Kaiyuan Chen

Primary healthcare (PHC) plays an important role in the healthcare system, yet the shortage of qualified primary healthcare physician (PHP) restricts the quality of diagnostic and treatment services provided by PHC institutions. The rapid development and prevalence of artificial intelligence (AI) offer a solution to this challenge. Existing research primarily focused on two themes: evaluating the performance of AI tools and investigating PHPs’ attitudes and use intention to such AI. There are three main modes used to evaluate the performance of AI tools: evaluation of single AI tool, comparison across different AI tools and comparison between AI tool and physicians. Confusion metrics and examination are two types of indicators primarily used to assess these products. PHPs generally hold positive attitudes and use intentions towards AI tools. According to the technology acceptance model, influential factors of PHPs’ attitudes and use intention can be categorised into usefulness and ease of use. Limitations and future research directions are also discussed in this article, which provides insights for future research and practice to improve AI tools assisting PHPs in diagnostic and treatment tasks and promote AI adoption in PHC.

In this work, we propose a trajectory planning framework for wheeled mobile robot (WMR) navigation in environments with sloped terrain. The framework adopts a hybrid strategy that combines sampling-based and optimization-based methods to plan stable trajectory for the robot on a known elevation map. It adopts a two-layer hierarchical structure to ensure fast and stable convergence of the planning process. In the first layer, a traversability probabilistic roadmap (T-PRM) is constructed to perform a coarse-level feasible route search, which serves to identify traversable regions and determine the task sequence. In the second layer, the planning problem is formulated as a multi-stage optimal control problem (MOCP), taking into account obstacle constraints, terrain constraints, and kinematic constraints. Notably, the coarse trajectory generated in the first layer defines the connectivity of the MOCP, significantly accelerating the convergence of the optimization. Finally, extensive results demonstrate the effectiveness of the proposed framework.

The rapid detection of abnormal body temperatures in urban populations is essential for managing public health risks, especially during outbreaks of infectious diseases. Multi-drone thermal screening systems offer promising solutions for fast, large-scale and nonintrusive human temperature monitoring. However, trajectory planning for multiple drones in complex urban environments poses significant challenges, including collision avoidance, coverage efficiency, and constrained flight environments. In this study, we propose an enhanced Trust Region Sequential Convex Optimization (TR-SCO) algorithm for optimal trajectory planning of multiple drones performing thermal screening tasks. Our improved algorithm integrates a refined convex optimization formulation within a trust region framework, effectively balancing trajectory smoothness, obstacle avoidance, altitude constraints and maximum screening coverage. Simulation results demonstrate that our approach significantly improves trajectory optimality and computational efficiency compared to conventional convex optimization methods. This research provides critical insights and practical contributions toward deploying efficient multi-drone systems for real-time thermal screening in urban areas.

To solve the attitude trajectory tracking problem for hypersonic vehicles in the presence of system constraints and unknown disturbances, this paper designed a nonlinear robust model predictive control (RMPC) scheme, which can produce near-optimal tracking commands. Unlike the existing designs, the proposed scheme is less conservative and successfully prioritizes the solution optimality. The established RMPC follows a dualloop structure. Specifically, in the outer feedback loop, the reference attitude angle profiles are optimally tracked, while in the inner feedback loop, the control moment commands are produced by optimally tracking the desired angular rate trajectories. Besides, an adaptive disturbance observer (ADO) is designed and embedded in the inner and outer RMPC controllers to alleviate the negative effects caused by unknown external disturbances. The recursive feasibility of the optimization process, together with the input-to-state stability of the proposed RMPC, is theoretically guaranteed by introducing a tightened control constraint and terminal region. The derived property reveals that our proposal can steer the tracking error within a small region of convergence. Finally, the effectiveness of the proposed scheme is demonstrated by performing simulation studies.

This article studies a mobile robot trajectory optimization problem where the control path constraints are affected by stochastic parameters and the obstacle information is limited. Based on online measured data, an upper confidence bound (UCB) is designed such that the estimated obstacle regions are guaranteed to cover the real obstacles with a predefined probability, thus avoiding collisions. Using the proposed approach, the control path constraints are transformed into deterministic constraints using analytic approximations. By adjusting the parameters in the proposed approximation function, the feasible set of the approximate optimization problem converges to the real feasible set conservatively. Such a property leads to both feasibility and suboptimality of the solution obtained by solving a deterministic optimization problem. Numerical results demonstrate that the designed UCB and approximation function are effective for trajectory optimization problems with control bounds and obstacle avoidance chance constraints. Comparative studies verified that the approximation function ensures the feasibility of the original chance constrained problem and reduces conservatism.

This article presents a novel online obstacle avoidance trajectory planning method for autonomous ground vehicles (AGVs) based on long short-term memory-attention (LSTM-Attention) networks. The proposed method can guide AGVs to perform emergency maneuvers when encountering sudden and moving obstacles, while also ensuring high levels of real-time performance and optimality. It consists of two parts: 1) offline training and 2) online planning. In the offline training phase, an AGV obstacle avoidance trajectory dataset is generated using numerical trajectory optimization methods to train the LSTM-Attention network. This training allows the network to capture the mapping between the relative information of the vehicle and the obstacles and the optimal control actions. The trained network is then used for online trajectory planning to achieve optimal feedback obstacle avoidance control for AGVs facing sudden obstacles. Furthermore, to address situations involving sudden obstacles in different directions and moving obstacles, a rotation coordinate system method is proposed, significantly expanding the application scenarios of the proposed approach. The effectiveness and real-time performance of the designed method are comprehensively validated through extensive simulation and physical experiments.

In this article, a novel global path planning approach with rapid convergence properties for autonomous ground vehicles (AGVs) named neural sampling rapidly exploring random tree (NS-RRT*) is proposed. This approach has a three-layer structure to obtain a feasible and near-optimal path. The first layer is the data collection stage. Utilizing the target area adaptive rapidly exploring random tree (TAA-RRT*) algorithm to establish a collection of paths considering the initial noise disturbance. To enhance network generalization, an optimal path backward generation (OPBG) strategy is introduced to augment the dataset size. In the second layer, the deep neural network (DNN) is trained to learn the relationships between the states and the sampling strategies. In the third layer, the trained model is used to guide RRT* sampling, and an efficient guarantee mechanism is also designed to ensure the feasibility of the planning task. The proposed algorithm can assist the RRT* algorithm in efficiently obtaining optimal or near-optimal strategies, significantly enhancing search efficiency. Numerical results and experiments are executed to demonstrate the feasibility and efficiency of the proposed method.