In a two-dimensional environment:


  • a starting point (x1,y1)
  • an initial velocity (speed and 2D heading/direction)
  • a constant rate of turning (yaw, since there is no roll or pitch in only two dimensions)
  • a constant acceleration of the object from itself (not like gravity where the direction of the acceleration is relatively constant - "down" - but like a rocket that is propelling itself)
  • a time change t1 to t2

What is the formula for the final point (x2,y2)? What kind of physics problem do I have on my hands? Is this just a variation on kinematics with variable acceleration (varying direction of acceleration instead of amount), or is it qualitatively different?

  • $\begingroup$ Intuitively I think that this should result in a path which might be visualized as a quarter of an ellipse and I was trying to work off of physicsclassroom.com/class/1DKin/Lesson-6/Kinematic-Equations and numericana.com/answer/ellipse.htm but I'm not sure this is the correct direction/method. $\endgroup$ – Stephen T. Robbins Jan 8 '16 at 0:21
  • $\begingroup$ Is this vehicle like a car with wheels, so it cannot slide sideways, or is it like a rocket, that can go sideways? $\endgroup$ – Mike Dunlavey Jan 8 '16 at 1:52
  • $\begingroup$ Like a car, no sliding. (turning without accelerating) If it is headed north/up from (0,0) and turns 90 degrees to the right at a speed of pi/10 per second, then after 10 seconds it will be at (2,2). $\endgroup$ – Stephen T. Robbins Jan 8 '16 at 2:23
  • $\begingroup$ OK, this is like a car. Now the constant rate of turning. Acceleration comes into play. Do you mean the steering wheel is held at a constant angle, so the front wheels are at a constant angle, so there is a fixed center about which the car is turning, in which case the curve is a circular arc? Or do you mean that as the speed increases the steering wheel is turned back toward 0 so as to keep the yaw rate constant? Just trying to clarify the question. $\endgroup$ – Mike Dunlavey Jan 8 '16 at 13:33
  • $\begingroup$ That is a great question that I had not thought to clarify. In this case, I definitely want the yaw rate to be constant. $\endgroup$ – Stephen T. Robbins Jan 8 '16 at 18:58

OK, If acceleration is constant over the interval, that means velocity $v$ changes at a constant rate over the interval. If steering wheel position (and thus path curvature) is $\phi$, and yaw rate is proportional to $v\phi$, that means $\phi$ must vary as $1/v$. To put it another way, turning radius $r$ must vary proportionally to $v$.

This describes some kind of spiral, which might be well-known, but I don't know what it is (where both $v$ and $r$ vary linearly with time and $v/r$ is constant).

Without knowing that (or even if you do know it) I would use a suitable ODE solver, such as Runge-Kutta, or even Euler if I didn't mind the accuracy vs. time tradeoff.

  • $\begingroup$ Ah, yeah, you're right. It is going to be a spiral if it keeps going. I confused myself by only considering one short section. $\endgroup$ – Stephen T. Robbins Jan 8 '16 at 19:59
  • $\begingroup$ @StephenT.Robbins: I'm tempted to say Achimedean spiral where radius is linear in $\theta$, but I'm not sure. $\endgroup$ – Mike Dunlavey Jan 8 '16 at 20:10
  • $\begingroup$ I found the Achimedean spiral pretty quickly and it does look at least very close though I'm still trying to nail down the correct equations to be using. Still, it helps a lot to at least be looking at the correct shape set! I am hoping it turns out to be the Archimedes spiral because that would make things simpler/easier. $\endgroup$ – Stephen T. Robbins Jan 8 '16 at 21:31
  • $\begingroup$ @StephenT.Robbins: I just realized, it $d\theta/dt$ is constant, that means $dr/dt$ is constant, so $dv/dt$ is constant, and that's your constant acceleration. So yes, the spiral is Archimedean. $\endgroup$ – Mike Dunlavey Jan 9 '16 at 15:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.