Lei Song

Heterogeneous robots capitalize on the diverse strengths of different nodes, enabling them to collaboratively perform tasks unfeasible for homogeneous robots. However, variations in perception capabilities and dynamic models complicate coordination, and potential failures present risks to the safety of the entire coordinated topology. To this end, this article proposes a resilient topology configuration framework for heterogeneous robots with fixed-time coordination. The framework solves the weighted topology configuration with $k$ -node connectivity to tackle the node failure problem and incorporates a $p$ -partite graph to solve the compatibility constraints caused by heterogeneous nodes. Building upon this optimized topology, a fixed-time convergence control barrier function is designed. Furthermore, a constraint-driven optimization framework is proposed to achieve fixed-time coordination of heterogeneous nodes with minimum control cost, thereby achieving synchronous configuration while ignoring initial errors, and being capable of recovering topological resilience after node failure. Comparative simulation results demonstrate the effectiveness of the proposed method.

To fully utilize the knowledge in visible and infrared images and perform pedestrian detection on infrared images, we propose a multi-level multi-source domain adaptive pedestrian detection method. Firstly, we use the Generative Adversarial Network (GAN) method to generate images with the content of the source domain and the style of the target domain as additional domains to promote the model to learn the pedestrian features with multi-domain invariance, thereby effectively alleviating the problem of feature distribution differences between domains. Next, different scales of regional features are extracted to fuse and join, and then image-level domain adaptation and instance-level domain adaptation are performed to the multi-scale part of region features so that the network can learn the pedestrian features more comprehensively. Finally, we obtain the prediction of the target domain images by a certain trained model and make it into pseudo labels to enhance the subsequent model training. Experimental results on KAIST and LLVIP multi-spectral pedestrian datasets show that our proposed method obtains better results than the baseline detector and other existing works, and the missing rate is reduced by 3.0% and 3.37% respectively.

In the collaborative application of multiple autonomous underwater vehicles (AUVs), collision avoidance is essential for ensuring the physical safety of AUVs, while a specific connectivity topology is required for information exchange. This dual requirement presents challenges for controller synthesis. Thus, this study focuses on collision-free connectivity topology switching and maintenance during collaborative missions. For the primary control goal of reaching predefined locations, a backstepping-based controller is designed to ensure system stability. Collision avoidance is achieved through the application of high-order zeroing control barrier functions (ZCBFs), which guarantee the forward invariance of the desired safety set. To facilitate high-bandwidth information exchange, establishing a specific connectivity topology within an effective range is necessary. High-order finite-time convergence control barrier functions (FCBFs) are employed to form new topologies within finite time and then maintain the forward invariance of the desired set. The controller is synthesized using quadratic programming optimization, incorporating collision-free topological constraints and minimally modifying the backstepping-based controller. Numerical simulations and field experimental results demonstrate the effectiveness of the proposed controller, showing that topology switches faster than in existing studies while ensuring collision avoidance under disturbances.

This paper focuses on the ship control system and studies the problem of unknown control directions. Considering that the traditional Nussbaum gain method has to consider the complex situation where the gain converges to both positive and negative infinity when proving the stability of a system, a unilateral Nussbaum function is defined in this paper. By constructing this function, the design and proof process of the adaptive Nussbaum gain method are simplified. Taking the ship course–keeping control system as the research object, a course angle tracking controller is designed by combining neural network, robust adaptive, and sliding mode control techniques. A dual-input RBF single-output neural network is used to approximate the uncertain part of the system, and the robust adaptive control is adopted to deal with the unknown disturbance. The simulation results at the end of the article show that when the direction suddenly switches, the overshoot of the system reaches 40%, and the adjustment time is approximately 3 s. However, the system can still adapt to the change of the control direction and maintain stability, indicating that the method proposed in this paper is reasonable and effective. And the proposed method can effectively cope with the problems of the unknown control direction and its jump, keeping the system stable, which has great theoretical and engineering application value.

This paper focuses on the ship control system and studies the problem of unknown control direction. Considering that the traditional Nussbaum gain method has to consider the complex situation where the gain converges to both positive and negative infinity when proving the system stability, a unilateral Nussbaum function is defined in this paper. By constructing this function, the design and proof process of the adaptive Nussbaum gain method are simplified. Taking the ship course - keeping control system as the research object, a course angle tracking controller is designed by combining neural network, robust adaptive and sliding mode control techniques. A dual - input RBF single - output neural network is used to approximate the uncertain part of the system, and the robust adaptive control is adopted to deal with the unknown disturbance. The simulation results show that the proposed method can effectively cope with the problems of unknown control direction and its jump, keeping the system stable, which has great theoretical and engineering application value.