On Jun 15, 8:33�pm, pnachtwey <pnacht...@gmail.com> wrote:> On Jun 8, 8:30�pm, Roy <brewer....@gmail.com> wrote: > This was done with a Parallax Propeller Chip. �That is considerably > more processing power than an 8 bit micro controller.http://www.youtube.com/watch?v=4IZ5eAeneWk > I doubt that sample time is an issue with the Propeller chip. As > pointed out above there are probably other devices that are much more > limiting when it comes to response. > > Peter NachtweyTo be fair, I've seen similar performance out of vehicles that use 8- bit chips (PICs or AVRs). And its a pretty likely that he's using MEMs sensors. - Roy
Sample rate selection for digital control project
Started by ●June 9, 2009
Reply by ●June 15, 20092009-06-15
Reply by ●June 15, 20092009-06-15
On Jun 15, 9:21�pm, Malachy Moses <malachy.mo...@gmail.com> wrote:> > The "right type" of sensors are integrating-type sensors, whose output > is delta-V for accelerometers (i.e, not acceleration) and delta-theta > for gyros (i.e., not rate). �These sensors work by maintain a mass (or > spinning mass) at some null position using an internal control loop. > The output of the sensor is the control signal needed to maintain the > null position. > > With such a sensor, low rates for the control loop are possible, since > the sensor itself has integrated out all the high frequency details > that would otherwise cause problems (Nyquist etc). >Yes we will be using MEMS sensors, at least at first because they are the "right price". This project is essentially a radio controlled toy. Can you give me more details on the distinction between "delta-V" and acceleration, or "delta-theta" and rate gyro outputs, and how such sensors are able to get around the Nyquist "problem"? Thanks Roy
Reply by ●June 16, 20092009-06-16
On Mon, 15 Jun 2009 19:09:18 -0700, Roy wrote:> On Jun 15, 9:21 pm, Malachy Moses <malachy.mo...@gmail.com> wrote: > > >> The "right type" of sensors are integrating-type sensors, whose output >> is delta-V for accelerometers (i.e, not acceleration) and delta-theta >> for gyros (i.e., not rate). These sensors work by maintain a mass (or >> spinning mass) at some null position using an internal control loop. >> The output of the sensor is the control signal needed to maintain the >> null position. >> >> With such a sensor, low rates for the control loop are possible, since >> the sensor itself has integrated out all the high frequency details >> that would otherwise cause problems (Nyquist etc). >> >> > Yes we will be using MEMS sensors, at least at first because they are > the "right price". This project is essentially a radio controlled toy. > > Can you give me more details on the distinction between "delta-V" and > acceleration, or "delta-theta" and rate gyro outputs, and how such > sensors are able to get around the Nyquist "problem"?Rate integrating gyros have an electromechanical assembly with a rate input, and a control loop that maintains the gyro wheel centered within the case. The rate output from the system is a filtered version of the rate command to the gyro itself. Nowadays this rate output is integrated for the duration of one sample time, so aside from control loop bandwidth issues it a fairly close measure of the angle through which the gyro has passed in that sample time. Force balance accelerometers have a force (i.e. acceleration) input to the electromechanical assembly, and a control loop that maintains the proof mass centered within the case. The acceleration/delta-V output is very analogous to the delta-angle output of a rate integrating gyro. One approach wit MEMS sensors would be to sample your sensors _fast_ (i.e. well above their bandwidth), sum up their readings within the controller's sample period, then decimate at the controller's sample period. It's basically a one-step CIC (cascade integrator comb?) filter. Another approach is to just sample, and accept a high level of noise. This may be good enough, if you're just looking to keep the thing upright mostly. -- www.wescottdesign.com
Reply by ●June 16, 20092009-06-16
On Jun 16, 10:54�am, Tim Wescott <t...@seemywebsite.com> wrote:> One approach wit MEMS sensors would be to sample your sensors _fast_ > (i.e. well above their bandwidth), sum up their readings within the > controller's sample period, then decimate at the controller's sample > period. �It's basically a one-step CIC (cascade integrator comb?) filter. > > Another approach is to just sample, and accept a high level of noise. � > This may be good enough, if you're just looking to keep the thing upright > mostly. > > --www.wescottdesign.comUsing a CIC filter (or some sort of averager) was our concept as well. I think it may have to wait for the 32 bit processors - I doubt the 8-bit can handle the iteration rates, and the ARMs seem to have some capability to do A/D and perhaps even averaging in the "background" (without requiring direct s/w intervention). We'll probably always be stuck with MEMS sensors, however. I'm fairly confident we can get something to behave "good enough" with the AVR (similar to the video that Peter posted) from a radio- controlled standpoint. For fully autonomous operations we'll almost certainly need the 32 bit chips. Thanks again Roy
Reply by ●June 19, 20092009-06-19
Roy wrote:> On Jun 12, 7:41 pm, Tim Wescott <t...@seemywebsite.com> wrote: > >> Good 16-bit processors just aren't that much more expensive than 8-bit >> ones in hobbyist quantities; even the 32-bit ARM chips from Luminary/TI >> aren't _that_ much more. No matter what you do you'll spend ten times >> more on the non-processor part of your BOM -- why cut corners on the >> processing? >> >> --www.wescottdesign.com > > I don't disagree. My partner is the EE/board designer/microcontroller > guy, and he has experience with the AVR, so that's the first board he > made (I guess I'm the Aero E/controls/theory guy). I tried to talk him > out of it. He did like the Luminary Board, so that may be our next > board. To be fair, there are several examples of hobby and university > research/lab quadrotors out there that make use of 8-bit AVRs or PICs > as the main controller (but they are pretty much r/c only - not > autonomous).You want autonomous? On the cheap? Seriously? Consider the difference in processing needed by an RC sailboat and an autonomous one. Controlling rudder and sail set is easy. Deciding what to set to is the hard part. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●June 20, 20092009-06-20
On 19 Jun, 19:52, Jerry Avins <j...@ieee.org> wrote:> Roy wrote: > > To be fair, there are several examples of hobby and university > > research/lab quadrotors out there that make use of 8-bit AVRs or PICs > > as the main controller (but they are pretty much r/c only - not > > autonomous)....and they have lots of very smart but inexperienced students, who work 24/7 for weeks on end (with - mind you! - no pay), to get those things implemented. No need to follow the universities' lead unless you are troubled by scores of idle PhD-level cybernetics students hanging around the front gate.> You want autonomous? On the cheap? Seriously? Consider the difference in > processing needed by an RC sailboat and an autonomous one. Controlling > rudder and sail set is easy. Deciding what to set to is the hard part.Not necessarily impossible, though. You could set a number of way-points, let the thing loose and have the autopilot visit them all in sequence. Rune
Reply by ●June 20, 20092009-06-20
Rune Allnor wrote:> On 19 Jun, 19:52, Jerry Avins <j...@ieee.org> wrote: >> Roy wrote: >>> To be fair, there are several examples of hobby and university >>> research/lab quadrotors out there that make use of 8-bit AVRs or PICs >>> as the main controller (but they are pretty much r/c only - not >>> autonomous). > > ...and they have lots of very smart but inexperienced students, > who work 24/7 for weeks on end (with - mind you! - no pay), to > get those things implemented. > > No need to follow the universities' lead unless you are troubled > by scores of idle PhD-level cybernetics students hanging around > the front gate. > >> You want autonomous? On the cheap? Seriously? Consider the difference in >> processing needed by an RC sailboat and an autonomous one. Controlling >> rudder and sail set is easy. Deciding what to set to is the hard part. > > Not necessarily impossible, though. You could set a > number of way-points, let the thing loose and have the > autopilot visit them all in sequence.Indeed, but I doubt that an 8-bitter that also acts as autopilot and rudder and trim servos is up to the added requirements. In an R-C boat, the "skipper" controls the rudder and sail positions, appropriate to wind and intended destinations (your waypoints). An autonomous boat needs to sense the wind direction and make those decisions by computation. Updating the waypoints via GPS may not be accurate enough. Will we need onboard video to replace the skipper's eyes? Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●June 20, 20092009-06-20
On 20 Jun, 17:46, Jerry Avins <j...@ieee.org> wrote:> Rune Allnor wrote: > > On 19 Jun, 19:52, Jerry Avins <j...@ieee.org> wrote: > >> Roy wrote: > >>> To be fair, there are several examples of hobby and university > >>> research/lab quadrotors out there that make use of 8-bit AVRs or PICs > >>> as the main controller (but they are pretty much r/c only - not > >>> autonomous). > > > ...and they have lots of very smart but inexperienced students, > > who work 24/7 for weeks on end (with - mind you! - no pay), to > > get those things implemented. > > > No need to follow the universities' lead unless you are troubled > > by scores of idle PhD-level cybernetics students hanging around > > the front gate. > > >> You want autonomous? On the cheap? Seriously? Consider the difference in > >> processing needed by an RC sailboat and an autonomous one. Controlling > >> rudder and sail set is easy. Deciding what to set to is the hard part. > > > Not necessarily impossible, though. You could set a > > number of way-points, let the thing loose and have the > > autopilot visit them all in sequence. > > Indeed, but I doubt that an 8-bitter that also acts as autopilot and > rudder and trim servos is up to the added requirements.I have no idea what is required to get an airplane to fly. Maybe some sort of 'dithering' of the controls? Lack of precision in individual corrections is compensated by sheer number of corrections?> In an R-C boat, the "skipper" controls the rudder and sail positions, > appropriate to wind and intended destinations (your waypoints). An > autonomous boat needs to sense the wind direction and make those > decisions by computation. Updating the waypoints via GPS may not be > accurate enough. Will we need onboard video to replace the skipper's eyes?You questioned the OP's ambition of coming up with an autonomous aircraft, and then compared it with controlling a sailboat. I commented on the OP's ambition for the aircraft and disregarded your comparision with the sailboat. Anyway, the premise for the comparision is flawed: The aircraft has an on-board power plant while the sailboat does not. Rune
Reply by ●June 20, 20092009-06-20
On Sat, 20 Jun 2009 11:46:43 -0400, Jerry Avins wrote:> Rune Allnor wrote: >> On 19 Jun, 19:52, Jerry Avins <j...@ieee.org> wrote: >>> Roy wrote: >>>> To be fair, there are several examples of hobby and university >>>> research/lab quadrotors out there that make use of 8-bit AVRs or PICs >>>> as the main controller (but they are pretty much r/c only - not >>>> autonomous). >> >> ...and they have lots of very smart but inexperienced students, who >> work 24/7 for weeks on end (with - mind you! - no pay), to get those >> things implemented. >> >> No need to follow the universities' lead unless you are troubled by >> scores of idle PhD-level cybernetics students hanging around the front >> gate. >> >>> You want autonomous? On the cheap? Seriously? Consider the difference >>> in processing needed by an RC sailboat and an autonomous one. >>> Controlling rudder and sail set is easy. Deciding what to set to is >>> the hard part. >> >> Not necessarily impossible, though. You could set a number of >> way-points, let the thing loose and have the autopilot visit them all >> in sequence. > > Indeed, but I doubt that an 8-bitter that also acts as autopilot and > rudder and trim servos is up to the added requirements. > > In an R-C boat, the "skipper" controls the rudder and sail positions, > appropriate to wind and intended destinations (your waypoints). An > autonomous boat needs to sense the wind direction and make those > decisions by computation. Updating the waypoints via GPS may not be > accurate enough. Will we need onboard video to replace the skipper's > eyes? > > JerryAs Rune mentioned Roy's problem is probably simpler, or at least less subtle. I suspect that if you know the boat's speed, position and direction, and if you know the wind speed and direction relative to the boat, that you have all the information that you need to at least competently handle the boat -- doing it with real flair would take either lots of programming, or a really good human skipper. -- http://www.wescottdesign.com
Reply by ●June 20, 20092009-06-20
Rune Allnor wrote:> On 20 Jun, 17:46, Jerry Avins <j...@ieee.org> wrote:...>>>> You want autonomous? On the cheap? Seriously? Consider the difference in >>>> processing needed by an RC sailboat and an autonomous one. Controlling >>>> rudder and sail set is easy. Deciding what to set to is the hard part. >>> Not necessarily impossible, though. You could set a >>> number of way-points, let the thing loose and have the >>> autopilot visit them all in sequence. >> Indeed, but I doubt that an 8-bitter that also acts as autopilot and >> rudder and trim servos is up to the added requirements. > > I have no idea what is required to get an airplane > to fly. Maybe some sort of 'dithering' of the controls? > Lack of precision in individual corrections is compensated > by sheer number of corrections? > >> In an R-C boat, the "skipper" controls the rudder and sail positions, >> appropriate to wind and intended destinations (your waypoints). An >> autonomous boat needs to sense the wind direction and make those >> decisions by computation. Updating the waypoints via GPS may not be >> accurate enough. Will we need onboard video to replace the skipper's eyes? > > You questioned the OP's ambition of coming up with an > autonomous aircraft, and then compared it with controlling > a sailboat. I commented on the OP's ambition for the aircraft > and disregarded your comparison with the sailboat.A boat moves in two dimensions, an aircraft in three. I thought that a simpler problem might make the difficulty of autonomy clearer. Despite a few years of DARPA-sponsored competition among university teams, no autonomous automobile is robust enough for serious consideration.> Anyway, the premise for the comparison is flawed: The > aircraft has an on-board power plant while the sailboat > does not.So consider an autonomous power boat, then. Power or sail, autonomy is the sticky part. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
