_head-heading;

if (n_head<0.2 && n_head>-0.2)
   facing=true;
else
   facing=false;

if (!facing)
   cmdvel.linear.x=0.0;

if (n_head>-3.14)
   cmdvel.angular.z=-0.25;
else if (n_head>-1.6)
   cmdvel.angular.z=-0.10;
else if (n_head<3.14)
    cmdvel.angular.z=0.25;
else if (n_head>1.6)
    cmdvel.angular.z=0.10;
else if (n_head>3.14)
     cmdvel.angular.z=-0.10;
else if (n_head<-3.14)
     cmdvel.angular.z=0.10;

if (facing && !reached)
   {
     cmdvel.angular.z=0.0;
          if (left_dist>0.5)
           {
            cmdvel.linear.x=-0.2;
                usleep(1000000);
       }
          else if (left_dist>0.2)
          {
           cmdvel.linear.x=-0.15;
           usleep(1000000);
      } 
         else if (left_dist>0.08)
         {
          cmdvel.linear.x=-0.05;
          usleep(1000000);
     }
  }
 else if (reached)
 {
    cmdvel.linear.x=0.0;
    n_head=0.0;
    ROS_INFO("Turtlebot has reached the destination");
}

Hope anybody can help me out. Thank you!

edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted
1

answered 2014-05-14 11:07:49 -0500

updated 2014-05-14 11:09:52 -0500

Hey, I believe there is lil problem with logic, atan(y,x) gives you angle mapped from -3.14 to 3.14. So at any point of time your curr_head and heading value will be between -pi to pi. and n_head = curr_head-headding (lets take a cases) 1) both values from 0 to pi/2 -> diff will be > -3.14 so it will satisfy first condition and wont go for next (coz of else if) (this will behave same way for any angles with abs(diff) < pi ) It will move round and round in most cases .. and you should increase your threshold because the subscribed topic callback depends on the frequency you give. so if its not high enough then it will skip some poses. you can rearrange if conditions for your first problem, ( first try to write extreme conditions like >pi or <-pi)

NOTE :Actually, n_head value ls between -2pi to 2pi, you should consider those cases too. Its better if you map the data first to 0 -> 2*pi , and use abs(diff) to get diff in value and abs(diff)/diff to get direction. it would be lot easier with lot less equations.

Hope this helped Sudeep

edit flag offensive delete link more

Comments

The angles package has a few functions that are very useful for this sort of angle-based math: normalize_angle and shortest_angular_distance. Much cleaner and easier than writing your own angle normalization.

ahendrix gravatar imageahendrix ( 2014-05-14 14:52:32 -0500 )edit

Thank you for both suggestion!

nadal11 gravatar imagenadal11 ( 2014-05-15 01:08:11 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Question Tools

1 follower

Stats

Asked: 2014-02-03 04:48:27 -0500

Seen: 1,508 times

Last updated: May 14 '14