Base Defense Using a Task Assignment Framework

Abstract

As the capability of unmanned systems increases and the availability of human resources decreases, the automation of manpower-intense tasks becomes more attractive. Base defense tasks such as perimeter surveillance, alarm response, and area security can be performed by a team of automated agents, allowing humans to become directors of the surveillance system rather than simply sensors in that system. This paper poses a base defense scenario and automates certain aspects of it using a cooperative task assignment framework suitable for teams of Unmanned Air Vehicles (UAVs). Specically, we focus on area and perimeter surveillance and alarm servicing. Our method allows for a team of heterogenous UAVs to operate in environments with no-y zones. A main feature of our algorithm is its ability to be run in real-time which allows solutions to be rapidly computed as air space and tasks dynamically change. We present results from a hardware experiment involving three UAVs. Present day operational UAV systems most often use multiple operators per UAV. Even man-packable UAV systems and their Ground Control Stations are primarily set-up for single-ship operation where one operator controls and digests information from one vehicle. These systems have been eective because there is a large need for Intelligence Surveillance and Reconnaissance (ISR) tasks that can be accomplished by single-ship UAV operation and a willingness to use personnel for these tasks. However, with increasing desire to reduce the operator workload and extract more information from the huge data input stream, and to address problems where cooperative, time-coordinated tasks are key; then the inadequacy of present-day operations becomes apparent. Our work is focused on the denition of a system and tools that will change this operator-to-UAV ratio to one where one operator can manage the actions of multiple UAVs. The base defense problem is one where there is a need to have unmanned systems cooperate and coordinate their actions with each other and with manned systems. In the base defense problem there is a need allow the operator to spend more eort on how to use the information being gathered and less on the details of how to operate the vehicles. The purpose of this article is to detail how certain aspects of the base defense problem can be automated using a team of UAVs via a cooperative task assignment algorithm. Section II contains a general description of the problem and an outline of our approach to solve it. Section III describes the algorithm used for nding the optimal task assignment and associated UAV routes in a cluttered environment. An example scenario and simulation results are shown in Section IV. Finally, experimental results are described in V and our conclusions are given in Section VI.