Ask HN: How do I get into robotics?
14 comments
One of the first projects I would start is look for Arduino related projects and other small hardware projects. Its best for cheap and simple electronics understanding. Learn how to use different sensors, modules or controllers.
After some time, you can take a look at more serious projects, such as robot-spiders, or robots manipulators.
Might you also be interested in buying cheap 3D printer, to make the parts such as gears or cases for your hardware projects.
After you are quite comfortable with electronics and hardware, you can make your own 3D printed robot, such as CM6 - Compliant 3D printed robotic arm [1] or WE-R2.4 Six-Axis Robot Arm [2].
Happy building!
[1] https://hackaday.io/project/180588-cm6-compliant-3d-printed-... [2] https://www.thingiverse.com/thing:3327968
After some time, you can take a look at more serious projects, such as robot-spiders, or robots manipulators.
Might you also be interested in buying cheap 3D printer, to make the parts such as gears or cases for your hardware projects.
After you are quite comfortable with electronics and hardware, you can make your own 3D printed robot, such as CM6 - Compliant 3D printed robotic arm [1] or WE-R2.4 Six-Axis Robot Arm [2].
Happy building!
[1] https://hackaday.io/project/180588-cm6-compliant-3d-printed-... [2] https://www.thingiverse.com/thing:3327968
Commenting to revisit later! Thanks for these resources.
Awesome! There are a few different paths you can take, depending on what type of robotics you are interested in.
Consumer/Commercial robotics. These are things like Roombas and Robomows. The dominate type of engineer in these companies is embedded systems and firmware engineer. You should probably switch to an electrical or computer engineering and get experience with some robotics program at your school which is smaller scale and works with microcontrollers and C/C++.
Advanced robotics. Things like Boston Dynamics. These companies also have embedded system engineers, but it is more important to have high-level controls experience. You should look for a CS or other degree which includes control theory, Digital Signal Processing, and so on. A lot of universities have robotics programs (think DARPA challenge) and teams, definitely join one of those.
Good luck!
Consumer/Commercial robotics. These are things like Roombas and Robomows. The dominate type of engineer in these companies is embedded systems and firmware engineer. You should probably switch to an electrical or computer engineering and get experience with some robotics program at your school which is smaller scale and works with microcontrollers and C/C++.
Advanced robotics. Things like Boston Dynamics. These companies also have embedded system engineers, but it is more important to have high-level controls experience. You should look for a CS or other degree which includes control theory, Digital Signal Processing, and so on. A lot of universities have robotics programs (think DARPA challenge) and teams, definitely join one of those.
Good luck!
Full disclosure: I'm just a hobbyist game dev with experience with arduino, and volunteering at my local robotics competition.
I would recommend getting into arduino. It's fun to buy little components, build little circuits, and see how your ability to program can have a physical manifestation.
As for programming concepts -- yes kinematics and inverse kinematics will allow you to move robotic limbs to certain points in real space.
And you'll find yourself wanting a larger conceptual framework to put those technical bits. That's where things like finite state machines, hierarchical finite state machines, behavior trees, and/or goal oriented action planners may come in handy. This grouping of concepts is often called "ai programming" but whatever you decide to call it, they're algorithms to decide what to do at any given slice of time.
Obviously follow your own path in this subject; follow the bits that resonate with you.
I would recommend getting into arduino. It's fun to buy little components, build little circuits, and see how your ability to program can have a physical manifestation.
As for programming concepts -- yes kinematics and inverse kinematics will allow you to move robotic limbs to certain points in real space.
And you'll find yourself wanting a larger conceptual framework to put those technical bits. That's where things like finite state machines, hierarchical finite state machines, behavior trees, and/or goal oriented action planners may come in handy. This grouping of concepts is often called "ai programming" but whatever you decide to call it, they're algorithms to decide what to do at any given slice of time.
Obviously follow your own path in this subject; follow the bits that resonate with you.
If you truly want to learn, I suggest joining a local FRC Robotics team as a new/junior mentor. If you simply ask the students to show you what they have been working on, you may be inspired to join in and help. You'll learn together... mechanical, electrical, programming, etc.
A short overview of what FIRST robotics is: https://www.youtube.com/watch?v=Jd29kzjclV0
A web based tool to search for local teams: https://www.firstinspires.org/team-event-search#type=teams&s...
I've been volunteering as a FIRST mentor for the last nine years and have learned more about robotics from the students than I did in my master's program or on my own.
A short overview of what FIRST robotics is: https://www.youtube.com/watch?v=Jd29kzjclV0
A web based tool to search for local teams: https://www.firstinspires.org/team-event-search#type=teams&s...
I've been volunteering as a FIRST mentor for the last nine years and have learned more about robotics from the students than I did in my master's program or on my own.
Thanks for the resources Jon, although unfortunately I am in the UK. I will however investigate if there is something similar over here.
If anyone else has information pertaining to how to begin working on small projects that would be great. Even just a list from the top of your head of things to learn in some chronological order.
If anyone else has information pertaining to how to begin working on small projects that would be great. Even just a list from the top of your head of things to learn in some chronological order.
Check out James Bruton's excelent youtube channel https://www.youtube.com/c/jamesbruton/videos
The general process he uses in his projects is worth emulating, as well as many of his design choices. He also provides 3d print files and code.
As far as a specific project that would be a good entry point, I reccommend making some variety of inverted pendulum which is actively balanced by a microcontroller. Every skill needed to achieve this is fundamental to succeeding and having fun in robotics.
The general process he uses in his projects is worth emulating, as well as many of his design choices. He also provides 3d print files and code.
As far as a specific project that would be a good entry point, I reccommend making some variety of inverted pendulum which is actively balanced by a microcontroller. Every skill needed to achieve this is fundamental to succeeding and having fun in robotics.
Look into ROS and it's associated projects.
There are several different "compatible" arms which are good for learning.
https://www.ros.org/ https://github.com/ros
And as others have mentioned it wouldn't hurt to gain access to a 3d printer and possibly some other microcontroller kits.
Quite a few writeups out there if you do some searching around though.
There are several different "compatible" arms which are good for learning.
https://www.ros.org/ https://github.com/ros
And as others have mentioned it wouldn't hurt to gain access to a 3d printer and possibly some other microcontroller kits.
Quite a few writeups out there if you do some searching around though.
There is a field of study called Mechatronics that combined Mechanics and Electronics though that is not the only path
For more hobbyist stuff, join the Lego FIRST competitions though note that the Mindstorms product line has been recently discontinued though Educators may still be able to obtain the Spike set. Look into open source ROS operating system
E.g. control engineers can also get into robotics and it's better not to havea more broad degree in my humble opinion
For more hobbyist stuff, join the Lego FIRST competitions though note that the Mindstorms product line has been recently discontinued though Educators may still be able to obtain the Spike set. Look into open source ROS operating system
E.g. control engineers can also get into robotics and it's better not to havea more broad degree in my humble opinion
Take a look at the already assembled robot arms available on eBay, they vary a lot in price but even the cheapest ones could be used to explore the software side of controlling a robot arm. The main thing you'll learn from programming a cheap robot is what its limitations are (lack of fine control or repeatability, lack of strength, quick to wear out, difficulties in handling overload situations). Then you'll know what you want to study next and have some idea of terminology (always important to learn the special words technologists use in their field). So, for example, you'll learn that servos work, but are slow and not very powerful or repeatable. Stepper motors may be an improvement but have limited resolution due to moving in steps. Moving fast requires a lot more power. Adding sensors to measure the actual rotation at joints (or distance for linear actuators) can improve repeatability. If you want to learn the electronics side of robotics start building simple circuits using a solderless breadboard, you can get parts at JameCo, DigiKey or Mouser (do a web search). There are lots of web sites with projects you can build. The ones that explain the theory as well as show how to build the circuit will teach you the most. Electronics has a lot of detail in it, there are thousands of types of components, many ways of interfacing things, lots of limitations in power and frequency, it is somewhat like software programming in those respects. Jumping in and building things alongside more formal learning paths is a good way to get started. Stick with low voltage, low current circuits to start with until you learn how to safely deal with higher voltages and currents (which can be deadly). Learning mechanics it is probably simpler to start with theory rather than building things. There are not a lot of kits or examples for building mechanical things that I know of (Lego? Erector sets?) that also teach you theory (stress, deformation, mass versus speed, calculating maximum load at the end of an arm, trajectory computation, etc.) The ROS software library and some exposure to machine learning could be helpful as well. Higher level software topics like SLAM (simultaneous location and mapping) and path planning and machine vision you can look at to start learning and study in more detail later. Sensors are an entire field in themselves (mechanical, optical, chemical, RF, pressure, humidity, cameras, air pressure, color, distance,...) A robot is like a person, complicated and built of systems on top of systems and you need to learn at least something about all the levels to be proficient in designing a complete system that works the way you want it to. It's a lot to learn, though you can also specialize in those parts you find more interesting. Some people do nothing but path planning or machine vision.
The area of study you're looking for is called Kinematics and if you do want to go down the building an arm route, you'll need a more than passing familiarity with Electronics and microcontrollers, since you like python, a Raspberry Pi is probably the best choice.
Perfect, thanks!
See if you can find a nearby Makerspace - you'll likely find they have short informal courses in basic electronic tinkering and it will give you access to a variety of useful equipment (and hopefully some interesting people too).
Duckietown [1]
[1] : https://www.duckietown.org
[1] : https://www.duckietown.org
FYI: Second year maths degree, work in Python and SQL, entry level knowledge of C and CPP.
I was inspired by the fluid movement displayed in the litter picker sort of arm one of the Boston Dynamic robots has and I would like to learn how to go about creating such a robot as a side project (just the arm). I want to know what resources I should look at to get started? I have no real experience soldering, EE or working with sensors. So some entry level projects recommendations may be best?
Thank you.