9DOF Inertial Measurement Unit

9DOF Inertial Measurement Unit

OLogic designed this 9-degree-of-freedom (9DOF) inertial measurement unit (IMU) specifically for use in ground-vehicle applications. We have found that most robotics researchers, have been using aircraft grade IMU’s for augmenting GPS, and odometery, in ground-vehicle position estimation tasks. OLogic initially designed this unit for use in Odyssey, the Robomagellan robot, and it has slowly evolved to be a more complete position estimation engine.

Most commercial IMU units are just a sensor package. Doing something useful with the data, is difficult, so this newest generation unit has a complete computer onboard, to do processing and position estimation.

  • 3 Rotational Degrees of Freedom (Gyro)
  • 3 Linear Degrees of Freedom (Accelerometer)
  • 3 Magneto Degrees of Freedom (Magnetometer)
  • (Optional) 2-Axis compass port for differential compassing
  • +/-2G measurement
  • +/-90deg/sec rotational motion
  • 100hz update rate
  • GPS port for onboard integration
  • CAN-bus output
  • RS-232 output
  • Sharp ARM7 Processor (~50 MIPS)
  • 2meg Flash Memory
  • 512k SRAM
  • Real-Time Clock
  • (6-12v) supply
  • Low power (200ma Peak)
  • Low profile footprint – 2.5″W x 4″L x 2″H
  • Ultra-low-cost component design, using consumer-grade parts


IMU Research
Most inertial measurement units are built for aircraft, missile guidance, or spacecraft applications. As a result, the specifications, and tolerances are quite narrow, and the prices of the units can be in excess of $10k. For ground vehicle applications, this is completely unnecessary.

OLogic initially designed the 6 degree of freedom unit pictured to the right, for use in the Robomagellan robot Odyssey. It needed to be constructed from inexpensive parts, yet still work reasonably well for GPS augmentation, and position estimation tasks. What we learned from this exercise was that just a good sensor package isn’t enough. Basic GPS augmentation is pretty easy to do, but doing higher-level position estimation, requires more CPU horsepower. A dedicated processor is the best way to go, thus which lead us to the 2nd generation device featured above.