Terrestrial Attitude Estimation for the Formation Control Testbed (FCT)

In this paper we look at the problem of terrestrial (earth based) attitude estimation of a unique robotic vehicle using full three axis attitude measurements and three axis inertial rate sensors (gyros). The vehicle is completely autonomous and uses air bearings to simulate the drag free dynamic environment of space. An onboard infrared camera system is used to provide quaternion measurements representing the attitude of the robot relative to the room frame of the test facility. Fiber optic gyros are used to sense the inertial angular rates. To simulate the performance of the system, a stochastic model of the gyros was developed based on long term rate table data. The angle random walk, bias, and bias stability were determined to agree with the data provided in the manufactures specification sheet. We show that a 3times reduction in the standard deviation of the attitude estimates can be achieved by proper mixing of the two sensor measurements. The attitude estimation algorithm used in this paper also provides bias free estimates of the angular rate which can be used for control or other purposes. These results are established in both high fidelity simulations and experimentally using data taken during real time operation of the robot.