Re: Re: Suggestion for Version 1.0 motion application API...
Herman Bruyninckx <[email protected]>
| Newsgroups | gmane.science.robotics.orocos.user |
|---|---|
| Message-ID | <Pine.LNX.4.44.0405102329260.27522-100000@srv04.mech.kuleuven.ac.be> |
On Mon, 10 May 2004, Walter Verdonck wrote: > > [...] > > A Path is a sequence of ViaPoints; > > Isn't this a quite restricted definition? To me, a Path is more like a > trajectory, e.g. a sinusoid, which specifies the relation between time > and position. > The Path is fully defined by the ViaPoints, _and_ by the behaviour of the Components in the _application_. The (choice of) behaviour is not part of the data flow, the path-influencing information in the ViaPoints is. If you want a sinusoid there are various alternatives: - the ViaPoint contains the parameters of the sinusoid (amplitude, phase, time duration,..., but also spatial direction in which to execute the sinusoid), and a property that identifies the ViaPoint as a sinusoidal one. The _application_ will generate the sinusoid, provided its Generator has the corresponding functionality. - the Path contains ViaPoints that are sampled from a sinusoid. The ViaPoint's information can be more than just tangent direction and velocity: also higher order parameters are possible; hence, the sinusoid could be approximated to any desired accuracy. That's what happens in most current machine tools, I guess. - the ViaPoint could describe a "process" which results in the generation of a sinusoid. [...] > If I understand this well, a ViaPoint is the point you want to reach > exactly, sometimes called the Precision Point. Does your definition > also include fly-by points, i.e. point you want to approach with a > certain velocity or within a certain tolerance? Absolutely! Tolerances etc are (could be) _parameters_ of the ViaPoint. So, there will be a large collection of ViaPoints (or Paths, etc.). This is, in my opinion, nothing more but a faithful reflection of the reality... Note that part of the information of the resulting path is contained in the Generator of the application! The same ViaPoint given to different interpolators can result in different trajectories. I know there is still a lot of filling-in-the-ViaPoint-details to do :-) Herman