Loading ...
Sorry, an error occurred while loading the content.

Near field power transfer ideas needed

Expand Messages
  • <rlnansel@...>
    I m working on several sculptural installations incorporating lots of BEAM-ish robots. These will all be indoors, and for exhibition purposes I need a
    Message 1 of 13 , Feb 1, 2014
    • 0 Attachment
      I'm working on several sculptural installations incorporating lots of BEAM-ish robots. These will all be indoors, and for exhibition purposes I  need a significant fraction of the of the machines to be active and doing visually interesting things. Batteries are out because I might have as many as a hundred machines; past experience with robotic museum exhibits tells me if I have four battery-powered robots, after the first day three of them will have dead batteries and the fourth will have a mechanical malfunction.

      Concentrating on the power supply issue, it seems to me to boil down to a handful of choices:

      1) Use PV panels with really bright lights  positioned overhead. I know this works, but the lights might have to be brighter -- and  hotter -- than I'd like to get the sort of activity I want. But maybe I could supply IR floodlights matched more closely to the peak sensitivity of the PV cells. I'd still have some visible light floods, but the IR floods would provide most of the power. I'm thinking this should allow a more efficient power transfer and conversion with less waste heat. Has anybody tried this before?

      2) Use inductive windings embedded in the top surfaces of the display plinths. Each bot would possess its own pickup coil to draw inductively coupled AC power feeding a rectifier and energy storage system. The power source coils would be driven at an ultrasonic frequency, say 30kHz or so. The amount of power that could be coupled this way wouldn't be huge, but it would be steady and with standard solarengine circuitry could give regular pulses of activity. The trick might be making the inductive pickups light and compact.

      3) Similar to 2, but using near-field resonant coupling techniques. I have no experience with this and don't know a lot about it, but it looks to make higher coupling efficiency possible than plain vanilla inductive coupling.

      4) Something else. Maybe a fan blowing over the plinth, and robots having one or more propellor(s) driving high-efficiency PMDC motor/generators? Fixed position charging stations ("plants"), with any of the above methods providing auxiliary power to robots between "feedings"? Heck, a fan would allow some of the machines to be purely mechanical wind-powered robots. (Would they still qualify as BEAM 'bots, though, if they didn't have any electronics aboard? Hmmm...)

      The goal is to create something with the aesthetics of an ecosystem. Motion and purposeful behaviours are important parts of this, but sounds will be important, too. For the latter I'm imagining times during the cycle of the system where behaviours emerging akin to a "dawn chorus". Some parts of the whole installation (not necessarily incorporated inTO individual robots) will sense the presence of human viewers, so I plan to play around with various "alert" and "danger" calls.

      I'm a sculptor, but I also have a degree in robotics engineering, so I can handle just about any ideas y'all might throw out. (Well, probably not radioisotope batteries or fuel cells, but you get the idea.)

      (Sorry for the lengthy post. So many ideas buzzing in my head.)

      Thanks in advance for your help!

      -Bobby

    • <rlnansel@...>
      Sorry for double posting, but ideas keep a buzzing. More ideas: using something like conductive elastomer zebra strips (such as are used to connect LCD
      Message 2 of 13 , Feb 1, 2014
      • 0 Attachment

        Sorry for double posting, but ideas keep a buzzing.


        More ideas: using something like conductive elastomer "zebra strips" (such as are used to connect LCD electrodes on glass to PCB contacts) to make walker feet. The surface the walkers walk on is a tiled pattern of isolated copper squares or maybe hexagons, each connected to either B+ or B-. As long as the tile size is small enough compared to the walker foot size, each bot could be guaranteed to have each foot straddle at least two tiles of different polarity.


        It would require lots of full-wave rectifiers, one for each conductive segment in the zebra strips, and the impedance might be kind of high, thus limiting how much current could be drawn from each foot pad, but might this work?


        Another way would be to use pogo probes for direct point contact at each foot. The tiles would need enough isolation distance between them so a pogo probe foot couldn't short out adjacent tiles when it stepped on the "crack" between the tiles. This would mean that not every pogo probe foot could make contact every time it was planted, but for a 4-legged walker chances should be excellent that at least two of its feet would be planted on tiles of opposite polarity .


        A third possibility: a bumpy/corrugated surface where B+ and B- are available, everywhere, but with one on "hills" and one in "valleys". Pogo probes as above, with the horizontal distance between hills and valleys chosen so a pogo probe foot can contact only one or the other at a time. Have to take special precautions two or more 'bots planting their respective feet in adjacent hills or valleys can't create a short circuit should their legs be touching.


        -Bobby
      • winglabs_inc
        Interesting. But does this really cause the bots to have any purposeful behavior aside from walking, by which I mean behaviors that both contribute to the
        Message 3 of 13 , Feb 2, 2014
        • 0 Attachment

          Interesting. But does this really cause the bots to have any purposeful behavior aside from walking, by which I mean behaviors that both contribute to the bot's survival and that of its habitat. I have had a few far-fetched ideas for BEAM ecosystems before, though. My favorite one involves "predatory" walkers/crawlers. Since these bots are either photophobes and prefer to stay hidden, or simply unable to compete with the large number of other photovores for light as a power source, they turn instead to capturing or latching on to otherwise competing bots, and somehow drain their charge. If they were to prey on photopoppers, they would have to be extremely power conservative, if they were to survive from photopopper to photopopper. Also, they would have to be able to detect their prey's size and distance, and be smart enough to know whether they will gain energy by going for a kill. This could be achieved by simply comparing two weights, and if the +(prey's energy) outweighs the -(own energy), it triggers a taxi toward the prey. My guess is that the bot would be able to sense its prey's charge(and size) through either capacitive or inductive sensors(or both). From there, some interesting defensive behaviors could be observed from the photopoppers?

          Something like that would be very fun and interesting if plausible. And while these sorts of bots would be predatory, they would be more of exoparasites than carnivores, unless they actually disassembled their prey and made some use of it, which would be incredibly awesome, but we're not there yet, sadly. Perhaps one day.


          Enjoy, Connor.

        • <rlnansel@...>
          Purposeful behaviour, in this case, consists of: 1) looking good; 2) doing interesting things often enough that, in a swarm of a dozen or so bots, one or two
          Message 4 of 13 , Feb 3, 2014
          • 0 Attachment

            Purposeful behaviour, in this case, consists of:

             1) looking good;

             2) doing interesting things often enough that, in a swarm of a dozen or so 'bots, one or two of them will be in motion at any given time;

             3) surviving unattended for at least as long as a day on a purpose-made platform;

             4) exhibiting variable behaviours under varying conditions (i.e. an artificial "sun" might periodically pass overhead).


            I'm supposing the ultimate plan would be to have several different "species" of 'bots inhabiting the same area. How they interact is of considerable interest, of course. It's also important that they all be "compatible" with each other in the sense that their collective normal range of behaviours shouldn't cause outright destruction. A big-ass turbot flopping around and crushing little Herbies or Photovores in the process would be an example of an Incompatibility.


            -Bobby


          • Jurva-Markus Vehasmaa
            Looking good is not behavior...
            Message 5 of 13 , Feb 3, 2014
            • 0 Attachment

              Looking good is not behavior...

              3.2.2014 11.14 kirjoitti <rlnansel@...>:
               

              Purposeful behaviour, in this case, consists of:

               1) looking good;

               2) doing interesting things often enough that, in a swarm of a dozen or so 'bots, one or two of them will be in motion at any given time;

               3) surviving unattended for at least as long as a day on a purpose-made platform;

               4) exhibiting variable behaviours under varying conditions (i.e. an artificial "sun" might periodically pass overhead).


              I'm supposing the ultimate plan would be to have several different "species" of 'bots inhabiting the same area. How they interact is of considerable interest, of course. It's also important that they all be "compatible" with each other in the sense that their collective normal range of behaviours shouldn't cause outright destruction. A big-ass turbot flopping around and crushing little Herbies or Photovores in the process would be an example of an Incompatibility.


              -Bobby


            • <rlnansel@...>
              Sure it is. Not all of aesthetics comes in the static design; some of it should come from the *way* the bots do their thing. For instance, I would consider an
              Message 6 of 13 , Feb 3, 2014
              • 0 Attachment

                Sure it is. Not all of aesthetics comes in the static design; some of it should come from the *way* the 'bots do their thing. For instance, I would consider an elegant-looking 'bot with an irregular, herky-jerky walking gait to be "looking bad" because it's behaviour is not consonant with it's appearance. Remember, this is a sculptural installation, and "just getting the job done" isn't good enough; it has to be done with some visual or audible *style*.


                -Bobby


                ---In beam@yahoogroups.com, <jurva.vehasmaa@...> wrote:

                Looking good is not behavior...

                3.2.2014 11.14 kirjoitti
              • Martin McKee
                This is where my writing-tutor self comes in -- with the niggliy little word stuff. In a general sense, yes, behavior is what something does. Which means
                Message 7 of 13 , Feb 3, 2014
                • 0 Attachment
                  This is where my writing-tutor self comes in -- with the niggliy little word stuff.  In a general sense, yes, behavior is "what something does."  Which means that whatever the bots are doing, that would be behavior.  Furthermore, given that the purpose of the installation is aesthetic, the behavior becomes a part of that aesthetics.  However, here on the very much biomorphically-based BEAM list ( nice alliteration! ), behavior carries a strong connotation of relating to survival.  As such, it seems, to some degree, incompatible with use as an aesthetic descriptor.  But, it is, as they say, all semantics.  Looking good is not behavior ( an adjective is not a noun ), but behavior can certainly be part of looking good ( one characteristic of the definition ).

                  I wish I had some better ideas for power distribution, but, though I've done a fair amount with PV, I've done nothing with near-field power distribution.  I think in either case, available power levels will be quite low.  The trick, therefore, is to make every action count ( from a visual standpoint ).  Another option would be to use multiple power systems - PV from the top, inductive or resonant from below.  I would stay away from the alternately charged grid ideas because I would more-or-less expect the robots to find their way into configurations that would simply short things out.  Power stations, however, seem like a good possible option -- any time you can collect power without actually expending much in the process, it's a good thing.

                  Putting on my artist cap ( darkroom photography and sculpture ), I would definitely work toward using some light producing bots ( perhaps the power plants ) and photovores.  This could add inexpensive ( certainly from an energy standpoint ) interest with a "firefly effect" and inspire more complex behavior if other bots are photovoric.  Of course, given the inclusion of bright lights for PV, it would be even better to have modulated LEDs with filtered sensors so that ambient light could be dealt with.

                  Sounds like a very interesting project.  I look forward to hearing about what you decide upon.

                  Martin Jay McKee


                  On Mon, Feb 3, 2014 at 2:40 AM, <rlnansel@...> wrote:
                   

                  Sure it is. Not all of aesthetics comes in the static design; some of it should come from the *way* the 'bots do their thing. For instance, I would consider an elegant-looking 'bot with an irregular, herky-jerky walking gait to be "looking bad" because it's behaviour is not consonant with it's appearance. Remember, this is a sculptural installation, and "just getting the job done" isn't good enough; it has to be done with some visual or audible *style*.


                  -Bobby


                  ---In beam@yahoogroups.com, <jurva.vehasmaa@...> wrote:

                  Looking good is not behavior...

                  3.2.2014 11.14 kirjoitti


                • winglabs_inc
                  Or you could include an aquatic environment within the ecosystem, in which all land based bots must be able to avoid water. But power transmission would be
                  Message 8 of 13 , Feb 3, 2014
                  • 0 Attachment

                    Or you could include an aquatic environment within the ecosystem, in which all land based bots must be able to avoid water. But power transmission would be easy enough for the aquatic bots. Simply course AC power into the water, and the bots should be able to collect it. You could have solar "algae" that collect solar energy and discharge it into the water.
                    However, I wouldn't just give unlimited energy to the habitat, like it seems you want to do. I mean, if you look at a real ecosystem, its entire structure is based upon limited resources. Ever organism has to contend with both a relatively scarce amount of energy, as well as competition for that same energy. If you just give every bot in the habitat as much energy as it wants without having to obtain it, you have no ecosystem, just a bunch of critters leeching off of your power socket. But if you have limited energy that isn't distributed evenly, then a ecological structure can be built from there, by imbuing each species with a set of behaviors that allow them to compete successfully.
                    Also, I would encourage you to introduce "incompatible" species with one another, like turbots and photopoppers, as they are only incompatible if you make them compete directly. Sure, if a turbot catches a smaller bot in its territory, that bot is done for, but in a structured ecosystem, they would have behaviors that help them avoid confrontation with a more dominant species. Meaning that if turbots are chasing the light, then the photopoppers have to do something other than going straight for the light. Like I have a design for a 4 eyed photopopper that circles around the light, rather than straight into it. This could be adapted to where the photopopper attempts to follow the light while staying under some cover.

                    Or even just have nocturnal photovores. Have moveable "plants"(which could be made to look very interesting using flexible solar cells) that have separate solar engines for each "leaf" on each side, and when night comes these plants light up an LED on the opposite side of where they received the most light. This LED light attracts nocturnal photovores, which dive into the plant and push it toward the area where the plant last sensed the most light came from. This way, a symbiosis is formed, seeing as the nocturnal photovores can move around, but don't know where to go to get the most light, and the plants know where to go, they just can't move.


                    Enjoy, Connor.



                    ---In beam@yahoogroups.com, <martinjaymckee@...> wrote:

                    This is where my writing-tutor self comes in -- with the niggliy little word stuff.  In a general sense, yes, behavior is "what something does."  Which means that whatever the bots are doing, that would be behavior.  Furthermore, given that the purpose of the installation is aesthetic, the behavior becomes a part of that aesthetics.  However, here on the very much biomorphically-based BEAM list ( nice alliteration! ), behavior carries a strong connotation of relating to survival.  As such, it seems, to some degree, incompatible with use as an aesthetic descriptor.  But, it is, as they say, all semantics.  Looking good is not behavior ( an adjective is not a noun ), but behavior can certainly be part of looking good ( one characteristic of the definition ).

                    I wish I had some better ideas for power distribution, but, though I've done a fair amount with PV, I've done nothing with near-field power distribution.  I think in either case, available power levels will be quite low.  The trick, therefore, is to make every action count ( from a visual standpoint ).  Another option would be to use multiple power systems - PV from the top, inductive or resonant from below.  I would stay away from the alternately charged grid ideas because I would more-or-less expect the robots to find their way into configurations that would simply short things out.  Power stations, however, seem like a good possible option -- any time you can collect power without actually expending much in the process, it's a good thing.

                    Putting on my artist cap ( darkroom photography and sculpture ), I would definitely work toward using some light producing bots ( perhaps the power plants ) and photovores.  This could add inexpensive ( certainly from an energy standpoint ) interest with a "firefly effect" and inspire more complex behavior if other bots are photovoric.  Of course, given the inclusion of bright lights for PV, it would be even better to have modulated LEDs with filtered sensors so that ambient light could be dealt with.

                    Sounds like a very interesting project.  I look forward to hearing about what you decide upon.

                    Martin Jay McKee


                    On Mon, Feb 3, 2014 at 2:40 AM, <rlnansel@...> wrote:
                     

                    Sure it is. Not all of aesthetics comes in the static design; some of it should come from the *way* the 'bots do their thing. For instance, I would consider an elegant-looking 'bot with an irregular, herky-jerky walking gait to be "looking bad" because it's behaviour is not consonant with it's appearance. Remember, this is a sculptural installation, and "just getting the job done" isn't good enough; it has to be done with some visual or audible *style*.


                    -Bobby


                    ---In beam@yahoogroups.com, <jurva.vehasmaa@...> wrote:

                    Looking good is not behavior...

                    3.2.2014 11.14 kirjoitti

                  • <rlnansel@...>
                    I don t follow how a bot might harvest power from an AC current flowing in a body of water. Either there has to be a voltage gradient to tap conductively
                    Message 9 of 13 , Feb 4, 2014
                    • 0 Attachment


                      I don't follow how a 'bot might harvest power from an AC current flowing in a body of water. Either there has to be a voltage gradient to tap conductively (capacitively?) or a well-defined current axis to tap inductively with a coil. Could you elaborate on what you have in mind?

                      ---In beam@yahoogroups.com, <connor_ramsey@...> wrote:

                      Or you could include an aquatic environment within the ecosystem, in which all land based bots must be able to avoid water. But power transmission would be easy enough for the aquatic bots. Simply course AC power into the water, and the bots should be able to collect it. You could have solar "algae" that collect solar energy and discharge it into the water. ...



                    • winglabs_inc
                      Probably not. I threw it out there, exactly because I m not sure how it would work. But it sounds plausible.
                      Message 10 of 13 , Feb 6, 2014
                      • 0 Attachment
                        Probably not. I threw it out there, exactly because I'm not sure how it would work. But it sounds plausible.
                      • Martin McKee
                        Serendipity allowed me to stumble across this video from linear technologies, http://www.eetimes.com/video.asp?section_id=124&doc_id=1320930. If nothing else
                        Message 11 of 13 , Feb 12, 2014
                        • 0 Attachment
                          Serendipity allowed me to stumble across this video from linear technologies, http://www.eetimes.com/video.asp?section_id=124&doc_id=1320930 .  If nothing else it demonstrates just how much is possible with a resonant RF power transmission system with minimal complexity ( using purpose built ICs ).  Looks like a selection of near-field power points could certainly be usable.

                          Martin Jay McKee


                          On Thu, Feb 6, 2014 at 8:06 PM, <connor_ramsey@...> wrote:
                           

                          Probably not. I threw it out there, exactly because I'm not sure how it would work. But it sounds plausible.


                        • <rlnansel@...>
                          Wow, I just ran across this chip earlier today. The datasheet can be found here http://cds.linear.com/docs/en/datasheet/4120fa.pdf. I could do a lot with 400mA
                          Message 12 of 13 , Feb 12, 2014
                          • 0 Attachment
                            Wow, I just ran across this chip earlier today. The datasheet can be found here.

                            I could do a lot with 400mA ...

                            -Bobby
                          • Martin McKee
                            There s no doubt that LT makes some awfully fun toys. Given that the pickup coil is integrated into the PCB in the reference design, it looks like something
                            Message 13 of 13 , Feb 12, 2014
                            • 0 Attachment
                              There's no doubt that LT makes some awfully fun toys.  Given that the pickup coil is integrated into the PCB in the reference design, it looks like something fairly easy to build into a small mobile robot.  This device, though, is not in what I would call a fun package ( QFN-16 ); and, like most LT stuff, it's not cheap ( about 5usd a piece ).  But, yes, 400mA might just as well be tens of amps for most BEAM robots.  It would be an almost unlimited power source.

                              I can't help but think that multiple power supplies with many different types of robots wouldn't create the most interesting "ecology."  Each of the power sources is limited and so the bots would have to compete for that limited resource.  However, if each could use two types of power ( say near-field and solar ) they would have the choice to fill up "now" or fill up slowly.  They could have extremely simple power usage strategies -- some aggressive, some conservative.  In any case, the different behaviors and their interactions, would, to some degree, be driven by the availability of power, so more power == more activity.  Combine that with power providing "plants" that turn on and off... there could be lots happening.

                              Martin Jay McKee


                              On Wed, Feb 12, 2014 at 6:43 PM, <rlnansel@...> wrote:
                               

                              Wow, I just ran across this chip earlier today. The datasheet can be found here.


                              I could do a lot with 400mA ...

                              -Bobby


                            Your message has been successfully submitted and would be delivered to recipients shortly.