Nebula - n-Prize competition entrant No. 1

You are here: Home > Design

Design
Navigation Electronics

Home	>	About
	>	Mission Profile
	>	Simulations
	>	Design Navigation Telemetry Abort & Recovery Engine Fuel & Oxidiser Satellites
	>	Production
	>	Testing
	>	Crew
	>	Launch Data
	>	Contact

Links

XE1610-OEMPVT Data Sheet

ADIS 16100 Data Sheet

LIS3LV02DL Data Sheet

Kalman Filtering

Navigation Package

The navigation package comprises three units that taken together are called the IMU (Inertial Management Unit - in deference to the Space Shuttle). However, before discussing them in detail we might consider why we need an IMU at all.

There are three main reasons for including an IMU in the design. Firstly for safety - if the rocket orientation or position is sufficiently off the predicted values, then an abort procedure can be initiated to safeguard people on the ground.

Secondly, for recovery. Part of the design of this rocket system includes a mechanism for automated recovery to a specific location necessitating the use of a reasonably accurate navigation system.

Finally, an IMU is needed for correct orbit insertion. Simulations have shown that the window of error (in terms of position and velocity vectors) for an accurate and predictable orbital trajectory is too tight to rely on non-guided means. The mission profile assumes that we are able to predict the position of the rocket/satellite combination in orbit for at least part of the mission.

Mission Patch

Unit 1 - Global Positioning System (GPS)

You may say "OK, GPS - that's all you'll need", but you'd be wrong. Non-military GPS won't work at the kind of altitudes we're aiming for, and even if it did it wouldn't work at the speeds we're trying to reach. The best data I have is that commercial GPS will work at up to 515 m/sec - nothing like the 7800 m/sec necessary for orbit.

However, GPS is valuable for a number of reasons. It allows the errors in the inertial platform (next section) to be characterised while in-flight so that its output is much more believable when GPS isn't available. Its also very useful in the terminal guidance recovery phase.

The GPS unit that we've plumped for is the XE1610-OEMPVT low-cost OEM GPS receiver.

Unit 2 - Inertial Navigation (IN) Platform

Six degrees of freedom need to be catered for: 3 axes of rotation, plus 3 axes of translation. The rotation is taken care of by 3 orthogonally mounted solid-state (MEMS) gyroscopes. There are many types to choose from these days, but I've opted for Analogue Devices' ADIS 16100 because of its good price/performance and drift characteristics.

Velocity vectors are obtained by integrating the outputs from 2 3-axis accelerometers. The devices in use are ST's LIS3LV02DL.

Unit 3 - Navigation Integration Filter (NIF)

Now for the important bit. Without a good means of integrating the navigation data from the GPS and IN platforms, we needn't bother with any of it.

Our system implements a 9-state Kalman filter. Using Bayesian mathematics, the kalman filter takes position fixes from the GPS platform and uses them to update confidence values (the covariance matrix) for the IN platform output. In this way, both position and velocity values (the state vector) can be maintained with a high degree of confidence while GPS is updating, but when it is no longer available the drifts in the IN platform should still be within known error bounds.

Because of cost considerations, the NIF is implemented on the same processor as the rest of the rocket control and telemetry software. The device being used in this role is Infinion's XC886 8-bit microcontroller.