A localization technique used for teams of unmanned vehicles (UVs) is presented. The proposed technique is meant to be scalable with respect of the size of the team, while allowing estimating position even in the presence of errors in the available information provided by one or more UVs comprising the team. The UVs are considered to be susceptible to errors in their sensors and collected data. The position estimation is achieved by using observations of local environment and signal strength information received from teammates broadcasting an estimate of their location using radio signals. This information, along with the model of the radiation pattern of the radio transmission from the team members help any UV within the team estimate its position, with precision in a decentralized manner.

In any UV application localization systems are important to enable UVs to operate autonomously. An example of such applications is the exploration of unknown unstructured environments, for example a disaster zone after an earthquake or an unexplored planet (where a GPS network is not available). Due to the harsh environment encountered within these scenarios, errors or variations in sensor data are present and unavoidable at all times. Obtaining a priori information of the environment or building a radio map is not an option in such situations. Thus, a localization technique that can operate effectively even in the presence of errors is highly desirable and presented herein.