This paper presents a sophisticated vision-aided flocking system for unmanned aerial vehicles (UAVs), which is able to operate in GPS-denied unknown environments to realize exploring and searching missions, and also able to adopt two types of vision sensors: day and thermal cameras, to measure relative motion between UAVs in different illumination conditions without using wireless communication. In order to realize robust vision-aided flocking, an integrated framework of tracking-learning-detection on the basis of multi-feature coded correlation filter has been developed. To achieve the long term tracking, a re-detector is trained on-line to adaptively re-initialize target for global sensing. An advanced flocking strategy is developed to address the autonomous multi-UAVs cooperative flight. LiDAR-based navigation modules are developed for autonomous localization, mapping and obstacle avoidance. Flight experiments of a team of UAVs have been conducted to verify the performance of this flocking system in GPS-denied environment. The extensive experiments validate the robustness of proposed vision algorithms in challenging scenarios.

A neural network-based distributed adaptive approach combined with sliding mode technique is proposed for vehicle-following platoons in the presence of input saturation, unknown unmodeled nonlinear dynamics, and external disturbances. A simple and straightforward strategy by adjusting only a single parameter is proposed to compensate for the effect of input saturation. Two spacing polices (i.e., traditional constant time headway policy and modified constant time headway policy) are used to guarantee string stability and maintain the desired spacing. Chebyshev neural networks (CNN) are used to approximate the unknown nonlinear functions in the followers online, and the implementation of the basic functions of CNN depends only on the leader's velocity and acceleration. Furthermore, unlike existing approaches, the nonlinearities of consecutive vehicles need not satisfy the matching condition. Finally, simulations are carried out to illustrate the effectiveness and the advantage of the proposed methods, first using a numerical example, followed by a practical example of a high speed train platoon.

Purpose: Cooperative control of a group of unmanned aerial vehicles (UAVs) is an important area of research. The purpose of this paper is to explore multi-UAV control in the framework of providing surveillance of areas of interest with automatic loss detection and replacement capabilities. Design/methodology/approach: The research is based on the concept of the multi-agent system. The authors present the framework of the multi-agent and protocol design for monitoring the network of a group of UAVs. Findings: If one or more UAVs which is conducting a high priority surveillance task is lost, the system can self-arrange for another UAV to replace the lost UAV and continue to execute its task. This research provides an excellent design protocol for UAV loss detection and replacement scheme. Research limitations/implications: One of the major limitations of this research is that we have only two types of priority levels, high or low. If the priority is more than two levels, for example, high priority 1, high priority 2, or high priority 3, the replacement has not yet been implemented. Originality/value: This paper contributes to the following two aspects of the scientific knowledge. The first contribution is the design of an agent model which jointly considers system architecture, communication, control logic and target monitoring. The second contribution includes the decentralized and automatic UAV loss detection and replacement algorithm.

It is well-known that collision-free control is a crucial issue in the path planning of unmanned aerial vehicles (UAVs). In this paper, we explore the collision avoidance scheme in a multi-UAV system. The research is based on the concept of multi-UAV cooperation combined with information fusion. Utilizing the fused information, the velocity obstacle method is adopted to design a decentralized collision avoidance algorithm. Four case studies are presented for the demonstration of the effectiveness of the proposed method. The first two case studies are to verify if UAVs can avoid a static circular or polygonal shape obstacle. The third case is to verify if a UAV can handle a temporary communication failure. The fourth case is to verify if UAVs can avoid other moving UAVs and static obstacles. Finally, hardware-in-the-loop test is given to further illustrate the effectiveness of the proposed method.