For those interested in the theory that goes into the practice, here is an article discussing the limits on [flying] maneuverability in cluttered envrionments. Inspired by abilities of birds, but ‘simple’ enough to soon be available for use in UAVs.
This is last years information, but I ran accros a reference, and do not remember seeing anything about it here. 40 minute video. Would be fun to have one of those PR2 robots to play with!
The talk mentioned being able to use ROS with Arduino, so I looking up this information:
Not enough to be able to build a complete bot running ROS on an Arduino, but enough to make it easy to use Arduino as part of a bigger system using ROS. A much more complete implementation will run on Andriod [in Java].
This 16:47 video includes many clips of quadracopters that have been posted here before, being talked about by a researcher from University of Pennsylvania researcher. The final ‘music video’ clip, I have not seen before.
Thanks to Gina for pointing this Ted Talk out to me.
I’ve been following the online AI class from Stanford. Included in that was a video with descriptive commentary of the robot that won the DARPA challenge. The video is also on youtube. The ‘trick’ was to overlay very accurate but short range laser sensor data, with much less accurate but long range camera vision, to ‘infer’ what was further ahead than it could see with the lasers.
I found a news article about a GPS receiver that is smaller than a penny, and weighs only 0.3 grams. Even with support electronics for data logging, remote downloading, scheduling, and battery, it still only comes to 10 grams. This may not be quite small enough to attach to insects, but they are designed to be used to track the foraging habits of bats.
I could not see any pricing on the manufacturer ? website, but that is small enough that weight and power requirements should no longer be a limitation for adding position sensing to almost any robotics project.
Definitely something to check into for our rocketry members. A 10 gram payload weight should be doable. If added to an existing sensors / recording package, it could add as little as 0.3 grams.
Here is a link to an article from Technology Review about GPS style positioning technology that use small [it says the size of a hardback book] land based units. Accuracy down to a few centimeters. This version uses the same frequency [band] as wi-fi, but apparently with higher power. A single unit could cover several kilometers. This is real technology already being used.
One of the company co-founders is quoted as saying
Tracking goods and machines with high accuracy can enable greater use of robotics and automation.
Assuming a “location hotspot” is available, or can be setup at reasonable cost, this looks like an opportunity for robotics location awareness. Depending on thoses costs, down to the consumer level. The existing home robots that need to have hardware “electronic walls” installed could potentially use position sensing and a virtual wall instead.
- a sumobot that knows its position relative to the ‘ring’ well enough that it does not need to completely rely on optical sensing of the white edges?
- a line follower that memorizes the path based on absolute position information? The high speed line followers could ‘plan’ acceleration, deceleration, and turns based on their position
- a Darpa style challenge where the ‘contestant’ is given only the geo-location of the goal