[Rock-dev] CopterCommand to base/types

Matthias Goldhoorn matthias.goldhoorn at dfki.de
Wed May 30 10:15:12 CEST 2012


On 30.05.2012 10:05, Felix Rehrmann wrote:
> For sure there is no point in cluttering base/types with a diversity of
> command types.
>
> But on the other hand if there are general types to be reused by a
> number of packets it would
> be nice to have them in a central place. Example: Definig some command
> type in a Quadrocopter
> driver and then the use the type in an airplane driver later,  would
> feel not so organized, too.
>
> I think a generic command type is not the right thing to do
> (Circle-Driving could be given as radius and speed).
> And for safety reasons it might be good to have command types that
> really fit the system, so that a command cannot
> exchanged with another command that is not meant for it.
this could be archived even with an specialized vehicle controller, that 
only accepts the right combination of values and goes otherwise in 
exception state.

Not sure at all about these at all, so we vote again?

1. creating generalized message like the rbs
2. splitting vehicle commands in an "base-command-types" package and 
move the existing ones to there
3. putting it in vehicle specific packages, and leave the existing ones 
where they are.

My vote:
1+ for #1, (seems for me the best way to create 
rock-standart-controllers and share code)
strong -1 for #3, because we couldn’t create generic-shared controller 
then in rock
meaning for #2 it's okay, but still prefer #1


>
> So command-type packets would be nice. In addition we could define
> command types in the specific packet. And if it is about
> to become more generic, it could be moved to a generic command-type packet.
>
> Felix
>
> Am 29.05.2012 22:12, schrieb Matthias Goldhoorn:
>> What about designing an ONE DataItem like the rbs?.
>>
>> could be:
>>
>> double[3] translation_velocity;
>> double[3] position;
>>
>> double[3] rotation_velocity;
>> double[3] target_orientation;
>>
>> this should cover all cases... (setting nan/inf simliar to the rbs)
>>
>> Otherwise we should move all types to specialized classes like
>> RoverCommands, WaterVehicleCommands etc...
>> I'm against real System specific commands, i vote for generic vehicle
>> topic based packages, that should be part of rock (but not of base).
>>
>> Matthias
>>
>>
>> Am 29.05.2012 17:33, schrieb Jakob Schwendner:
>>> On 05/29/2012 05:22 PM, Leif Christensen wrote:
>>>> On 29.05.2012 17:10, Matthias Goldhoorn wrote:
>>>>> I also need an ASV Motion Command like:
>>>>>
>>>>> double translation[2];
>>>>> double heading;
>>>>>
>>>>> We should discuss about generalize "motion commands"
>>>>>
>>>>>
>>>>> Matthias
>>>>>
>>>>> On 29.05.2012 16:58, Jakob Schwendner wrote:
>>>>>> On 05/29/2012 04:44 PM, Felix Rehrmann wrote:
>>>>>>> Hello,
>>>>>>>
>>>>>>> I would like to add a type CopterCommand to base/types.
>>>>>>> It is to give commands to flying machines and will have 3 axis commands
>>>>>>> and one thrust command:
>>>>>>>
>>>>>>> struct CopterCommand {
>>>>>>>            double axis_command[3];
>>>>>>>            double thrust;
>>>>>>> };
>>>>>>>
>>>>>>> The value range would be from -1.0 to 1.0 for the axes and from 0.0 to
>>>>>>> 1.0 for the thrust. Any opinions?
>>>>>>>
>>>>>> seems very system specific to me. Any reason this couldn't be defined in
>>>>>> the module that requires it?
>>>>>>
>>>>>> Jakob
>>>> I think I would not try to generalize motion commands at all, because of
>>>> the wide range of inputs even for vehicles of similar classes. Think for
>>>> example of all the different possible inputs for a quite narrow
>>>> specified class of planar ackerman vehicles. Specifying such classes +
>>>> deriving from them seems more work to me than directly define it in the
>>>> vehicle specific modules.
>>> +1 even the ones that are currently there could probably go into vehicle
>>> class specific modules.
>>>
>


-- 
  Dipl.-Inf. Matthias Goldhoorn
  Unterwasserrobotik

  Hauptanschrift Standort Bremen:
  DFKI GmbH
  Robotics Innovation Center
  Robert-Hooke-Straße 5
  28359 Bremen, Germany

  Phone: +49 (0)421 178 45-4193
  Fax:   +49 (0)421 178 45-4150
  E-Mail: robotik at dfki.de

  Weitere Informationen: http://www.dfki.de/robotik
  -----------------------------------------------------------------------
  Deutsches Forschungszentrum fuer Kuenstliche Intelligenz GmbH
  Firmensitz: Trippstadter Straße 122, D-67663 Kaiserslautern
  Geschaeftsfuehrung: Prof. Dr. Dr. h.c. mult. Wolfgang Wahlster
  (Vorsitzender) Dr. Walter Olthoff
  Vorsitzender des Aufsichtsrats: Prof. Dr. h.c. Hans A. Aukes
  Amtsgericht Kaiserslautern, HRB 2313
  Sitz der Gesellschaft: Kaiserslautern (HRB 2313)
  USt-Id.Nr.:    DE 148646973
  Steuernummer:  19/673/0060/3




More information about the Rock-dev mailing list