As a significant tool to lead a new round of technological revolution and industrial transformation, the humanoid robots equipped with humanoid dexterous hands having the capability of human-like grasping are expected to assist or even replace humans in performing delicate and complex tasks in various application scenarios, such as industrial production, household services and counter-terrorism. The stable grasping methods for humanoid dexterous hands in different application scenarios are comprehensively reviewed in this paper. The implementation forms of stable grasping configurations for humanoid dexterous hands with different structures are outlined, the methods for analyzing and estimating the contact states based on exteroceptive sensing / proprioceptive sensing for humanoid dexterous hands are summarized, and the techniques for detecting and recognizing the grasping states of humanoid dexterous hands when engaging with objects in global relative motion / local relative motion are categorized. And then the commonly used control strategies are presented for stable grasping with humanoid dexterous hands. The strengths and limitations of the aforementioned methods in the application of humanoid dexterous hands during contact and grasping processes are We analyzed, the challenges in the research of stable grasping methods posed by factors such as soft materials, dynamic environments are discussed, and an outlook on research trends and their applications in humanoid robots is provided.

For the dynamic control challenges caused by the coupling effect of multiple power sources and high nonlinearity in distributed electric drive vehicle, a model predictive torque coordination control strategy based on particle swarm optimization and ant colony optimization is proposed, which uses a 7-degree-of-freedom vehicle dynamics model as the prediction model. The simulation and actual vehicle test platforms were built, and the multiple operating conditions were test. The test results show that the proposed torque coordination control strategy can be used to adjust the control mode according to the experimental conditions, thus achieving a comprehensive optimal control effect of power, economy, and handling stability.

The design and control of continuous variation capability of robot bionic joint stiffness are the difficult problems in flexible drive technology. In this paper, a bionic joint with continuously variable stiffness is proposed, in which the continuous change from low stiffness to high stiffness is achieved by using a variable moment arm. The main and secondary motors of the bionic joint have large power ratio and compact structure. In order to analyze the relationship between the elastic deformation of bionic joint and the stiffness variation of load, a static model with load conditions is established, and the relationship between the joint stiffness and the angle of main and secondary motors is obtained through simulation. The experiments of torque tracking, step response and stiffness tracking are conducted on the bionic joint. The experimental results show that the bionic joint with continuously variable stiffness has a maximum error of 1.23 N·m with a variance of 0.19 N·m at a sinusoidal output torque of 16 N·m, with the maximum error being 7.7% of the total amplitude, and when the stiffness tracking is in the range of 0.4 N·m/(°) to 1.6 N·m/(°) for and the output torsion angle is 5°, the maximum error is 0.09 N·m/(°), accounting for 9.0% of the total mean stiffness.

Envelope control originated in the aerospace industry, which provides the safety guarantees and movement limit and brings a better performance in aircraft control. A vehicle dynamics controller, which pushes the vehicle in handling limit, is proposed based on 8×8 distributed driving unmanned ground vehicle and the core idea of envelope method. Firstly, a novel method is proposed to evaluate the driving force status of vehicle by establishing a tire slip circle. The tire slip status and “g-g” diagram are combined to achieve approaching the vehicle handling limit under autonomous driving, and to perform the vehicle's mobility and maneuverability. On the other hand, the controller is used to insure the tracking accuracy during trajectory tracking which can complete tasks more accurately and efficiently. Subsequently, in considering the impact of external environmental uncertainties on the stability of vehicles in extreme states, the stability phase planes under different conditions are obtained based on the analysis of stability characteristics of vehicle lateral dynamics model, and the mathematical expressions of them are summarized according to change mechanism. A stability maintaining controller with yaw moment output is proposed by analyzing the stability phase plane. Finally, based on the output of the upper-level controller, a lower-level torque distribution control strategy is designed to achieve full performance through the optimal allocation of actuators. The integrated control strategy is deployed on an 8×8 prototype vehicle and was tested with multiple subjects under the condition of off-road. The test results show that the vehicle has better dynamic performance and safety in trajectory tracking under high-speed condition.

In recent years, the task offloading has been becoming a research hotspot. It is one of the key technologies to ensure the efficient cooperative operations of unmanned aerial vehicle (UAV) cluster, aiming to overcome the constraints of insufficient computing power and limited energy of a single platform. The purpose of reducing cost s and increasing efficiency is achieved by offloading the computing tasks to the servers of edge network for processing. In this paper, the UAV cluster-assisted air-ground integrated cooperative reconnaissance is taken as the combat scenario, and the complex wartime electromagnetic environment and the time-varying network topology of the cluster is considered. The long-term task offloading is decoupled into an online Markov decision process via Lyapunov optimization. To solve the problems of difficult convergence in hybrid action space and low learning efficiency, a multi-agent reinforcement learning offloading decision algorithm driven by data-model bi-level optimization is proposed by combining the convex optimization and multi-agent deep deterministic strategy to solve the power allocation and task allocation problem hierarchically. Numerical experiments show that the proposed algorithm can adaptively adjust the agent task offloading strategy according to the time-varying battlefield environment to improve the performance of traditional algorithm and optimize the complex multi-dimensional objectives.

This paper aims to improve the transient and steady-state performances of autonomous vehicle systems with input saturation and unknown perturbations. Firstly, a coordinated controller based on the sliding mode control and the prescribed performance control is designed considering the coupling between the lateral and longitudinal motion dynamics. To address the possible input saturation, an auxiliary system is designed to adjust the prescribed performance boundaries when saturation occurs, so that the tracking errors always adhere to the performance constraint. Consequently, it avoids the possible instability when the errors cross the performance boundaries. Finally, the neural network is introduced to approximate and compensate for the model uncertainty and external interference, and an online identification scheme based on a composite learning algorithm is proposed to train the neural network. The stability of the closed-loop system is strictly proved by Lyapunov approach, and the effectiveness of the proposed identification and control scheme is verified by simulation. The coordinated controller can be used to ensure the prescribed trajectory tracking performance in the presence of strong coupling characteristics, model uncertainty, and external interference.

In response to the insufficient consideration of terrain features in the existing path planning methods, a path planning method based on the estimation of terrain traversability for unmanned tracked vehicles is proposed. In the proposed method, a ConvLSTM-based deep neural network is used to extract the spatial and temporal correlation features of LiDAR point clouds from continuous trajectories, fuse the vehicle motion state features, and estimate the terrain traversability. Based on the terrain traversability, the node expansion mode and cost function of A^* algorithm are improved, and the discrete waypoints that meet the collision-free constraints and the low traversability cost are output. A gradient-free iterative smoothing algorithm is used to reduce the cost of path relaxation and traversability. Then a cubic B-spline curve is used to generate a smooth reference path which is used to establish the Frenet coordinate system. A safety corridor based on the traversability is constructed in the Frenet coordinate system. On the premise of meeting the collision-free constraints and the low traversability cost, a smooth path meeting the vehicle kinematics constraints is generated in the corridor. The experimental results indicate that the proposed method can fully consider the terrain features and improve the stability of path planning results and the traversability of path.

With the in-depth research on military unmanned ground systems, it is difficult for a single unmanned ground vehicle (UGV) or mission payload to meet the needs of the modern battlefield. Only with the coordinated development of mission payload and UGVs, the ground unmanned system can truly form combat effectiveness in the battlefield. In order to further promote the development of collaborative technology for the mission payload and UGVs, the development background, research status and technical characteristics of military ground unmanned systems with mission payload are summarized, and then the key technologies of UGV system are elaborated from three aspects: multilevel and multi-dimensional environmental modeling, traversability assessment based on multimodal data, and collaborative planning and control optimization method based on multi-agent collaborative modeling. Relevant research framework and focus of each aspect are summarized, and the future development direction of military ground unmanned systems with mission payload is prospected.

The collaborative positioning and environmental awareness technologies are the cornerstone of autonomous navigation of unmanned swarm. However, due to the limitations of computing, load, bandwidth and other resources of small-sized individual platforms in large-scale unmanned swarm systems, many related technologies are difficultly deployed and applied in practice. In order to achieve the accurate positioning and environmental awareness of large-scale unmanned swarm under resource constraints, a lightweight collaborative visual SLAM algorithm based on semi-direct method is proposed, and a semi-direct feature point tracking method that combines the optical flow method and the direct method is designed. The centralized two-way communication strategy is used to make the large-scale unmanned swarm system have a high fault-tolerant rate in the face of communication interference and delay, and the unmanned swarm system. The comparative experiments were conducted on the algorithm based on the EuRoC dataset and actual physical environment. The results show that the real-time performance of the proposed algorithm is improved by an average of 60%, which is significantly better than other feature-based collaborative visual SLAM algorithms. In a low-quality communication environment where the packet loss rate is less than 40% and the communication delay is less than 0.1 second, the proposed algorithm has higher positioning accuracy and stronger robustness.

For the false data injection attacks against sensor and actuator in cyber-physical systems, the reconstruction problem of the attack signals is studied by using the observer technique. It is difficult for a defender to directly use the estimation errors of the attack signals. As a result, more accurately reconstructed attack signals are obtained hardly. Therefore, a virtual auxiliary system containing the information of the estimated errors of attack signals is established with the help of the system input and output data , and then a cooperative interaction observer is constructed by using the cooperative interaction strategy. Through the cooperative interaction with the virtual auxiliary system, the observer uses the information of the estimated errors of attack signals to improve the reconstruction accuracy of the attack signals and obtains more accurate attack signals. It is proven by Lyapunov approach that the reconstructed attack signals can converge to a small set around real attack signals. Then, the effectiveness of the new scheme is demonstrated through numerical simulation.

A three-dimensional leader-follower cooperative guidance law is proposed for the salvo attack of multiple missiles against a maneuvering target with the desired angles of impact. A three-dimensional nonlinear guidance model for the multiple missiles is established based on the relative motion relationship of missile and target, in which it is not necessary to assum a small lead angle. The three-dimensional guidance laws of leader and followers in the normal and lateral directions to the LOS are designed, respectively, based on the second-order sliding mode control (SMC) theory and designed finite-time convergence SMC surface, which can improve the system convergence speed while suppressing chattering phenomenon. The proposed guidance law innovatively transforms the time consistency problem of the leader-following missiles into a second-order multiagent consensus tracking in the LOS direction of the followers, which achieves the time consistency of leader and followers and, effectively solves the practical problem of low guidance precision caused by the time-to-go estimation. Moreover, the proposed guidance algorithms are proved by strict Lyapunov stability, respectively. The simulated results show that the proposed three-dimensional leader-follower cooperative guidance law can effectively control the leader and followers to attack a maneuvering target with high precision at the desired angles of attack.

In response to the limitations of the traditional Sand Cat Swarm Optimization(SCSO) algorithm, including inadequate global search capability and susceptibility to local optima, an improved Sand Cat Swarm Optimization (LVSCSO) algorithm is proposed. The proposed algorithm introduces a nonlinear adjustment mechanism to better encapsulate the search and attack processes inherent in SCSO algorithm. Moreover, an adaptive Levy flight mechanism is incorporated to effectively enhance the algorithm's global search capability and capacity to escape local optima. A grid-based approach is used to establish the wilderness and complex urban environment models for unmanned aerial vehicles(UAVs). A composite fitness function, considering the factors such as path length, flight altitude, and flight angles, serves as the evaluation metric. The algorithm is validated through simulation.The results show that, in the wilderness environment model, the improved algorithm achieves the enhancements of 56.40% and 22.06% over the traditional SCSO algorithm and the particle swarm optimization algorithm, respectively. In the complex urban environment model, the improvements are 56.33% and 61.80% compared to the traditional SCSO algorithm and the particle swarm algorithm, respectively. These findings highlight the efficacy and superiority of the improved SCSO algorithm in the context of path planning.

An online cooperative path planning method for maneuvering target positioning in two typical mission scenarios is presented to improving the accuracy of maneuvering target positioning. Firstly, a direction-finding cross-target positioning model is established. The factors affecting the positioning accuracy are determined based on the geometric dilution of precision, and the effect of the influencing factors on the positioning accuracy is analyzed. Secondly, for the positioning of low-speed maneuvering target, an online cooperative track generation method based on variable curvature Dubins curve is proposed with a geometric accuracy factor as the evaluation index, and the fast track generation of the optimal configuration of cooperative positioning is realized. For the cooperative positioning of high-speed maneuvering targets, an optimal control model for target positioning accuracy is established, and a penalty function cooperative path planning method based on interior point method is proposed. Finally, the digital and semi-physical simulation verification is carried out. The results show that the target cooperative positioning accuracy is improved by 36.9% and 23.5%, respectively, in two typical task scenarios, which verifies the effectiveness of the proposed method. This method has the engineering application value for cluster cooperative target positioning.

Autonomous cooperative combat of UAV cluster is an important combat mode in future wars. The study on autonomous collaboration mechanism of UAV cluster is conducive to reveal the formation and evolution pattern of its cooperation for providing the theoretical support for its application in combat. Firstly, an autonomous collaboration information interaction network for UAV cluster is constructed, and the UAV interaction levels based on topological potential are defined. An UAV cluster evolutionary game model based on the public goods game is established by analyzing the mapping relationship between autonomous cooperation of UAV cluster and evolutionary game, and the overall expected gain function, average expected gain function and improved aspiration dynamic strategy based on UAV interaction levels are proposed. Then, the evolution process of UAV cluster is quantitatively described by Markov chain, and the average abundance function which characterizes the evolutionarily stable distribution of UAV cluster is deduced theoretically. Finally, the feasibility and effectiveness of the autonomous collaboration mechanism of UAV cluster based on public goods game are verified by simulation, and the influences of the changes in the selection intensity, interaction gain factor, enhancement factor and aspiration level on the collaboration of UAV cluster are analyzed, which provides the theoretical support for autonomous cooperative parameter regulation design of UAV cluster. The results show that the proposed autonomous collaboration mechanism can promote the emergence of UAV cluster cooperation to a greater extent than the point and punishment mechanisms, and provide technical support for UAV cluster autonomous collaboration theory to be applied to combat.

To study the operational concept of all-optical UAV swarm in unknown complex-terrain battlefield environment, an evaluation method is proposed. The method is used to evaluate the effectiveness and the system-of-systems (SoS) contribution rate of all-optical UAV swarm and is based on Agent-based modeling and simulation (ABMS). Firstly, the evaluation index system is established according to the combat characteristics of all-optical UAV swarm. Then, the modular agent model of each combat unit is given. Finally, the simulation of patrol and attack mission in ultra-low altitude is carried out. Analyzing the influences of all-optical UAV perception performance, all-optical UAV communication performance, swarm size and swarm configuration on swarm survival rate, tasks completion rate, battle damage rate and SoS contribution rate. The research results show that the all-optical UAV can bring new capabilities to the combat system and effectively improve the combat effectiveness of SoS. Results also show that a larger swarm size and a higher proportion of all-optical UAV can increase the combat effectiveness of SoS. And the contribution brought by the improvement of the performance of all-optical UAV is not significant.

A three-dimensional cooperative guidance method is presented to investigate the problem of leader and seeker-less follower flight vehicles attacking on the targets by sector coverage. An impact time control guidance law with fixed-time convergence is presented for the leader to reach a target at desired impact time. A sector coverage configuration against target with dynamic self-adjusting center and radius is proposed with the leader as baseline. Finally,on the basis of the desired converge sector, the cooperative guidance law is given for the followers to reach the target simultaneously with multi-directional sector blockade. A theoretical analysis is given to prove the stability of the proposed method. The simulated results show that the proposed method can be used to make all the flight vehicles hit the target at appointed impact time, and the leader and the seeker-less followers can form the coverage situation towards the target.

For the failure caused by the deterioration of the dynamic characteristics of external oil pipe in the integrated transmission device of armoured vehicle, the dynamic characteristics and structure strength change laws of external oil pipe under the dynamic excitation of different vehicle speeds are studied, and the effective optimization and improvement measures for the external oil pipe are proposed. The dynamic characteristics of typical external lubricating oil pipe and oil supply pipe system on the integrated transmission device are analyzed. The load excitation of key parts of the oil pipe is obtained by test. The transient dynamic simulation and theoretical dynamic modeling are conducted to study the influence of key parameters on the dynamic characteristics of oil pipe system, and an optimization and improvement scheme of the dynamic characteristics is proposed. The vibration amplitude of key parts of the oil pipe system have decreased by more than 15%.

Operational effectiveness prediction is a great significance to the weapon equipment system in the whole process from construction, production to actual combat. The cross-validation method for data is optimized based on the Stacking ensemble learning model. For the problem of sparse input vector of secondary learner in the original model, the input polynomial characteristics and the combat simulation data indicators (raw data) after PCA dimension reduction are increased to the learning layer, A prediction method for the operational effectiveness of equipment system with improved Stacking ensemble learning model is proposed. The effectiveness of the method is verified by an example of the operational effectiveness prediction of a synthetic battalion seizing a position.

A human-machine collaborative control algorithm based on model prediction and policy learning is proposed for the optimal decision-making and high maneuvering motion control of intelligent vehicles in complex environments. The algorithm takes advantage of the human driver's understanding of the environment and comprehensive processing ability to assist the machine in local trajectory planning at the decision planning level, including speed adjustment and dynamic path generation, to achieve the human-machine collaboration .For the timeliness of the online optimal planning and control of vehicles with high maneuverability, on the one hand, a long sampling interval and a simplified dynamics model are used to design a local trajectory planning method based on model predictive control at the local planning level in order to achieve efficient online trajectory optimization. On the other hand, a learning-based predictive control method based on rolling time-domain reinforcement learning is used to optimize the control strategy in the control layer in order to improve the computational efficiency and adaptability of online optimal control. In the driving simulation on the mountain highway with the driver in the loop, the proposed method not only complies with the driver's acceleration and deceleration commands and steering commands to generate a safe and smooth planning trajectory for human-machine cooperation, but also can accurately control the vehicle to travel along the desired trajectory in real time. In the human-machine cooperative control mode, the time to complete the same driving task is reduced by 8.3% on average and the steering operation load is reduced by 51.1% compared with the manual driving by six ordinary drivers.

A passive obstacle-surmounting six-wheeled rocker-type mobile robot is roposed to improve the obstacle-surmounting performance of mobile robots in complex terrain environment. The obstacle-surmounting performance of the robot is studied for climbing steps. The relationship between the structural parameters and the obstacle-surmounting height when the robot does not have the bottom support phenomenon is obtained from the analysis of geometric conditions, and the key obstacle-surmounting process of the robot is analyzed. The obstacle-surmounting dynamics models of front wheels, middle wheels and rear wheels are established. The influences of the robot's speed, acceleration, adhesion coefficient and the position of the center of gravity on the output torque of hub motor are analyzed based on the dynamics models. The passing capability of robot is verified by using the prototype experiment and ADAMS simulation software, and the output torque curve of robot's hub motor is obtained to guide the selection of motor. The research results show that the robot has good passing capability for climbing steps, which provides a design reference for improving the obstacle-surmounting performance of robot in complex terrain.

A target assignment model for high-power microwave (HPM) weapon system and midium- and short-range air defense weapons is proposed for the challenges of air defense operations brought by the development of weapon equipment technology. In the target assignment model, the combat scenario is set, the soft and hard killing effectivenesses of HPM weapon system and the cooperative combat among HPM weapon system with medium-range air defense missiles, short-range air defense missiles and terminal close-range anti-aircraft guns are considered, and the interception efficiency of HPM and medium- and short-range air defense weapons is analyzed. A simulated annealing method is used to give a solution method of the target assignment model, and obtain an air defense weapon allocation scheme that maximizes the interception efficiency. The results of simulation experiments show that, compared with Monte Carlo method and particle swarm algorithm, the simulated annealing method can find a solution with higher interception efficiency in a shorter time, and can effectively solve the problem of target assignment in cooperative combat. It lays the foundation for the integration of HPM weapon system into the multi-weapon coordinated air defense combat system.

The deviation and even safety problems of vehicle are easily caused if the unilateral motor of the tracked vehicle driven by two-sided motor fails. The safety control of vehicle braking and obstacle avoidance in the failure mode of unilateral motor is studied to ensure the safety of vehicles. The vehicle safety is analyzed based on the control mode that one side of the real vehicle fails and the other side is in the failure mode in time. Therefore, an obstacle avoidance control strategy in the braking process is added based on the current control mode. A vehicle safety control strategy of tracked vehicle driven by double-side motor coupling in the mode of single-side motor failure is presenred, and it is verified by RT-LAB hardware-in-the-loop real-time simulation. The results show that this control strategy can be used to realize the steering control of vehicles with different relative steering radii at different speeds in the mode of single-side motor failure according to the driver's intention, and it can stabilize the steering and ensure the safety of tracked vehicle in the face of continuous obstacle avoidance requirements.

Few-shot object detection(FSOD) aims to enable the detector with a small number of training samples. Typical FSOD method takes Faster R-CNN as the basic detection framework, and uses a convolutional neural network to extract the image features. However, the pooling operation used in the convolutional neural network inevitably leads to the loss of image information. Therefore, a hybrid dilated convolution is introduced into the backbone network to ensure a larger receptive field and minimize the loss of image information. A support feature dynamic fusion module is proposed to further utilize the given support data in k-shot setting, which adaptively fuses the support features with the weight of the correlation between each support feature and query feature to obtain stronger support clues. Experimental results show that rhe proposed method achieves good and state-of-the-art FSOD performance on public Pascal VOC and MS-COCO datasets.

For the dynamic firepower allocation in the cooperative air defense operation of strategic locations on the sea, the characteristics of air defense operations in strategic locations on the sea are comprehensively analyzed to establish the dynamic firepower allocation problem based on the Markov decision model, and an optimization model with the damage expectation and interception cost as the indexes is constructed. Considering the problem that the Markov decision model is easy to fall into the disaster of dimensionality, an approximate dynamic programming method is proposed to explore the validity of the solution, and a least squares temporal difference algorithm based on reinforcement learning is given to solve the problem. The simulated results of 80 cases in four typical offensive and defensive scenarios show that, compared with the traditional matching algorithm, genetic algorithm and particle swarm optimization algorithm, the proposed model and algorithmin this paper are more scientific, reasonable and effective, which can provide a certain basis for the firepower allocation in the cooperative air defense operations of strategic locations on the sea.

The higher requirements are put forward for the environmental adaptability of special unmanned vehicles in the future battlefield. In order to meet the requirements of multi-working conditions, high maneuverability and low-cost development of special unmanned vehicles, the truss body structure of unmanned vehicles is optimized by the idea of multi-working conditions correlation design and lightweight optimization. However, considering that there are many design variables, large design space and too many simulation times in multi-working conditions, a lightweight optimization method of body structure based on multi-working conditions correlation is proposed; In the proposed method, the strategy of reducing design variables interval is used to reduce the design space. Gaussian process surrogate model is introduced to replace the simulation analysis for realizing the rapid performance evaluation of structural design scheme, and the optimization of the scheme is realized using genetic algorithm. The experimental results show that the mass of the final optimization scheme is 14.12% lower than that of the initial scheme and 8.87% lower than that of the scheme optimized only by Gaussian process under the condition of multi-disciplinary performance simulation verification, such as stiffness, strength and mode.