Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Mainstream treatment for relativistic Doppler effect under accelerated motion?

  1. Jan 20, 2007 #1
    I would need a pointer to a mainstream treatment of the general relativistic Doppler effect when source and receiver are accelerated wrt each other. Book/paper/wiki . No crank papers , please (i.e. no Apeiron and such). Thank you.
  2. jcsd
  3. Jan 20, 2007 #2
    accelerated Doppler and others

    I was interested in the problem because in the case of the Doppler Effect with accelerating observers we should take into account the nonlocality in the period measurement by accelerating observers. The results of my investigations are presented in "Radar echo, Doppler Effect and Radar detection in the uniformly accelerated reference frame"
    Authors: Rothenstein, Bernhard; Popescu, Stefan
    The uniformly accelerated reference frame described by Hamilton, Desloge and Philpott involves the observers who perform the hyperbolic motion with constant proper acceleration gi. They start to move from different distances measured from the origin O of the inertial reference frame K(XOY), along its OX axis with zero initial velocity. Equipped with clocks and light sources they are engaged with each other in Radar echo, Doppler Effect and Radar detection experiments. They are also engaged in the same experiments with an inertial observer at rest in K(XOY) and located at its origin O. We derive formulas that account for the experiments mentioned above. We study also the landing conditions of the accelerating observers on a uniformly moving platform.
    Comment: 15 pages, 8 figures, includes new results on radar detected times and distances
    If you consider arXiv among "such" then delete my thread. If not please have a critical look at my paper you can download from arXiv.
    sine ira et studio

  4. Jan 21, 2007 #3

    Chris Hillman

    User Avatar
    Science Advisor

    Mainstream treatments of freq. shifts for accelerated receivers/emitters

    Hi, nakurusil,

    I am not sure I know what you mean by "general relativistic Doppler effect"; one major point I have tried to emphasize in several threads here is that this stuff really has nothing to do with general relativity or even curved versus flat spacetime, but rather with the geometry of congruences (especially null congruences), whether in flat or curved spacetimes.

    So let me restate the question as: "what are some mainstream treatments of frequency shift phenomena, involving a source and an receiver, when either the source or the receiver (or both) are accelerating"?

    Some obvious places to begin are the first edition of Taylor and Wheeler, Spacetime Physics, or another book which treats the k-calculus (IIRC, there is one by Rindler which does so), then Section 2.8 and Chapter 6 of Misner, Thorne & Wheeler, Gravitation (MTW) for two scenarios featuring accelerating observers and or sources.

    Then you can see "Frame fields in general relativity", "Rindler coordinates", "Born coordinates", "Ehrenfest paradox", "Bell's spaceship paradox" in the versions listed at http://en.wikipedia.org/wiki/User:Hillman/Archive; [Broken] these have citations to published review papers where you can find many more references. Books like Nakayama, Geometry, Topology, and Physics, might be useful in following the first article; there is also some material on frame fields in MTW, and you can find these techniques treated in standard monographs such as Hawking & Ellis, The Large Scale Structure of Space-Time.

    Note well! Wikipedia is inherently unstable, and that articles in this specific area have been dogged in the past year by a single cranky dissident who managed to chase out of WP least one contributor with a Ph.D. in physics, plus myself (Ph.D. in mathematics), who incorrectly maintains in the face of all evidence that the mainstream view is his view. For this reason, current versions of the articles I mentioned may be better than the last ones I contributed to, or they may be much worse, so I recommend that you start with the ones I cited and compare carefully with subsequent editions. To repeat something I find myself saying with distressing frequency at PF: I will not "discuss" physics with specific cranks or their supporters; it should not be neccessary to for me to explain why not and I will not do so. Ultimately, you are on your own in terms of evaluating specific versions of specific WP articles; it will not always be easy to tell simply from obvious clues whether or not the article faithfully and accurately reflects the current scientific mainstream. Ultimately, only someone who has read the literature with adequate insight and understanding may be able to tell.

    And don't overlook the obvious: MTW itself has a huge bibliography which you can use to find some older but still important, useful, and relevant papers.

    Note well! Everyone should be aware that published papers vary widely in quality; as you yourself obviously already appreciate (good!), some "journals" seem to function as trashcans which collect papers rejected by more rigorous journals. In addition, while the arXiv is an invaluable resource, it does not yet function as a refereed journal. The "endorsement" system is only analogous to "moderation" in a newsgroup; it cannot and does not prevent cranky eprints from being posted to the arXiv. Thus, the quality of eprints posted there varies even more widely than does the quality of papers in the published literature. So you should be cautious about anything you read outside a highly reputable and widely used textbook such as MTW until you know more.

    Unfortunately, I must add a specific caveat. I and others have noticed that the arXiv is particularly uneven in the case of papers on alleged "foundational issues" centering around relativistic physics, and I emphatically intend to include the treatment of accelerated observers as a known "problem area" in the arXiv where eprints have a better than even chance of being partially or completely incorrect. It is crucial to understand that most physicists (at least those who often deal with relativistic physics), upon being asked to pick out the bad eprints, would have little trouble doing so; there is fact wide concensus on right and wrong ways to treat these problems, but a small group of noisy dissidents came sometimes create an incorrect impression which might fool casual observers unfamiliar with standard mathematical techniques. Let there be no mistake: the appropriate mathematical techniques are very well known and widely used outside of these particular problems; the issues in questions come down to computations and there is no ambiguity about the fact that the incorrect claims are in fact mathematically incorrect.

    [EDIT: having just noticed another post in this thread, perhaps I should caution against "guessing games" about which specific arXiv eprints I might have in mind. That would be profitless and in self-defense I will not respond to queries of that sort.]

    You can also search PF for a recent thread (last two weeks or so) in which I posted some computations of frequency shift phenomena for various pairs of observers, including some accelerating receivers, in the Schwarzschild vacuum solution.
    Last edited by a moderator: May 2, 2017
  5. Jan 21, 2007 #4

    "general" in the sense of arbitrary orientation (i.e. arbitrary angle [tex]\theta[/tex] and arbitrary relative motion between source and receiver). Sorry about confusing you. I will look for your posts on "some computations of frequency shift phenomena for various pairs of observers, including some accelerating receivers, in the Schwarzschild vacuum solution.". A pointer from you would be welcome.
    Last edited by a moderator: May 2, 2017
  6. Jan 21, 2007 #5

    The paper you quote does not appear to be correct since it is an attempt to plug in variable [tex]v[/tex] into the relativistic Doppler formulas. The oobjections I have to the paper are that:

    1. It does not derive the relativistic Doppler effect from base principles.
    2.Instead, it uses the formulas derived for non-accelerated motion and it plugs in the values computed for hyperbolic motion

    Looking thru your papers, I think this one is better
    The objections are smaller:

    1. It is not clear what are the improvements over reference [3]
    2. It deals only with source-receiver angle of motion [tex]\theta=0[/tex], i.e. is not general enough.
  7. Jan 21, 2007 #6


    User Avatar
    Staff Emeritus
    Science Advisor

    While no particular papers come to mind, the required technique seems fairly straightforwards to me. Since Chris has already given you some suggestions for papers, I'll just add in my $.02 on the techniques I'd suggest.

    You have a source, moving through some specific path through space-time, that emits signals as a constant time interval [tex]d\tau_1[/itex].

    The signals themselves follow null geodesis

    You have a destination, moving through some specific path through space-time, that receives the signals. We are interested in the interval between pulses (null geodesics) as a function of time. Note that the notion of simultaneity will, in general, depend on the coordinate system one uses, of course.

    If you happen to be in flat space-time, regardless of whether or not your observers are acclerating, the answer is easy enough to compute using inertial coordinates, for the null geodesics will be straight lines in any inertial coordinate system.

    This approach will basically demonstrate, when carried to its conclusion, that there is no doppler shift due to acceleration as long as one works in an inertial frame. There is only SR doppler shift in inertial frames in flat space-time, any other doppler shift comes from the choice of coordinates.

    I.e. pick a frame momentarily comoving with the transmitter. Draw the null geodesics in this frame. The time interval between reception will depend entirely on the velocity of the receiver relative to the inertial frame we defined, the "angle" at which the receiver's wordline "cuts across" the congruence of null geodesics.

    By computing the null geodesics for other geometries (say the FRW expanding universe - I've seen this partricular case done in textbooks for cosmological redshift and could probably dig up a reference in MTW), one can compute the doppler shift for those non-flat geometries as well. Rescaling the metric to "conformal time" can help this process a lot, greatly simplifying the geodesic equation.

    \frac{d^2x^\lambda }{dt^2} + \Gamma^{\lambda}_{~\mu \nu }\frac{dx^\mu }{dt}\frac{dx^\nu }{dt} = 0

    It's a lot easier to pick suitable coordinates to simplify the above equation than it is to solve it in the general case. Most particularly one can often simpolify this equation greatly by by rescaling time so that T = f(t) - i.e. "conformal time". This will often make light cones "straight lines", at least if one starts out with spatial coordinates that are the same in all directions (i.e. isotorpic coordinates for the Schwarzschild metric).

    The next most useful trick in dealing with geodesics is to take advantage of conserved quantities - every Killing vector in the source geometry generates a conserved quantity.

    The least useful technique is to attempt to solve the differential equations above by "brute force".
    Last edited: Jan 21, 2007
  8. Jan 21, 2007 #7
    Nonlocality in the Doppler Effect experiment

    Thank you for the attention you payd to my papers. I allways say "read the paper and not the place where it is published. Not all of us have access to top journals." My comments are inserted in your message. I would be happy to continue the discussion with you and let me know please the results you obtain in the field. Arxiv is a very democratic place where physicists could present the results they obtain.
  9. Jan 21, 2007 #8
    doppler effect and nonlocality

    You have a source, moving through some specific path through space-time, that emits signals as a constant time interval [tex]d\tau_1[/itex].

    That is the problem. What you propose is known as very small period assumption Most papers I know treat the problem making that assumption which obscures some peculiarities of the Doppler Effect.
  10. Jan 21, 2007 #9


    User Avatar
    Staff Emeritus
    Science Advisor

    I don't see it as any more problematical than in any other case where one takes a limit.

    One way of looking at doppler shift geometrically is that there is some function that maps [itex]\tau_1[/itex], the proper time of the emitter of some signal when that signal is emitted, into [itex]\tau_2[/itex], the proper time of the receiver when that same signal is received.

    The slope of that curve can be regarded as the doppler shift, ie

    [tex]\frac{d\,\tau_2}{d\,\tau_1}[/tex] (or perhaps the inverse).
  11. Jan 21, 2007 #10
    You know, even the top journals will accept papers from anyone, just as long as they eventually pass a peer review, and you can still keep a copy accessible on the ArXiv afterwards. (My comment would normally be off-topic, but I think the OP's plea for non-crackpottery deserves answer without reference to one's own arxiv-only writings.)
  12. Jan 21, 2007 #11
    Thank you, I'll give it a try.
  13. Jan 21, 2007 #12
    I think you missed my point: you cannot substitute v=variable in the formula for the realtivistic Doppler effect that was derived for inertial motion and expect to get a correct result.
  14. Jan 21, 2007 #13

    Chris Hillman

    User Avatar
    Science Advisor

    Just wanted to point out that this is most often employed in connection with "effective potentials", but is most successful in spacetimes with several independent Killing vector fields.

    However, techniques from computational commutative algebra involving Groebner bases, which can be roughly summarized as analogous to triangularization of a system of linear equations via Gaussian reduction, often are applicable, even in the case of systems of nonlinear ODEs (as is generally the case here, for reasons which should be obvious from the form of the geodesic equation as written out by pervect above). See Stephani, Differential Equations: Their Solution Using Symmetries, Cambridge University Press, 1989. Maple unfortunately has not yet provided a one command implementation of the scheme provided by Stephani, but there is no real difficulty in using his method. (I never seem to get around to providing a tutorial...)

    Another technique which can be useful is group analysis, in which one starts with a system including some undetermined function, and looks for particular choices which ensure that the system acquires extra symmetries, which often leads to exact solutions of special cases. Stephani has some remarks about this. Further introducutions to symmetries and differential equations include:

    1. Peter J. Olver, Applications of Lie groups to differential equations, Springer, 1993.

    2. Bluman and Kumei, Symmetries And Differential Equations, Springer, 1989.

    3. Brian J. Cantwell, Introduction to symmetry analysis, Cambridge University Press, 1992.

    4. Nail H. Ibragimov, Elementary Lie group analysis and ordinary differential equations, Wiley, 1999, and CRC handbook of Lie group analysis of differential equations, CRC Press, 1994.

    I'd recommend reading all of these.

    Incidently, Killing equations are very simple to attack by CAS. In Maple, the key command is "casesplit", which carries out the triangularization mentioned above (as in linear algebra, this can be tedious for humans, but computers enjoy such mindless tasks).
  15. Jan 21, 2007 #14
    doppler with acceleration and changing angle of incidence

    I did not. My humble oppinion is that the formulas which account for the Doppler Effect, in all its scenarios, we find in the literature of the subject, hold only in the case of the very high frequency assumption assuming that the observer receives two successive wave crests from the same point in space. In the particular case of the longitudinal Doppler Effect with constant relative velocity it is not necessary to make that assumption, because it is independent from the magnitude of the period. As you can see in my papers devoted to the subject, I derive formulas which are free of the very high frequency assumption accounting for all successive pair of wavecrests, in accordance with your point of view. I would be pleased to continue our discussion on that fascinating subject.
  16. Jan 21, 2007 #15


    User Avatar
    Staff Emeritus
    Science Advisor

    One "wrinkle" I glossed over - under some circumstances, it's possible to view a point source via multiple paths. Each of these paths can have its own associated doppler shift.
  17. Jan 22, 2007 #16


    User Avatar
    Staff Emeritus
    Science Advisor

    Interesting, thanks for the info and references. I've done enough with the problem to appreciate how messy solving the geodesic equations can be, so I'm glad to hear about new techniques and methods.
  18. Jan 22, 2007 #17
    You are still missing my point, you can't stick v= variable into the original relativistic Doppler effect. The relativistic Doppler effect is derived for v=constant , using the Lorentz invariance of [tex]k,\phi[/tex] . If you are working with accelerated motion, you cannot use the original formulas anymore, you need to derive things from base principles, as you did in your second paper. Nothing to do with "high frequency".
  19. Jan 22, 2007 #18
    doppler acceleration

    That is not only your point of view! I quote in my paper others who share the same point of view.
  20. Jan 22, 2007 #19

  21. Jan 22, 2007 #20
    I was just trying to help you see the error in the paper, I guess that you want to cover your ears ans shout : "la-la-la". Your choice....
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook