Below are some pictures of a new biped that I just started putting together with Lynxmotion parts. This is going to be a long-term project for me and I expect it to easily take many months if not years.
An interesting aspect of the robot is that I’ll be using it as a test bed for the OpenServo. The OpenServo takes stock analog RC servos and converts them with new PCBs to digital servos using open hardware and software. The OpenServo feature an AVR ATmega MCU, serial interface (currently I2C), feedback of position and current and other features geared towards robotic control. The OpenServo is still very much under active development and right now I would encourage you to only build your own if you are very brave, but hopefully building my own robot will help evolve them quicker to better meet the needs of hobbiest roboticists.
I’m very impressed with the flexibility and quality of the Lynxmotion servo erector set. It allowed me to easily build legs with six degrees of freedom that appear to be much more human-like in their movement than many other very expensive biped robot kits I have examined.
The torso looks a bit clunky (it’s not Lynxmotion parts, but a hacked aluminum project case), but my daughter thinks it looks cute. To me it looks like a walking Shuttle PC. Over time it will likely house a Nano-ITX motherboard for robot control and the batteries to keep it running.
My main goal is to focus on development of the servos and to use them to create a very stable walking gait for the robot. If things work out I’ll plan on eventually building a robot with stronger servos and an entire humanoid body.
I’ll periodically update this thread with progress reports. Wish me luck as I’ll need it .
This is the first example I have seen using the duel servo bracket for the legs. I was wondering how this kind of setup would look like. It looks good!
Looking at the photos, I can’t help but think of the early Honda ASIMO prototypes where it started out as a pair of legs mounted to a box like torso. I think, although it may look boxy, it can still look excellent with arms mounted to the torso. The added arms would give a Mech Warrior look to it.
Really great looking project - sure to keep you busy for a long time! Good luck
I came across your work on the OpenServo via Andy Lippitt’s work on his hexapod project - I’m very impressed and keen to try constructing some in the future.
I think that this is exactly what I’ve been waiting to see.
I’m just you’re ordinary undergraduate who’s never taken an electronics class in his life (which will change, come spring ) who just so happened to drop 2 grand on a humanoid biped.
Andy, who’s been a great help to me (and many others) has frequently pointed me towards Open Servo.
But, being a newby roboteer (I’ve never touched a microcontroller in my life!), the betaness (probably not a word) of the Open Servo design scares me.
Proof that it could be done in a real robot was exactly what I had been looking for, and that you picked a biped as your design platform is even more perfect.
I wish you loads of luck!
Let us know if we can help.
My goal is to have it walking by the end of the summer or so. Right now I need to assemble a dozen or more OpenServos and then start working on the various sensors. Some aspects of the robot that I’m planning for now include:
Robot controller is unknown at this point. I’m really eying the VIA 12cm x 12cm NanoITX motherboards to make PC based development much easier. It would fit perfectly. However, it may be overkill and require too much power. In early development the robot will be tethered to a PC that will read the sensors and control the servos.
Batteries will probably represent at least 25% of the robots mass. To leverage this mass I intend to come up with some type of mechanism that the CG of the batteries can be shifted above the right and left leg during the walking gait. This is to minimize the exaggerated hip action often seen in other biped robots as the robots weight is shifted between the legs. This is why it was important for me to get the robots legs very close together and to reduce the moment as much as possible on the two servos that control the ‘spread’ of the legs.
I want at least four sensors on the bottom of each foot – probably force sensing resistors at each corner. This is so the robot can sense the location of the ZMP (zero moment point) relative to the weight bearing foot. Like the servos, the foot sensors boards will likely be based on an AVR MCU and have an I2C interface.
I’ll probably use some of the accelerometer/gyro boards from SparkFun and interface them to an AVR MCU with an I2C interface as well. These sensors will be placed as high on the robot as possible.
My software goal is for all motion of the robot to be closed looped with feedback from the various sensors and the servos themselves. I don’t have a lot of interest in open loop control of a biped robot such as commonly seen in the RoboNova or similar robots. This will make the project much more difficult, but hopefully result in a more sophisticated robot.
.
