Vision-Aided, Cooperative Navigation for Multiple Unmanned Vehicles

In the recent past, a shift has taken place from manned to unmanned Intelligence, Surveillance, and Reconnaissance (ISR) missions. This shift has lead to an increase in the number of unmanned vehicles (UV) operating in a theater. Additionally, removal of the crew allows for a reduction in the scale of the vehicles, which in turn leads to an increased ability to operate in GPS degraded environments. These environments include but are not limited to urban canyons, caves, and building interiors. With the loss of GPS signal the vehicles must rely on Inertial Navigation Systems (INS) which when reduced to an appropriate size are inherently inaccurate. As the reliance on these vehicles increases so does the desire for ever more accurate navigation and targeting. In this paper, we endeavor to exploit three attributes of the increase in UV use for ISR missions. The first attribute is the increased numbers of UVs in a theater, while the remaining two attributes of on-board vision systems and wireless communication are inherent in the ISR mission. This paper's focus is the development and validation of a cooperative navigation system based on the measurement of UV position relative to shared landmark position estimates. Each UV in the network locates landmarks using it's on-board vision system and transmits the data to all other system UVs. After receiving data from the other UVs, the system uses the landmarks as additional measurements. The system is evaluated using Matlab simulation. Three simulations are created to perform the evaluation. The first provides individual position and attitude estimates without interaction between the UVs. The second provides the same estimates using the cooperative system without ranging information between UVs while the third includes ranging. The results of the simulations are used to demonstrate the improvements provided by the utilization of shared landmarks.