The authors propose a complete software and hardware framework for a novel spherical robot to cope with exploration in harsh and unknown environments. The proposed robot is driven by a heavy pendulum covered by a fully enclosed spherical shell, which is strongly protected, amphibious, anti‐overturn and has a long‐battery‐life. Algorithms for location and perception, planning and motion control are comprehensively designed. On the one hand, the authors fully consider the kinematic model of a spherical robot, propose a positioning algorithm that fuses data from inertial measurement units, motor encoder and Global Navigation Satellite System, improve global path planning algorithm based on Hybrid A* and design an instruction planning controller based on model predictive control (MPC). On the other hand, the dynamic model is built, linear MPC and robust servo linear quadratic regulator algorithm is improved, and a speed controller and a direction controller are designed. In addition, based on the pose and motion characteristics of a spherical robot, a visual obstacle perception algorithm and an electronic image stabilisation algorithm are designed. Finally, the authors build physical systems to verify the effectiveness of the above algorithms through experiments.

Motivated by the excellent performance of proportional–integral–derivative controllers (PIDs) in the field of control, the authors injected the philosophy of PID into optimisation and introduced two types of novel PID optimisers from a continuous‐time view, which benefit from the idea that discrete‐time optimisation algorithm can be modelled as a continuous dynamical system/controlled system. For centralised optimisation, the authors discuss the idea of the first‐order PID optimiser and the second‐order accelerated PID optimiser. Furthermore, this framework is extended into distributed optimisation settings, and a distributed PID optimiser is proposed. Finally, some numerical examples are given to verify our ideas.

Solder joint quality inspection is a crucial step in the qualification inspection of printed circuit board (PCB) components, and efficient and accurate inspection methods will greatly improve its production efficiency. In this paper, we propose a PCB solder joint quality detection algorithm based on a lightweight classification network. First, the Select Joint segmentation method was used to obtain the solder joint information, and colour space conversion was used to locate the solder joint. The mask method, contour detection, and box line method were combined to complete the extraction of solder joint information. Then, by combining the respective characteristics of convolutional neural network and Transformer and introducing Cross‐covariance attention to reduce the computational complexity and resource consumption of the model and evenly distribute the global view mutual information in the whole training process, a new lightweight network model MobileXT is proposed to complete defect classification. Only 16.4% of the Vision Transformer computing resources used in this model can achieve an average accuracy improvement of 31%. Additionally, the network is trained and validated using a dataset of 1804 solder joint images constructed from 93 PCB images and two external datasets to evaluate MobileXT performance. The proposed method achieves more efficient localization of the solder joint information and more accurate classification of weld joint defects, and the lightweight model design is more appropriate for industrial edge device deployments.

A deep reinforcement learning (DRL) method based on the deep deterministic policy gradient (DDPG) algorithm is proposed to address the problems of a mismatch between the needed training samples and the actual training samples during the training of intelligence, the overestimation and underestimation of the existence of Q‐values, and the insufficient dynamism of the intelligence policy exploration. This method introduces the Actor‐Critic Off‐Policy Correction (AC‐Off‐POC) reinforcement learning framework and an improved double Q‐value learning method, which enables the value function network in the target task to provide a more accurate evaluation of the policy network and converge to the optimal policy more quickly and stably to obtain higher value returns. The method is applied to multiple MuJoCo tasks on the Open AI Gym simulation platform. The experimental results show that it is better than the DDPG algorithm based solely on the different policy correction framework (AC‐Off‐POC) and the conventional DRL algorithm. The value of returns and stability of the double‐Q‐network off‐policy correction algorithm for the deep deterministic policy gradient (DCAOP‐DDPG) proposed by the authors are significantly higher than those of other DRL algorithms.

Nuclear facilities have a regulatory requirement to measure radiation levels within Post Operational Clean Out (POCO) around nuclear facilities each year, resulting in a trend towards robotic deployments to gain an improved understanding during nuclear decommissioning phases. The UK Nuclear Decommissioning Authority supports the view that human‐in‐the‐loop (HITL) robotic deployments are a solution to improve procedures and reduce risks within radiation characterisation of nuclear sites. The authors present a novel implementation of a Cyber‐Physical System (CPS) deployed in an analogue nuclear environment, comprised of a multi‐robot (MR) team coordinated by a HITL operator through a digital twin interface. The development of the CPS created efficient partnerships across systems including robots, digital systems and human. This was presented as a multi‐staged mission within an inspection scenario for the heterogeneous Symbiotic Multi‐Robot Fleet (SMuRF). Symbiotic interactions were achieved across the SMuRF where robots utilised automated collaborative governance to work together, where a single robot would face challenges in full characterisation of radiation. Key contributions include the demonstration of symbiotic autonomy and query‐based learning of an autonomous mission supporting scalable autonomy and autonomy as a service. The coordination of the CPS was a success and displayed further challenges and improvements related to future MR fleets.