Group 1 (2005)

Student Members:
Nathaniel P. Barcelos
Christopher M. Baron
Adrian J. Gonsalves
Jacob S. Olson

Faculty Advisor:
Professor Robert Helgeland

Project Title:
GPS/INS Subsystem for DARPA Grand Challenge Vehicle

Project Description (provided by group):
The purpose of any navigational system is to provide the user with a location of whatever it is that the user may happen to need. In most cases, what the user needs is location of where they are so that they can avoid getting lost and help find their desired destination.

The main area of focus for our system is to provide the Dartmouth Autonomous Vehicle (DAVe) with the most accurate assessment of where we believe the vehicle is at all times, along with the confidence and current health status of our system with respect to the vehicle. We summarize all this information into the vehicle's state. State consists of latitude, longitude, altitude, vehicle orientation, velocity, and overall system health.

We incorporate different types of technology together in our system. We integrate a Global Positioning Sensor (GPS), an Inertial Navigation System (INS) which is comprised of Accelerometers and Gyroscopic sensors, along with a wheel encoder and binary counter to calculate the state.

Many problems exist with using only one of these types of technology. GPS devices typically provide accurate data for long distances and periods of time while providing inaccurate data for relatively short distances and periods of time. INS devices provide information that gets exponentially worse with respect to time and typically are only useful within the first few minutes after initialization. The wheel encoder outputs the wheel velocity which gives us accurate wheel speed data and provides us with another means of determining position by combining how far the vehicle moved with the direction in which it moved.

Combining all three of these devices enables us to dramatically improve the system accuracy as well as verify the confidence of each device output. If one device is returning faulty data, or no data, the other two devices will notice what is happening and pay little attention to it. We feel that this design gives us the most accurate results possible with the budget and timeline with which we were provided.