Cross-domain cooperative task allocation is a complex and challenging issue in the field of multi-agent task allocation that requires urgent attention. This paper proposes a task allocation method based on the multi-strategy improved dung beetle optimization (MSIDBO) algorithm, aiming to solve the problem of fully distributed multi-agent cross-domain cooperative task allocation. This method integrates two key objective functions: target allocation and control allocation. We propose a target allocation model based on the optimal comprehensive efficiency, cluster load balancing, and economic benefit maximization, and a control allocation model leveraging the radar detection ability and control data link connectivity. To address the limitations of the original dung beetle optimization algorithm in solving such problems, four revolutionary strategies are introduced to improve its performance. The simulation results demonstrate that our proposed task allocation algorithm significantly improves the cross-domain collaboration efficiency and meets the real-time requirements for multi-agent task allocation on various scales. Specifically, our optimization performance was, on average, 32.5% higher compared to classical algorithms like the particle swarm optimization algorithm and the dung beetle optimization algorithm and its improved forms. Overall, our proposed scheme enhances system effectiveness and robustness while providing an innovative and practical solution for complex task allocation problems.

This paper addresses the lateral motion control of a supercavitating vehicle and studies its ability to maneuver. According to the unique hydrodynamic characteristics of the supercavitating vehicle, highly coupled nonlinear 6-degree-of-freedom (DOF) dynamic and kinematic models are constructed considering time-delay effects. A control scheme utilizing radial basis function (RBF) neural-network-(NN)-based adaptive sliding with planing force avoidance is proposed to simultaneously control the longitudinal stability and lateral motion of the supercavitating vehicle in the presence of external ocean-induced disturbances. The online estimation of nonlinear disturbances is conducted in real time by the designed NN and compensated for the dynamic control laws. The adaptive laws of the NN weights and control parameters are introduced to improve the performance of the NN. The least squares method is utilized to solve the actuator control efforts with rolling restriction in real-time online. Rigorous theoretical proofs based on the Lyapunov theory prove the globally asymptotic stability of the proposed controller. Finally, numerical simulations were performed to obtain maximum maneuverability and verify the effectiveness and robustness of the proposed control scheme.

Radar error is a significant factor restricting the detection accuracy of radar. Improving the detection accuracy of radar is crucial to accurate target localization in an ocean environment. With a large amount of measured data, this paper presents the correlation analysis of radar errors, and gives the assumptions needed for radar error modeling. This provides theoretical support for calibrating the modeling method based on the radar error in the simultaneous observation of double targets. First, three error correction models were developed with constant, first-order, and second-order fitting methods, based on the origin and transfer of error in the radar system and the principles of error elimination and reduction in a cooperative platform. Subsequently, the advantages of using a cooperative platform including information sharing were used to detail the collection and preliminary processing of measured radar data as well as the establishment of datasets. Ultimately, the measured radar data were analyzed considering the random error of radar, the strategy for the selection of the optimal error correction model, and the detection error correlation of double targets. In particular, the second-order fitting model with the strongest azimuthal systematic error correlation calculated the root mean square error, which was 32.07% higher compared to the other two models. As revealed in the results of the error analysis, the random error of radar does not agree with the vector superposition characteristic, and it is strongly correlated with the system error of a single radar system that observes different targets within a short period. This provides a way for resolving some problems such as accurate target localization and time variance of error in the complex ocean environment.

Supercavitating vehicles have received significant attention in military applications due to their high underwater speed. This paper aims to establish analytic models of the encounter probability between a supercavitating vehicle and its target in two modes: straight-running mode (SRM) and turning-straight-running mode (TSRM). Mathematical models for the supercavitating vehicle to encounter the target in SRM and TSRM are proposed. An improved particle swarm optimization (PSO) algorithm based on local best topology with a penalty function is introduced to plan the TSRM path of the supercavitating vehicle. The particle swarm position is updated by using Cauchy and Gaussian distributions. The judgement indexes of the encounter probability analytic models in SRM and TSRM are determined through statistical analysis. Based on the law of error propagation, the analytic models of judgement index variances in the two modes are derived by using the implicit function differential method, and the integral intervals of their probability density functions are obtained according to the relative motion between the supercavitating vehicle and the target. Monte Carlo simulations and hypothesis testing are utilized to verify the accuracy and feasibility of the encounter probability analytic models in SRM and TSRM.

Next-generation intelligent consumer electronics, including autonomous recreational vessels and unmanned surface vehicles, require reliable situational awareness during abrupt maneuvers, sensor faults, and disturbance-like interference. However, traditional centralized tracking pipelines suffer from communication bottlenecks and have limited capabilities for proactive anomaly awareness. This study proposed a threat-hunting-oriented tracking framework for consumer maritime electronics based on a collaborative edge inference paradigm and a Variational Bayesian Adaptive Interacting Multiple Model (VBTA-IMM) engine for local inference. Proactive anomaly indication was achieved through a variational Bayesian update of the transition probability matrix. Mahalanobis-distance-based prediction errors rapidly down-weighted mismatched motion models and accelerated mode switching. In addition, a quasi-uniform B-spline fitting strategy was introduced to adaptively identify the turn rate and refine the constant-turn model set for unknown maneuvers. A square-root cubature information filter was further integrated into the IMM framework to enhance numerical stability and multi-sensor fusion on resource-constrained consumer devices. Simulation and field tests showed that the proposed method improved position and velocity accuracy by 28.64% and 44.28%, respectively, compared with conventional IMM variants. These results demonstrate its effectiveness for proactive tracking and anomaly detection in networked consumer systems. Nevertheless, broader real-world embedded deployments still require further validation under stricter onboard computing and communication constraints and more heterogeneous sensing conditions.