Cooperative multi-UAV collision avoidance based on distributed dynamic optimization and causal analysis

Abstract

A critical requirement for unmanned aerial vehicles (UAV) is the collision avoidance (CA) capability to meet safety and flexibility issues in an environment of increasing air traffic densities. This paper proposes two efficient algorithms: conflict detection (CD) algorithm and conflict resolution (CR) algorithm. These two algorithms are the key components of the cooperative multi-UAV CA system. The CD sub-module analyzes the spatial-temporal information of four dimensional (4D) trajectory to detect potential collisions. The CR sub-module calculates the minimum deviation of the planned trajectory by an objective function integrated with track adjustment, distance, and time costs, taking into account the vehicle performance, state and separation constraints. Additionally, we extend the CR sub-module with causal analysis to generate all possible solution states in order to select the optimal strategy for a multi-threat scenario, considering the potential interactions among neighboring UAVs with a global scope of a cluster. Quantitative simulation experiments are conducted to validate the feasibility and scalability of the proposed CA system, as well as to test its efficiency with variable parameters.