EmbeddedRelated.com
Forums

Reference about Network Topologys applyed to Robotics or Automation

Started by Sink0 January 22, 2011
Any one got any good reference for most common hardware topologies
applied to robotics?

Topologies i mean, how the sensors or actuators are connected to the
main controller. As an example, if the sensors and actuators are
connected directly to the embedded hardware (for microcontrollers) or
expansion cards (for embedded PCs like a PC/104) it would be a very
centralized topology.

I am not looking knowledge, just a good and solid reference, becouse
one points of my msc is about a hardware architecture for robotics
applications that is easily expandable, with a good throughput,
average latency, very reduced number of cables (a very important point
for a high reliability), and mainly, almost no change on the hardware
connected to the main controller for very different applications
(basically a multchannel, real multidrop, with distribuited processing
network). And i need this reference becouse all knowledge i got right
know comes from lots of articles and works, but nothing solid that
goes deeply on this subject.

Thank you!
On 22.01.2011 19:13, Sink0 wrote:
> Any one got any good reference for most common hardware topologies > applied to robotics?
No such thing.
> Topologies i mean, how the sensors or actuators are connected to the > main controller.
Whichever way works. And that depends a lot, if not entirely, on what those sensors and actuators actually are, how big the entire apparatus is, how complex it is, and what the main driving point of the design is: cost, power, speed, flexibility, etc...
> =A0And that depends a lot, if not entirely, on what > those sensors and actuators actually are, how big the entire apparatus > is, how complex it is, and what the main driving point of the design is: > cost, power, speed, flexibility, etc...
Thats true, thats the reason i was asking for some possible reference on this field becouse teachers seems to do not be happy about if some option is a good one. They always want some sort of comparison with other possibilities and they ask why i chose this and not others options. I can really argue why my project is a good one for my specific purpose (biomechatronics applications) going throught all its specific chareacteristics, but as there are too many variables for a comparison, this task seems almost impossible (and too much waste of time). Any way.. Other question.. any good theory of how to make comparisons of comunication protocos for deterministic aplications? I got this Article http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=3D&arnumber=3D898793 , But i dont know if there are other references? Basicaly It compares some characteristics as Frame Time, Propagation Time, Blocking Time, Efficiency, Average time delay and Utilization... for Ethernet, ControlNet, and DeviceNet.