There is a 10% chance my job gets automated, how about you? Take the test!
Robots are entering the workforce. Some will work alongside you. Others, sadly, will put you out of a job. The question is, which jobs are actually on the chopping block?
There are too few women in engineering, however projects and programs with strong societal benefits seem to attract them to the field.
With 1.3 billion transistors, my computer would have cost about $10.4 billion in the 1960s.
The topic everyone in the world should be talking about.
Great article discussing the future of AI, and how it will impact our lives. A good, fairly lengthy review of everything going on right now in the space.
I have often been asked: “how do I get into robotics?”. Well, here are a few related (and not so related) thoughts on that.
The first thing you want to look at is what kind of robots you are interested in: Industrial or Research.
Industrial robots, for now, tend to be concentrated in manufacturing. These are the ABB arms, Kinova stock handling, even Roomba vacuums, etc. Usually, these are fairly “dumb” robots in that they can be programmed to do a fairly simple set of tasks, over and over again. For this market segment, there are fairly entrenched standards of operation and design, others would know more about this than I. The product cycle tends to be fairly long, but the products are very reliable.
For Research robots, I am talking about robots which are not yet fully implemented in ongoing work. These robots are the ones that tend to be one-offs, in military, mining, human interaction, research, etc. This industry is currently growing faster, and has the potential to replace all current “dumb” robots. For this industry, a good systems level understanding is key. Of course, you need to be an expert in your field, but understanding the basics of software, mechanical and UX is a good idea as well. This concept was introduced to me at a talk a long time ago, as “T-shaped people”.
Another aspect that I would suggest is simply to get rid of the idea of “work-life balance”. Now I do not mean work yourself to death, (though there is the temptation to just fix that last bug before bedtime) I mean that a growing, rapidly changing field like robotics requires you to live and breathe it. Subscribe to journals, talk with industry experts, read blogs, reports (Automaton is a favorite of mine), make little robots in your spare time. Turn the question “how do I get into robotics?” into more of “where else can I learn more about robotics?”.
On the software side, I strongly suggest learning ROS, as it is a powerhouse in tying together the research robotics community, it’s free, and runs on little stuff like Arduino and Raspberry Pi.
Lastly, I would read some of Ray Kurzweil’s books. His “How to Create a Mind” and “The Singularity Is Near” touch on the subject of AI and the future of robotics.
Past, Present & Future
I like to think that the robotics industry today is as the computer industry was in the 80s. Just getting out of the military/industry and starting to make an appearance in everyday life. Right now, any garage hobbyist, with a little effort, can compete on the world stage in robotics. You don’t yet need millions of dollars to make an impact. In fact, many of the traditional companies trying to get into “mobile robots” and smart robotics are doing very poorly despite pouring in millions. The first “killer app” for robotics in the home seems to be the Roomba, but the next one will be much more serious. Figuring that out will make someone very rich. (like Bill Gates and the spreadsheet).
Another aspect is the decreasing cost and increasing power of computing and sensors. It is now fairly inexpensive to make a robot that can use depth sensing cameras (Kinect) to map a house, look up object data online, and navigate to pick up an object.
The downside is, both actuators and batteries have stagnated quite severely in comparison, so mobile robots remain weak and short-lived.
As always, drop me a comment if you want anything featured on the site!
Circuit Board printing at home
This week I want to talk about a great project another Waterloo
sufferer grad has created: A all-in-one circuit board printer, paste dispenser and re-flow oven. You can do circuit board printing at home!
Whenever you want to create an electronics project, you usually have to have some amount of circuit design. (Unless you are playing with
Lego Arduino) This usually cools the fire a bit on your project, as you have one of two options for PCB manufacture:
- Make it yourself
- Pay a circuit board house to make it
Traditionally, I had always done #1, with ferric chloride, a laser printer, and a lot of patience. I definitely recommend making circuit boards yourself as a student, because it is the fastest, cheapest, and best way to learn about the PCB process. However, as I have gotten relatively busier (and relatively less-poor), I had switched to getting my circuit boards made for me. (Check here for links) The upside is less time spent huffing questionably safe chemicals, the downside is a trade-off of time for money. The faster you want your board, the more you are paying for it.
- Cheap ++
- Fast +
- Labor Intensive +++
- 2 layer board possible, hard to do
- Cheap or Fast, not both
- High Precision
- Multilayer boards
- Easy +++
- Fast ++
- Easy ++
- “2 layer board” possible
* Note that Cheap in the Voltera case is only if you are making a significant number of boards. At $1.4k, the cost of the machine pays for itself in as little as 10 prototype boards made with an overnight service like APC.
Here is the killer app of the idea though: The swapable head. I give huge kudos to the Voltera team for developing an incredibly complex paste that can be printed, but is then also temperature stable enough for soldering at reasonable temperatures. Being able to press the start button and walk away is a huge boost to productivity. When I had to etch boards and solder paste stencils by hand, it was cheap, but required hours and multiple re-tries to get it right. A slight downside is the need to swap heads twice each print, once for traces, once for isolation print (to get 2 layers), and again for connecting traces.
After printing, the system can dispense solder paste to each pad. When I started making circuits, I hated solder paste and reflow, as carefully placing the components one by one seemed slow and more error prone than simply soldering each one. However, the magic of reflow is that each part will float a little, and correct its own position. This allows you to be less careful with positioning than you would be with hand soldering.
About the only downside I can call out about the printer is the lack of drilling ability. With the swapable head setup, it should be fairly easy to design in a small rotary bit attachment to automatically drill holes. This would allow vias and through-hole components to be used. Recently, Voltera posted an update with a step in the right direction for through hole components. The printer will leave a small non-inked circle in the middle of the through hole pad, to prevent the drill head from drifting while drilling.
With the right material provided by Voltera, I am sure that the next version (or maybe even software upgrade to this version) could allow for printing on both sides, drilling a hole through, and filling the hole with conductive paint. This would allow the setup to truly start to compete with traditional 2 layer boards. I suspect the Voltera team wisely chose not to overreach for the Kickstarter campaign, and may sell the drill attachment as an add-on later.
It remains to be seen how long printed boards survive in harsh environments. Things like vibration, corrosion, and even dust can destroy home printed boards. I would love to see if the printed material is flexible enough to use on flexible plastic. This would allow a whole range of great applications, from smart clothing to innovative sensors.
It is interesting to see a lot of the same movements that had been made in the 3D printer space now being applied to the electronics space. Makerbot and similar companies have proven that people are willing to spend a few thousand on making small 3D models. Will they be willing to do the same for electronics? A clever move for Voltera now would be to open a “Thingiverse” type setup for electronic circuits. Perhaps, let people upload a design and a BOM. Creating a marketplace where people can sell both designs and entire kits would be a great way for Voltera to generate ongoing revenue. The other question is whether or not Voltera will be able to hang on to their secret formula for conductive ink. If aftermarket inks start proliferating, it could bite into future sales.
You can check out the project here: Voltera
Their Kickstarter is running for another 6 days, so go now!
Thanks for the shout-out for the ruler 🙂
I have officially started a new position at Clearpath Robotics: Senior Electrical Designer!
Other than that, updates to this website are ongoing, with older projects being added to the portfolio page, and cross linked as a blog post.
Upcoming topics are:
– My experience doing small time hardware start-ups and pitches.
– Reverse Geocache gift idea
– Upcoming Gaming Chair fun project
– Hardware encryption, my concept of how it could work, and why we need it
– Group buys for stock markets, modeling and moving into real world (XKCD applies…)
If you have any topic suggestions, or want me to write about a topic of interest to you (as long as it is robotics related) drop me a line.
Business Insider has listed Clearpath Robotics founders as some of the Top 40 under 40 today!
I couldn’t agree more, this is a fantastic place to work!
Transport Canada, the governmental body regulating all aircraft in Canada, announced new and simpler rules for UAV (drone) usage in Canada on Nov 5th.
The most important effect is the clarification of how UAVs are to be regulated in Canada. In the past, any commercial use (including videotaping from quadrotor, posting to YouTube, and getting add revenue) required getting an SFOC. This was a fairly painful and bureaucratic process. Not getting the SFOC and getting caught could mean tens of thousands of dollars of fines. Transport Canada is now revising this rule.
The bullet points:
- UAVs cleared for any use under 2Kg (most quadrotors and RC planes out there) even commercial one like the aeryon scout.
- “certain operations involving UAVs under 25 kilograms” are also cleared, though exactly what that means is so far unclear
What does this mean?
Well for one thing, amateur aerial videographers can breathe a little easier. This new regulation actually makes Canada one of the most permissive countries when it comes to UAV use. Canada is now even more permissive than the United States for UAV use, as different states have wildly different laws about their use. One of the most controversial rules in the States is the requirement to only fly in line of sight. This makes the growing sport of FPV (First Person View) flying technically illegal in some states. Keeping the drone in sight is suggested in Canada, but not a legal requirement as far as I understand it.
When I was a co-op at Aeryon Labs in Waterloo, we always had problems getting an SFOC approved in a timely and reasonable matter. This could sometimes lead to delays in our flight plans. This new regulation should make life a lot simpler for all GIS and info gathering uses of quadrotors.
Note that this new regulation just removes the SFOC impediment, it does not relax the regulations covering UAV usage. It is still up to the pilot to use the drone safely. Here is a handy guide:
Overall the new regulations should see more drone usage in every day work, while not decreasing safety in any reasonable way.
Full list of Rules Here
The new rules should come into effect at the start of December.