"affordable motion sensors may be at hand"

http://www.eetimes.com/news/latest/showArticle.jhtml;jsessionid=GW53B4ZAN3QNAQSNDLSCKHA?articleID=207001827So for UAV booms to become the dot com booms of the future, sensors need to cost under $1. Basically another year, another wave of startups, but no affordable motion sensors. Interestingly, this year they're not making the mistake of attacking gyros like Invensense did. They're focusing on accelerometers, not saying they're accelerometers, and selling the application rather than the measurement.The artificial neural network experts category on Craigslist is still a few decades off.
E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones

Comments

  • The trend seems to be to integrate inertial sensing into GPS receivers. A Tyco receiver I looked at recently had a gyro input, and odometry inputs are typical as well. That's your mass market - portable devices with positional awareness. UAV is not a mass market, though it's certainly an interesting niche.
  • Yes, that's true. I guess what I meant to say is that MEMS gyros generally have more sources of error which you need to compensate for (earth's rotation, linear acceleration effect, generally more sensitive to temperature), and often just have a lower signal-to-noise ratio. It's certainly useful to be able to be able to measure angular velocity directly, but what I was getting at is that by using several accelerometers and some Kalman filtering, you're able to find the angular velocity and orientation by integration - also at a much higher update rate, considering many accelerometers have a bandwidth of around 1kHz.
  • Gyros aren't actually less accurate than accelerometers, they are just different, and are accurate in a different sense of the word. Accelerometers tend to be very good at detecting low frequency movement ( < 10 Hz or so) but are very noisy ( > 10 Hz or so). Gyros are very good at detecting high frequency rotation, but have a (sometimes severe) low frequency drift. If you are smart about it, gyros and accelerometers are both very helpful. The cost thing is certainly an issue.
  • I think the fundamental problem is that MEMS gyroscopes will always be perhaps an order of magnitude more expensive than MEMS accelerometers simply because of the greater complexity involved.

    I've been working on an navigation solution of my own which uses a Kalman filter to get a position estimate from four accelerometers, a 3-axis magnetometer and a GPS receiver. The advantage of this, in my opinion, is that by arranging the four accelerometers (or just 3, if you want) properly and numerically solving a system of differential equations, you have a full 6DoF IMU without using any gyros (which are less accurate than accelerometers anyway).
  • The issue with UAV's is more about $25 gyros and not $1.50 accelerometers. It wouldn't matter if accelerometers were free. Cnet & Dr. Dobbs won't start The UAV Corner until those $30 Picco'Z's in malls are doing autonomous panoramas.
  • 3D Robotics
    Bizarre.

    The problem with cheap accelerometers and gyros is that they're hard to use. What you really want is hardware integration and temperature calibration on the gyros, and hardware filtering on the accelerometers. Once you add that, you're talking hundreds of dollars--those integrated packages are what needs to become cheap.

    Otherwise, we have to do it all in software, and nobody knows better than you what a hassle that is!
  • 3D Robotics
    Accelerometers are $1.50 now. As a practical matter, what's the difference between that and $1.00?
This reply was deleted.