Orientation technology serves as the core foundation supporting for numerous navigation applications in the Internet of Underwater Things (IoUT). However, the unavailability of underwater GPS makes it challenging to mitigate orientation drift in small autonomous underwater vehicle (AUV). Inspired by the visual mechanism of mantis shrimp, underwater polarization navigation technology demonstrates significant potential for long-range and high-precision orientation tasks in AUVs, owing to its unique advantages of electromagnetic interference resistance and absence of cumulative error. However, factors such as bubble scattering, wave disturbances, and overexposure from intense light in complex underwater environments compromise the symmetry of the refractive polarization pattern. This causes the fitted solar meridian to deviate from the ideal reference direction, thereby significantly degrading the heading resolution accuracy and system reliability. To address the aforementioned polarization pattern degradation, this paper proposes an intelligent underwater IoUT orientation method based on polarization pattern restoration. By constructing a multi-step polarization feature reasoning mechanism, physical constraints are embedded into the reconstruction process to restore the polarization distribution characteristics compromised by interference. Furthermore, a line-symmetric polarization enhancement attention module and a solar meridian cosine loss function are designed, achieving an end-to-end solution from polarization information restoration to accurate solar meridian extraction. Additionally, a specialized underwater polarization pattern dataset is constructed for network training and performance evaluation. Simulation and AUV deployment experiments demonstrate that this method effectively suppresses polarization-based orientation errors in real underwater environments, offering a novel solution framework and practical reference for polarization-based orientation technology in complex underwater conditions. The code is available at https://github.com/zbdxdh/UPPRTE_PART.

This paper presents a prescribed time learning-based near optimal enclosing controller to ensure that unmanned aerial vehicles (UAVs) encircle around the specified target with prescribed performance constraints and minimum cost efforts. First, a basic enclosing controller is established to achieve the enclosing error stabilization and stable circumnavigation around a given target. Second, a new prescribed time behavior envelope that eliminates the availability on initial error is proposed. To render the satisfaction of performance constraints and optimal enclosing actions, a transformed enclosing error is obtained by enforcing state conversion on original error. Then, aiming at stabilizing the enclosing error to a prescribed accuracy within a prescribed time, a prescribed time learning-based near optimal enclosing controller under a critic-only adaptive dynamic programming (ADP) is explored, approximating the solution of a novel Hamilton-Jacobi-Bellman (HJB) equation via pursuing the minimum cost associated with transformed errors. Especially, a novel prescribed time learning rule driven by weight errors is elaborated by revisiting real-time and historical information, such that the convergence of weights is only determined by a user-defined time constant. The prominent merit is that the optimal enclosing with performance guarantees can be achieved by a prescribed time ADP. Lyapunov stability demonstrates that involved error variables are ultimately limited and resultant controller satisfies optimality. Finally, simulations verify the values and superiority of the proposed methodology.

This article solves an anti-collision near optimal enclosing control for Uncrewed Aerial Vehicles (UAVs) under a fixed-time learning based adaptive dynamic programming (ADP) context, containing an adaptive feedforward circling item and a collision-free optimal stabilization policy. First, an adaptive feedforward circling item is derived to steer UAVs to approximate and evolve along a desired circle under wind interferences. Specifically, a neural predictor is incorporated that enables a smooth and accurate uncertainty estimate without imposing transient chattering. Next, under a critic-only ADP, a collision-free optimal stabilization policy is developed to tackle static obstacle avoidance and preserve control optimality with the minimum energy cost. To achieve collision elimination, a novel barrier function that describes the hitting risk degree is augmented in the value function. Especially, based on integration of historical data, an innovative fixed-time learning rule driven by weight errors is elaborated, retaining a prescribed convergence free from initial weight selections. The unique innovation includes two aspects, one is that obstacle avoidance and enclosing maintenance are simultaneously guaranteed even with different weight initializations in a unified ADP learning paradigm, another is that energy consumption can be reduced by approximately 18.9% compared with artificial potential field. Involved errors are theoretically demonstrated to be convergent. The usefulness and merits of presented algorithm are accessed by abundant comparisons.

This article presents a concurrent learning-based adaptive critic formation for multirobots under safety constraints, which comprises of an initial formation consensus item and a collision-free adaptive critic policy. First, based on directed graph communication, an initial formation consensus item is designed to maintain the velocity agreement under a leader-follower setting. Particularly, a collision-free adaptive critic policy is developed that enables robots to preserve formation configuration with the minimum cost while excluding collisions caused by inter-robots and static/moving obstacles, wherein safety constraints encoded by an elegantly devised penalty function are enforced by converting constrained optimal control into unconstrained optimal control issue. Furthermore, by revisiting real-time and historical information, a concurrent weight learning rule is elaborated under a critic-only adaptive dynamic programming, improving the weight convergence without demanding the persistence excitation conditions. The remarkable benefits outperforming existing outcomes are safety-critical coordination with energy-saving performances is assured under a computationally efficient optimal learning paradigm. Involved errors are theoretically proved to be convergent. Finally, the values and superiorities are verified through extensive simulations on 2-D and 3-D multirobots.

This paper presents a safety-certified optimal for-mation control scheme for nonlinear multi-agents to realize de-sired formation configuration under safety constraints, guaran-teeing a compromise between safety-critical and energy-saving performances. Firstly, a self-learning optimal formation policy enables agents to achieve optimal formation configuration, wherein optimal performance is guaranteed via a computational-ly-efficient adaptive dynamic programming (ADP) framework. Furthermore, by revisiting real-time and historical information, a novel weight updating rule with fixed-time convergence is elabo-rated, such that rapid weight regulation is realized without de-pending on the initial choices. Secondly, a minimally-invasive safe control policy with high-order control barrier function con-straints is constructed in obstacles-clustered environments, wherein collision risk is excluded by ensuring the forward invari-ance of the safety set. It is strictly proved that closed-loop errors are uniformly ultimately bounded. Finally, extensive simulations are verified the values and superiorities of proposed method.

We developed an intelligent innovative orientation method to improve the accuracy of polarization compasses in harsh conditions: weak skylight polarization patterns resulting from unfavorable weather conditions (e.g., haze, sandstorms) or locally destroyed skylight polarization conditions caused by occlusions (e.g., buildings, trees). First, the skylight polarization status was determined with the degree of linear polarization threshold analysis method and a bionic polarization enhancement sensing model was constructed to simulate the enhanced perception mechanism identified in the Syrphidae visual neural pathway, highly efficient in dark or weakly illuminated environments. The bionic model successfully enhanced the information content extracted from weak polarization patterns. Second, polarization pixel interferences, caused by occlusions under locally destroyed skylight polarization conditions, were removed with a convolutional neural network for image segmentation and the sky area of interest was identified. Finally, the incomplete angle of polarization map derived after image segmentation was fitted using our optimized adaptive antisymmetric ring algorithm. On the basis of the strong angle-of-polarization antisymmetry along the solar meridian, information extracted from the sparse and irregular polarization pixels was analyzed to derive a high-accuracy polarization orientation solution. The whole method intelligently realizes pattern analysis and deep learning intelligent processing, efficiently rotates to manage polarization disorientation. The experimental results demonstrated the performance of the proposed method in compensating for reduced orientation accuracy under degraded polarization conditions, its robustness against perturbations, and its beneficial impact on the environmental adaptability of bionic polarization compasses.

This paper addresses an optimal elliptical enclosing problem of Unmanned Aerial Vehicles (UAVs) under prescribed constraints, whose objective is to steer UAVs to fulfill accurate target encirclement while complying with arriving time restrictions and minimum efforts. A novel learning-based approximate optimal control policy including two-stage designs is presented. At the first stage, a steady-state robust control protocol is developed to steer UAVs to precisely travel along a predefined elliptical path based on concise filtering. At the second stage, to address specified-time constraints and gain the online optimization ability, a single-critic based enhanced learning rule is explored to generate an approximate optimal regulator that stabilizes error dynamics and minimizes value functions, wherein specified-time constraints can be handled by encoding inequality conditions as skilled barrier functions, and by making full use of historical data and current information, a finite-time learning mechanism driven by weight errors rather than Bellman errors is proposed to approximate the solution of Hamilton-Jacobi-Bellman (HJB) equation with faster decaying. The distinct merit is that an improved reinforcement learning (RL) paradigm is formulated to prescribe an elliptical circumnavigation with assured time requirements and optimization behaviors, which can greatly outperform non-RL alternatives in maintaining the optimal performance index while exhibiting restriction handling ability via online learning. Lyapunov stability demonstrates that involved error variables are ultimately limited and resultant controller obeys optimality. The feasibility and values of presented algorithm are accessed by comparisons and simulations.

Autonomous velocity measurement technology based on optical flow plays an important role in applications of the Internet of Things. However, robust velocity measurement results need to provide a robust optical flow field and distance to the ground, especially for complex ground scenes. To solve this problem, this article proposes a visual velocity measurement method based on a$3\times3$camera-array multiaperture biomimetic compound-eye imaging system. The multiaperture optical flow field is first obtained from compound-eye through multiscale analysis and Bayes threshold processing. The real distance to the ground is then obtained by extracting the disparity information of nine apertures for adaptive depth estimation. The velocity measurement error is less than 0.6 m/s in low-altitude (8 and 13 m) flight scenarios with multiple obstacles.

In an actual indoor acoustic environment, the signal processing technique to extract the low-noise and low-reverberation speech signal of a particular speaker from the mixed audio signals is crucial for the back-end speech recognition, speech emotion perception judgment, voiceprint recognition, and other artificial intelligence systems that can be used for IoT connectivity. This paper proposes a method solving the problem of speaker source separation in the presence of both noise and reverberation by using two-step networks for separation, denoising and dereverberation. To ensure the high quality of the input signal used in the dereverberation stage, various separation networks such as Sepfomer were used as training targets with different signals types, and exhibited good convergence in the training set, an obvious separation effect in the validation set, and good generalization. An improved dereverberation method based on time convolution network (TCN-WPE) is proposed. This method uses various improvement strategies such as employing the scale-invariant signal-to-distortion ratio (SISDR) as the network loss function, adopting a transposition mechanism for the input signal, and employing an additional residual mechanism in the network unit, and significantly improves the dereverberation compared with the traditional WPE and DNN-WPE. The Sepformer with the best separation and denoising effect was cascaded with the TCN-transposed-residual. The experiments confirmed that the proposed method can achieve high-quality speaker speech separation and enhancement within a limited corpus, which enables it to be used in IoT-oriented applications such as automatic speech recognition systems.

Accurately obtaining the navigation information of the device is crucial for realizing various emerging Internet of Things (IoT) applications, and a multi-source fusion navigation system is the key to achieving this goal. A distributed integrated inertial navigation system (INS), polarization compass (PC), and geomagnetic compass (MAG) enhanced direction approach is presented to improve the accuracy and robustness of the multisource fusion navigation system in complex environments. To estimate the time-varying measurement noise covariance in a nonlinear multi-source fusion navigation system, the traditional federated Kalman filter (FKF) is improved. In the FKF framework, the third-order spherical radial cubature rule and variational Bayesian theory are introduced, and a variational Bayesian federated cubature Kalman filter (VBFCKF) is proposed. Furthermore, a distributed information monitoring and compensation algorithm based on residuals is developed to address issues like anomalous measured values and asynchronous multi-rate problems. Finally, an experimental platform for unmanned vehicle navigation is designed, and the tests are conducted to confirm the efficacy of the suggested approach. The experimental results show that the system can precisely estimate values based on the measurement quality of sub-filters during navigation. It effectively adjusts measurement noise covariance during updates, thereby mitigating the negative impact of interferences like occlusions and electromagnetic noise on the multi-source fusion navigation system in complex environments. This can strengthen the accuracy and robustness of the navigation system.

Autonomous orientation technology has major engineering significance for intelligent transportation systems (ITS) especially for the intelligent vehicle. The sky polarization characteristics offer a wealth of navigation data. At present, the most advanced polarization navigators can output this navigation information and do not require complex optical structures. However, these current navigation methods are limited severely by the sky conditions, and it is difficult to achieve orientation accurately under interference from reflected light. Here, we report a sky recognition algorithm based on a region prior approach to reduce the influence of the reflected light. Furthermore, to solve the sun fuzzy problem, a morphometric template matching in transform domain (MTMTD) strategy is proposed based on the angle of polarization (AOP). The application scope and practicability of the method are improved effectively by using all the effective pixels as a navigation unit. In addition, this method turns these pixels into curves, which means that even when only one pixel is observed, the sun fuzzy problem can be solved exactly. Our results efficiently verify the feasibility of the proposed strategy, which may provide an interesting solution for heading measurement of intelligent vehicle.

Accurate and reliable navigation data is a key component in Internet of Things (IoT). High-precision and stable autonomous orientation has attracted considerable attention regarding environments in which the global navigation satellite system signal is unreliable. This study proposed a robust orientation method based on the atmospheric polarization mode called weather weighting sparse coding that includes weather classification, sparse coding, and fitting. In comparison with previous studies, the proposed method is effective in restoring the angle-of-polarization image and achieving accurate orientation under complex weather. Specifically, to achieve high-precision orientation in complex weather, a polarization compass was designed, which used the characteristics of the different levels of destruction of atmospheric polarization images under different weather conditions for classification and denoising. The proposed strategy was used to process atmospheric polarization images obtained in complex weather. Experimental results showed that the orientation accuracy was better than 0.31° (root-mean-square error) under conditions of overcast, sandstorms, clear with overexposed areas, and smog with tree-obscured and overexposed areas.