One late winter evening, a group of a dozen men and women gathered around a couple tables in the cramped upstairs of a hackerspace. It was the inaugural meeting of the Robot Club, and there had already been some lobbying and promoting of projects. It was generally deemed that the group should work on related projects so as to help each other. The front runner, and first project undertaken was the quadcopter.
This was an ambitious undertaking. We wanted these to be able to carry a GoPro, which meant it would be heavier and more expensive than typical learner flight craft. Though we weren’t experienced enough to know this at the time… After a few mad weeks of research, we had a BOM (Bill Of Materials) of something that might, in theory, fly. Ravi provided some much needed expertise with robotics and RC. John and Matt pulled together the specs, hoping that the calculations for propellers, motors, and battery would be okay. Parts were ordered to build 10, and the waiting began…
But while we waited, we invited people in to demo various robots and help us learn what might be needed. Ali demoed a robot from a UW Engineering class that does maze solving using ultrasonic sensors and SLAM (Simultaneous Localization and Mapping). One of the primary challenges of robotics: dealing with uncertain data and the disparity between reality and the model.
Many thanks to Jim, who put together a CopterVm, and presented a lot of research around which FOSS software to use for flight control.
When the parts finally arrived, we were able to assemble and they were able to fly with surprisingly few tweaks. The first few flights were just up, hover for a while, then land. We didn’t really expect it to go well, so it was still impressive…
After a few test flights, some UFO sightings were reported. ☺
Many tweaks were simply hacks, like soldering the right battery connectors, which didn’t match. There were a lot of neat innovations stemming from the original project! Ryan demonstrated the importance of not flying indoors. Neil designed a shock absorbing landing gear. Jim outfitted his with remotely dimmable lights. Enrico built an emergency parachute.
We flashed SimonK onto the ESCs — taking care as we discovered that some of these were the newer models with inferior response time, so we had to wait for a new version that didn’t cause these to melt and fall from the sky. Kanoa replaced the frame with a more rugged one and mounted both a FPV (remote First Person View) and GoPro.
After the summer, the projects diversifies a little for two reasons: we didn’t want to fly in the cold and dark, and others were coming out with their robots. Chris has been working on a balance bot. David has made some really cool muscle controlled actuators. Shaun hacked a RC car to turn it into a bot. There have been multi leg walkers and several other robots.
The winter project was a tiny robot, which autonomously does challenges. We only got around to doing some line following, though looked into several other sensors too.
Here’s an ESP8266 hooked up to 9 axis sensors (accelerometer, gyroscope, magnetometer) sending the information to ROS.
We’ve learned a lot about a lot of tech. We’ve used Arduino, Teensy, RaspPi, ESP8266, and MultiWii. Sensors include accelerometer, gyroscope, magnetometer, GPS, IR reflectance sensor, ultrasonic rangefinder. We’ve used 3D printers and a laser cutter. Software includes: ArduPilot, Robot OS, OpenSCAD, Vagrant, APM Planner, MAVProxy, mavlink, Linux, OpenCV (Computer Vision). Hope to have more frequent updates going forward!
from SpaceBlogs http://ift.tt/1T5SfUD
via IFTTThttp://ift.tt/1G532rm
No comments:
Post a Comment