Are flamethrowers and missile launchers right out? They can be dramatic and very interactive.
Erco, of course not. Can we step it up a bit, ala Burning Man style and use a 5-gallon propane tank and 12V solenoid? Do I need to be sending you an Arlo right now? I think you might have the time with the recent turn of events.
Here is a type of bumpers from DeAgostini Robot. This system is another type of whiskies. When the robot hit an obstacle it gives a signal in order to stop the robot. Although it is effective I think that is very fragile. A strong crash is able to damage the entire mechanism.
So for a big robot like the Arlo I think that the bumpers must be from aluminum, very strong with springs for elasticity, and must also have a type of switch in order to understand the collisions and give the appropriate signal to the Robot (image on post #28).
What about a bluetooth module with a simple AndroidApp with some buttons/sliders etc. so that you can easily control some functions of the ARLO with your phone.
A user-friendly GUI. Phipi did a wonderful job on GUIs for Scribbler and S2, so he's your go-to guy. Think of Arlo as a giant reconfigurable S2. If Arlo was as easy to program as an S2, you would sell a lot more. Isn't that a good thing?
This would require standardizing the package with encoders & various sensors. Also good for sales. You guys are entitled to make a profit for all your hard work. Big bonuses for all Parallax employees!
A user-friendly GUI. Phipi did a wonderful job on GUIs for Scribbler and S2, so he's your go-to guy. Think of Arlo as a giant reconfigurable S2. If Arlo was as easy to program as an S2, you would sell a lot more. Isn't that a good thing?
This would require standardizing the package with encoders & various sensors. Also good for sales. You guys are entitled to make a profit for all your hard work. Big bonuses for all Parallax employees!
I totally second that. Even a simple GUI for displaying sensor values and having control over motors would be indescribably useful during construction and troubleshooting.
Although I really think my idea of offering a vision board and writing some Prop C code for it should get the grand prize (a lifetime supply of chocolates), I will second Nikos' idea for fenders/bumpers. Not only is this the kind of thing our Contest Master, MattG, likes to do, it is also functionally useful, not to mention wise. The mechanism will need to be far more robust that Nikos' follow up on post #35, but the general idea is sound.
Sourcing parts shouldn't be too difficult. Small shock absorbers are common for RC cars, and I'm sure Parallax can get a wholesale arrangement with one of the bigger hobby distributors. I'd recommend two each front and back, with additional bars for mechanical support.
The sensor can be some simpler arrangement than what most people usually envision. Just don't use standard pushbutton switches. They'll be destroyed in no time. Piezo film or discs are a nice idea, and cheap, but they won't always detect a slow depression against an object.
Isn't there a small sensor that can detect heat? Mounted on a mobile robot R/L and a servor for U/D couldn't something like this give a robot to detect human beings? IIRC, someone had a thread around here on cheap (very low res) thermal imaging. A Propeller's video ability combined with a moving sensor would be very interesting if this could be made to work.
Isn't there a small sensor that can detect heat? Mounted on a mobile robot R/L and a servor for U/D couldn't something like this give a robot to detect human beings?
Well known limitations to these.
If you're going to make a killer robot, its vision can't just be thermal. Yul Brenner proves it.
What about a wrap around sensor holder/bumper sensor?
It could be made of plastic, with a rubber bumper around the outside, like a rub rail, that is able to sense when it is hit/pushed and signal where along it's length it was triggered.
Then the plastic part could have holes pre-built into it for PING and IR sensors to provide 360 degrees of Sonar and InfraRed vision.
It could also have holes aimed and ready for mounting the same to focus down to act as cliff detectors.
The entire assembly could be made to wrap around and attach to the Arlo.
Here is a picture of something similar for another robot:
The sensor holder part itself, like above, isn't too amazing, although it would be a nifty addition to the Arlo if properly designed, but if you could also design a wrap around bumper with triggering electronics into it, that would be amazing.
But a much more versatile and useful option would be something where the fence itself was passive and required no power. That way you could just install it anywhere, like across the exit to your classroom, or the top of a set of stairs, and leave it.
This would require a fancier sensor of course.
It could be visual, color, pattern, ultraviolet, etc. That has benefits because you could also see if your fence was "working" visually, although it could be easy to fool a visual sensor with changing light or other objects in the robot's normal area.
It could be something like NFC, but it would have to be able to be deployed in a "stripe" and the sensor would have to be able to detect it from at least the ground clearance height of the robot. No dragging NFC readers on the floor.
If non-powered was not possible at least something that could run for periods of time no battery power would be more useful that something that required AC power.
Robot Beacon:
Another option, along the same lines, but also useful in other ways, would be a beacon of some sort. It could be IR or a radio frequency. This could be placed in a location and the robot could interact with it, either being told to find it or to avoid it, or even to retrieve it!
Both would require hardware and software and give students and hobbyists more tools to work with when writing their own applications.
Arlo needs a smart bumper: use a softpot (a long mylar strip potentiometer switch) as a wraparound perimeter bumper. Any contact is readable on one pin, ADC or RCtime tells you exactly where you're hitting. Hi resolution yields accurate object/obstacle location and is simpler than having numerous individual bump switches.
To extend an idea someone else had, I would really like to see a programmable GUI, instead of a set GUI.
To be more specific, it would be really cool to be able to program GUI windows in the SimpleIDE editor. The available graphics would include windows that resemble the windows that the programmer can create using the JComponent class in Java.
I have a number of ideas currently that would work, and look, a lot better if I could create something like a JLabel.
Soft pots are fairly fragile. I'd opt for a mechanical switch with a stop in the bumper itself to prevent damage to the plunger and contacts. Basically what every modern car has in the brake pedal to activate the lights, or the clutch to serve as a starter interlock. Those switches are hardy, but the mechanism also uses a stop to keep them from being jammed beyond recognition.
A length of surgical tubing sealed at one end and connected to a pressure sensor could provide the sensing and springiness at the same time.
I've tried that, and it sorta works, but it's not as straightforward as it looks. I discovered (at least I think) the air inside has to be pressurized, not just sealed at atmospheric pressure. Otherwise there's not enough to trigger the sensor, I believe because there is very little volume of air in the tube. If would be great if the system could use the pressure sensor Parallax already sells. Using a pressurized hose makes for a more complicated mechanism. Too, the hose has to be very thin-walled to present as little resistance as possible. Might be fairly prone to damage.
I've tried that, and it sorta works, but it's not as straightforward as it looks. I discovered (at least I think) the air inside has to be pressurized, not just sealed at atmospheric pressure. Otherwise there's not enough to trigger the sensor, I believe because there is very little volume of air in the tube. If would be great if the system could use the pressure sensor Parallax already sells. Using a pressurized hose makes for a more complicated mechanism. Too, the hose has to be very thin-walled to present as little resistance as possible. Might be fairly prone to damage.
My thought was to use larger diameter 1/2" silicone tubing between the robot and a rigid bumper, relatively thick walled to provide the springiness. I can't image that a pressure sensor could not detect the tube being squeezed. A pinhole might be required so that it equalizes with ambient pressure. Some time ago when Parallax asked about pressure sensors I strongly suggested to use one with a nipple as it would be more versatile. Oh well...
You might try it. Your combination of sensor and tubing could do the trick. it's been a while since I experimented, but with the setup I had the sensor wasn't reliable.
A contour edge might help (though I don't think Parallax has the extra axis on their CNC), as would (probably) a more sensitive sensor. I used one I got surplus, and they're fairly common nowadays with more to chose from.
When I researched these, I found the ones they use (or used to use) at gas stations were pressurized.
The add on object I would recommend would be a camera. The camera would have build in capability for face detection. Uses would include recognition of students it comes in contact with which it can compare to images in the digital yearbook. It can then wirelessly transmit the name and location of the student to a central server that keeps track of attendance. Should a student ever go missing or be wandering where it shouldn't be, like in the case of Avonte Oquendo, the robot can send an alert to the faculty network in real time.
Telepresence package to make Arlo into a security patrol bot: wifi, camera with pan/tilt, monitor, and enough sensors to detect light, sound, motion and navigate a preprogrammed path autonomously and also driven remotely. Also need an autodock recharger.
But don't send it - drive it up here with your bike in the car.
Actually, that sounds like a great idea. I'd love to have a few days to brainstorm with your people up there. Any time you can get PhiPi, Mike Green, Chip and Tracy Allen on-site, I'm all over it! I'll easily be the dumbest and least talented guy in the room, but I'm good with that.
There have been many great "game changer " ideas suggested here which might require more development time than Parallax has. I suspect the winner here may be something very useful but mundane like the power distribution board. Nothing wrong with that. But if you like an idea here that doesn't win, just do it yourself. Nothing stopping you from creating the next big thing that might go viral on Kickstarter. Just do it!
Comments
Erco, of course not. Can we step it up a bit, ala Burning Man style and use a 5-gallon propane tank and 12V solenoid? Do I need to be sending you an Arlo right now? I think you might have the time with the recent turn of events.
Ken Gracey
You're a good man, Ken! But please no Arlo, I don't need distractions right now. I'm seizing this opportunity to finish my big idea.
Maybe I should be sending you my robot.
Maybe so! Carpe diem, amigo!
But don't send it - drive it up here with your bike in the car.
Ken Gracey
Here is a type of bumpers from DeAgostini Robot. This system is another type of whiskies. When the robot hit an obstacle it gives a signal in order to stop the robot. Although it is effective I think that is very fragile. A strong crash is able to damage the entire mechanism.
So for a big robot like the Arlo I think that the bumpers must be from aluminum, very strong with springs for elasticity, and must also have a type of switch in order to understand the collisions and give the appropriate signal to the Robot (image on post #28).
-Matt-changing-the-rules-as-we-go-G
Yes I was being snarky. We are eagerly anticipating the power distribution board...
This would require standardizing the package with encoders & various sensors. Also good for sales. You guys are entitled to make a profit for all your hard work. Big bonuses for all Parallax employees!
I totally second that. Even a simple GUI for displaying sensor values and having control over motors would be indescribably useful during construction and troubleshooting.
Sourcing parts shouldn't be too difficult. Small shock absorbers are common for RC cars, and I'm sure Parallax can get a wholesale arrangement with one of the bigger hobby distributors. I'd recommend two each front and back, with additional bars for mechanical support.
The sensor can be some simpler arrangement than what most people usually envision. Just don't use standard pushbutton switches. They'll be destroyed in no time. Piezo film or discs are a nice idea, and cheap, but they won't always detect a slow depression against an object.
Well known limitations to these.
If you're going to make a killer robot, its vision can't just be thermal. Yul Brenner proves it.
It could be made of plastic, with a rubber bumper around the outside, like a rub rail, that is able to sense when it is hit/pushed and signal where along it's length it was triggered.
Then the plastic part could have holes pre-built into it for PING and IR sensors to provide 360 degrees of Sonar and InfraRed vision.
It could also have holes aimed and ready for mounting the same to focus down to act as cliff detectors.
The entire assembly could be made to wrap around and attach to the Arlo.
Here is a picture of something similar for another robot:
The sensor holder part itself, like above, isn't too amazing, although it would be a nifty addition to the Arlo if properly designed, but if you could also design a wrap around bumper with triggering electronics into it, that would be amazing.
Like this article: http://www.parallax.com/news/2014-04-09/parallax-prepares-usa-science-and-engineering-festival-washington-dc-april-26-27
But a much more versatile and useful option would be something where the fence itself was passive and required no power. That way you could just install it anywhere, like across the exit to your classroom, or the top of a set of stairs, and leave it.
This would require a fancier sensor of course.
It could be visual, color, pattern, ultraviolet, etc. That has benefits because you could also see if your fence was "working" visually, although it could be easy to fool a visual sensor with changing light or other objects in the robot's normal area.
It could be something like NFC, but it would have to be able to be deployed in a "stripe" and the sensor would have to be able to detect it from at least the ground clearance height of the robot. No dragging NFC readers on the floor.
If non-powered was not possible at least something that could run for periods of time no battery power would be more useful that something that required AC power.
Robot Beacon:
Another option, along the same lines, but also useful in other ways, would be a beacon of some sort. It could be IR or a radio frequency. This could be placed in a location and the robot could interact with it, either being told to find it or to avoid it, or even to retrieve it!
Both would require hardware and software and give students and hobbyists more tools to work with when writing their own applications.
Per http://forums.parallax.com/showthread.php/132260-Smart-Bumper-for-Robot
That idea was ahead of its time!
Agreed, in theory. But in the real world, but neither IR nor sonar (ultrasonics) are foolproof. Contact switches are the last line of defense.
To be more specific, it would be really cool to be able to program GUI windows in the SimpleIDE editor. The available graphics would include windows that resemble the windows that the programmer can create using the JComponent class in Java.
I have a number of ideas currently that would work, and look, a lot better if I could create something like a JLabel.
As Erco sez, IR only sees things that reflect infrared light. Place a thick black sock near an IR sensor. Odds are it won't see it.
Soft pots are fairly fragile. I'd opt for a mechanical switch with a stop in the bumper itself to prevent damage to the plunger and contacts. Basically what every modern car has in the brake pedal to activate the lights, or the clutch to serve as a starter interlock. Those switches are hardy, but the mechanism also uses a stop to keep them from being jammed beyond recognition.
I've tried that, and it sorta works, but it's not as straightforward as it looks. I discovered (at least I think) the air inside has to be pressurized, not just sealed at atmospheric pressure. Otherwise there's not enough to trigger the sensor, I believe because there is very little volume of air in the tube. If would be great if the system could use the pressure sensor Parallax already sells. Using a pressurized hose makes for a more complicated mechanism. Too, the hose has to be very thin-walled to present as little resistance as possible. Might be fairly prone to damage.
My thought was to use larger diameter 1/2" silicone tubing between the robot and a rigid bumper, relatively thick walled to provide the springiness. I can't image that a pressure sensor could not detect the tube being squeezed. A pinhole might be required so that it equalizes with ambient pressure. Some time ago when Parallax asked about pressure sensors I strongly suggested to use one with a nipple as it would be more versatile. Oh well...
A contour edge might help (though I don't think Parallax has the extra axis on their CNC), as would (probably) a more sensitive sensor. I used one I got surplus, and they're fairly common nowadays with more to chose from.
When I researched these, I found the ones they use (or used to use) at gas stations were pressurized.
Thanks,
Marcus
Everything Rovio should have been.
Lots of good entries. Will get to judging RSN!
-Matt "ka-bloooey & 90 minute tow" G.
If you favor unreliable cars, I'll make you a righteous deal on a '67 Corvair. JK!
-erco "unsafe at any speed" O.
Actually, that sounds like a great idea. I'd love to have a few days to brainstorm with your people up there. Any time you can get PhiPi, Mike Green, Chip and Tracy Allen on-site, I'm all over it! I'll easily be the dumbest and least talented guy in the room, but I'm good with that.
There have been many great "game changer " ideas suggested here which might require more development time than Parallax has. I suspect the winner here may be something very useful but mundane like the power distribution board. Nothing wrong with that. But if you like an idea here that doesn't win, just do it yourself. Nothing stopping you from creating the next big thing that might go viral on Kickstarter. Just do it!