Warfare missions are conducted in dangerous and unpredictable environments, where human assets are the highest priority. It is desirable to replace the human assets with unmanned vehicles. However, currently the unmanned vehicles require a great deal of human intervention to operate. For global kill-chain operations the amount of interaction must be reduced. One way to do this is to increase the autonomy of the unmanned agents to perform the missions. Imagine a team of ground and aerial vehicles performing together to accomplish a mission in a hostile environment without a threat to our human assets. In order for this scenario to be realized, a robust multi-vehicle architecture that optimizes resources for a wide range of systems and provides accurate navigation state data must be developed.
The joint AFIT/ENG and AFRL/RYRN Intelligent Navigation, Sensing and Cooperative Tasking (INSeCT) project is pursuing this goal. The project seeks to provide autonomous control for multiple vehicles to complete search and rescue, surveillance, or combat operations in dynamic conditions with near-GPS precision in urban and indoor environments. To accomplish this goal, INSeCT includes the development of technologies that provide for both autonomous single and multi-platform control and accurate navigation.
The autonomous control and cooperative tasking system provides the foundation for multiple autonomous vehicles to perform many unmanned, automated tasks. The multi-robot architecture is designed to emphasize individual independence while providing mechanisms for enabling cooperation among autonomous vehicles. The independence allows for large, heterogeneous teams to cooperate effectively on a task, be implemented on multiple systems simultaneously, and be robust to system loss.
The individual independence of the robots is made possible through several systems that combine high level reasoning with the much finer grained sensor and action components. The high level reasoning focuses on the mission tasks and the milestones needed to complete the mission. Particular attention is paid to power management and optimization. By modeling the current goals in a decision process, an overall predictive power management scheme is constructed. This provides power conservation when high fidelity sensing is not required and increases power for future tasks that require these resources. The end result is an autonomous vehicle with increased battery life that can loiter on station longer, remain active in a low-power state, or accomplish additional goals.
The high level plan is dynamically translated into low-level control components called behaviors that accomplish the goals. Because of the dynamic generation, the system can be implemented on any system regardless of the system’s intended tasks or functionality. The behaviors provide real-time reactive response to immediate changes in the environment. As a component of the hybrid control architecture on the agent, the unified behavior framework provides a common interface shared by all behaviors, leaving the higher order planning and sequencing elements free to interchange behaviors during execution to achieve high level goals and plans. Additionally, the modular design of the behavior framework: 1) simplifies development and testing; 2) promotes the reuse of code; 3) supports designs that scale easily into large hierarchies while restricting code complexity; and 4) allows the behavior based system developer the freedom to use the behavior system he or she feels will function the best.
In order to provide an accurate navigation state, the INSeCT project integrates other technological developments from within the Advanced Navigation Technology (ANT) Center. The system builds on the work on vision-aided IMU indoor navigation, making it real time, and integrating this with a Lidar simultaneous localization and mapping system and on-board vehicle odometry. Together with GPS, these integrated navigation technologies provide a means to maintain accurate navigation solutions in urban and indoor environments.
The developed system establishes vehicles that are robust and able to perform well in multiple environments, including high-risk warfare-related areas. The architecture can be implemented on multiple types of autonomous vehicles and they will perform well individually and cooperatively. The individual performance includes both increased survivability due to internal resource management and better performance due to robust behavior sequencing. The cooperative performance includes group scalability, robustness to failure, and efficient task assignment and execution.
