Design of a Guidance Controller Using Network Topology

Abstract

In dynamic environments, the control of a formation of unmanned vehicles remains a significant control problem on both network and guidance dynamics levels. Currently, formations of unmanned vehicles require multiple ground operators to supervise the movement of a formation to effectively execute a task or mission objective. By designating a single vehicle as the lead agent of the formation, control over the entire formation reduces to a distribution of information through a single vehicle, thereby reducing the number of required ground operators. In such architectures, the location of the leader and the manner in which the control is distributed amongst the decentralized formation must suit the desired behavior requested by the ground operator. This work centers on the implementation of a single-leader formation of unmanned vehicles following a desired trajectory. Incorporating the desired formation behavior provided by a ground operator, the locally implemented distributed controllers satisfy a previously established global performance index, leveraging only local information exchange in the decentralized formation.

Consider a fleet of unmanned vehicles moving through an environment under the direction of a single lead agent communicating with a ground operator. The topology of the communication structure is unknown to the operator, knowing only that it communicates with a single vehicle in the formation and provides a desired behavior for the formation as it moves along a series of predetermined waypoints. The formation establishes its leader based on the desired response from the ground operator and the knowledge of the formation structure, as determined exclusively through nearest-neighbor communication.

The contributions of this work are twofold. First, a novel approach to the decentralized selection of a formation leader establishes a leader best suited to follow a desired trajectory while adhering to the behavioral constraints requested by the ground operator. Second, an optimal control approach to a distributed controller design incorporates trajectory tracking and formation keeping through the underlying communication topology of the unmanned formation. This distributed design provides the leader with feedback from its followers, thereby reducing control usage by the followers to retain the formation structure and allows the leader to track the desired trajectory from the ground operator.