by StuartL » Sat Apr 19, 2008 9:25 am
by StuartL
Sat Apr 19, 2008 9:25 am
For me robot engineering is all about experimentation, imagination and discovery. When I bought the Bioloid kit I bought it with the intention of building a variety of robots and this I've done. I've now got a modified version of the standard kit humanoid and I'm tinkering with the software.
Building a biped has many steps (pun intended) and each one may be easy or a challenge for you:
1) Figure out a conceptual layout of what you want to build. This with include approximate physical design, size, abilities and features. As will all home engineering you are likely to revise this over time as trial and error occurs.
2) Figure out how to build the physical form of the robot. Work out which actuators/servos go where, where the hinges are, if you're going to attempt ball joints, where to put the power source and brain(s). Try to bear in mind construction and maintenance so that you can remove and replace an actuator that's broken.
3) Figure out how to build the control electronics for the robot. This may be a "simple" Stamp controller or a complex array of processors and/or microcontrollers working together. It may include off-robot processing or be entirely self contained.
4) Revise the above as necessary.
5) Build it. You'll probably find there's bits you can do easily and bits you need assistance for. You may be lucky enough to own a CNC machine or you may choose to craft your joints out of plywood. Your family or friends may work in electronics and can help you out or you may choose to buy a PCB manufacturing system off ebay.
The biggest piece of advice I have for you is to start small. Build a line following robot before you attempt a maze solving robot. Build a closed loop control system in a small robot before trying to get the biped to walk. The lessons you learn small scale and with simple joints are all valuable for larger joints.
It's for this reason that I have two robots. The humanoid is the ultimate aim. This is temporarily stalled as we're currently programming the snake. What we're learning in the snake's simple motion is all relevant to the biped and is improving the control software daily without the frustrations of constantly falling over.
And it's a SNAKE!
For me robot engineering is all about experimentation, imagination and discovery. When I bought the Bioloid kit I bought it with the intention of building a variety of robots and this I've done. I've now got a modified version of the standard kit humanoid and I'm tinkering with the software.
Building a biped has many steps (pun intended) and each one may be easy or a challenge for you:
1) Figure out a conceptual layout of what you want to build. This with include approximate physical design, size, abilities and features. As will all home engineering you are likely to revise this over time as trial and error occurs.
2) Figure out how to build the physical form of the robot. Work out which actuators/servos go where, where the hinges are, if you're going to attempt ball joints, where to put the power source and brain(s). Try to bear in mind construction and maintenance so that you can remove and replace an actuator that's broken.
3) Figure out how to build the control electronics for the robot. This may be a "simple" Stamp controller or a complex array of processors and/or microcontrollers working together. It may include off-robot processing or be entirely self contained.
4) Revise the above as necessary.
5) Build it. You'll probably find there's bits you can do easily and bits you need assistance for. You may be lucky enough to own a CNC machine or you may choose to craft your joints out of plywood. Your family or friends may work in electronics and can help you out or you may choose to buy a PCB manufacturing system off ebay.
The biggest piece of advice I have for you is to start small. Build a line following robot before you attempt a maze solving robot. Build a closed loop control system in a small robot before trying to get the biped to walk. The lessons you learn small scale and with simple joints are all valuable for larger joints.
It's for this reason that I have two robots. The humanoid is the ultimate aim. This is temporarily stalled as we're currently programming the snake. What we're learning in the snake's simple motion is all relevant to the biped and is improving the control software daily without the frustrations of constantly falling over.
And it's a SNAKE!