The 2005 DARPA Grand Challenge was a driverless car competition sponsored by DARPA in October 2005. The challenge awarded a $2 million prize to the winning team. I was a pivotal member of the team from the University of Florida, Team CIMAR. Our vehicle was named the NaviGATOR and was purpose-built for the off-road desert challenge.
On the team I was involved in every major area of the team, the vehicle and system software. Due to rules requirements, the team was forced to develop, from scratch, a software SDK based on the Joint Architecture for Unmanned Systems (JAUS). I was very involved in the design and implementation of this SDK for the software development team. This codebase later went on to form the basis for the original OpenJAUS product.
Another key responsibility on the team was to develop the software API for a common sensor data output format based on JAUS. This lead to the Smart Sensor architecture deployed on all sensors deployed on the NaviGATOR system. Similar to the SmartMotor concept, a Smart Sensor encapsulates a sensor technology within a computing resource and unifies the output of sensors into a grid of traversability values. Within the NaviGATOR architecture, a centralized arbitration process took discrete sensor grid inputs and generated a single grid output which was used by the planning and control elements to produce vehicle movement.
My background in World Modeling and geospatial databases was put to use on two software processes known as “pseudo-sensors”. A Psuedo-Sensor takes a priori data (such as stored in a geospatial database) and paints that data into a traversability grid during runtime. As such, information which may influence the planning and control process (such as desired paths, boundaries, known obstacles) can be sent through the same interface and arbitration process as other sensor technologies.
The 2005 DARPA Grand Challenge was, at the time, the premier exposition of unmanned ground vehicle technologies and a great opportunity to discover real-world solutions to realistic problems in challenging environment. One of the best professional experiences in my history and a very exciting project!
