In order to effectively solve the inefficient path planning problem of mobile robots traveling in multiple destinations, a multi-destination global path planning algorithm is proposed based on the optimal obstacle value. A grid map is built to simulate the real working environment of mobile robots. Based on the rules of the live chess game in Go, the grid map is optimized and reconstructed. This grid of environment and the obstacle values of grid environment between each two destination points are obtained. Using the simulated annealing strategy, the optimization of multi-destination arrival sequence for the mobile robot is implemented by combining with the obstacle value between two destination points. The optimal mobile node of path planning is gained. According to the Q-learning algorithm, the parameters of the reward function are optimized to obtain the q value of the path. The optimal path of multiple destinations is acquired when mobile robots can pass through the fewest obstacles. The multi-destination path planning simulation of the mobile robot is implemented by MATLAB software (Natick, MA, USA, R2016b) under multiple working conditions. The Pareto numerical graph is obtained. According to comparing multi-destination global planning with single-destination path planning under the multiple working conditions, the length of path in multi-destination global planning is reduced by 22% compared with the average length of the single-destination path planning algorithm. The results show that the multi-destination global path planning method of the mobile robot based on the optimal obstacle value is reasonable and effective. Multi-destination path planning method proposed in this article is conducive to improve the terrain adaptability of mobile robots.

The accuracy of an articulated torque analysis influences the comprehensive performances of heavy-duty multi-legged robots. Currently, the extremal estimation method and some complex methods are employed to calculate the articulated torques, which results in a large safety margin or a large number of calculations. To quickly obtain accurate articulated torques, an analysis method for the articulated torque is presented for an electrically driven heavy-duty six-legged robot. First, the rearmost leg that experiences the maximum normal contact force is confirmed when the robot transits a slope. Based on the ant-type and crab-type tripod gaits, the formulas of classical mechanics and MATLAB software are employed to theoretically analyze the relevant static torques of the joints. With the changes in the joint angles for the abductor joint, hip joint, and knee joint, variable tendency charts and extreme curves are obtained for the static articulated torques. Meanwhile, the maximum static articulated torques and the corresponding poses of the robot are also obtained. According to the poses of the robot under the maximum static articulated torques, ADAMS software is used to carry out a static simulation analysis. Based on the relevant simulation curves of the articulated torques, the maximum static articulated torques are acquired. A comparative analysis of the maximum static articulated torques shows that the theoretical calculation values are higher than the static simulation values, and the maximum error value is approximately 10%. The proposed method lays a foundation for quickly determining accurate articulated torques to develop heavy-duty six-legged robots.

Heavy-duty legged robots have played an important role in material transportation, planet exploration, and other fields due to their unique advantages in complex and harsh terrain environments. The instability phenomenon of the heavy-duty legged robots often arises during the dynamic interactions between the supporting feet and the intricate terrains, which significantly impact the ability of the heavy-duty legged robots to move rapidly and accomplish tasks. Therefore, it is necessary to assess the mechanical behavior of foot–terrain interactions for the heavy-duty legged robots. In order to achieve the above goal, a systematic literature review methodology is employed to examine recent technical scientific publications, aiming to identify both current and prospective research fields. The characteristics of supporting feet for different heavy-duty legged robots are compared and analyzed. The foot–terrain mechanical models of the heavy-duty legged robots are discussed. The problems that need further research are summarized and presented, which is conducive to further deepening and expanding the research on the mechanical behavior of foot–terrain interactions for heavy-duty legged robots.

Gait planning is an important basis for the walking of a legged robot. To improve the walking stability of multi-legged robots and to reduce the impact force between the foot and the ground, gait planning strategies are presented for an electrically driven large-load-ratio six-legged robot. First, the configuration and walking gait of the electrically driven large-load-ratio six-legged robot are designed. The higher-stable swing sequences of legs and typical walking modes are respectively obtained. Based on the Denavit–Hartenberg (D–H) method, the analyses of the forward and inverse kinematics are implemented. The mathematical models of the articulated rotation angles are respectively established. In view of the buffer device installed at the end of shin to decrease the impact force between the foot and the ground, an initial lift height of the leg is brought into gait planning when the support phase changes into the transfer phase. The mathematical models of foot trajectories are established. Finally, a prototype of the electrically driven large-load-ratio six-legged robot is developed. The experiments of the prototype are carried out regarding the aspects of the walking speed and surmounting obstacle. Then, the reasonableness of gait planning is verified based on the experimental results. The proposed strategies of gait planning lay the foundation for effectively reducing the foot–ground impact force and can provide a reference for other large-load-ratio multi-legged robots.

The combination of gesture recognition and aerospace exploration robots can realize the efficient non-contact control of the robots. In the harsh aerospace environment, the captured gesture images are usually blurred and damaged inevitably. The motion blurred images not only cause part of the transmitted information to be lost, but also affect the effect of neural network training in the later stage. To improve the speed and accuracy of motion blurred gestures recognition, the algorithm of YOLOv4 (You Only Look Once, vision 4) is studied from the two aspects of motion blurred image processing and model optimization. The DeblurGanv2 is employed to remove the motion blur of the gestures in YOLOv4 network input pictures. In terms of model structure, the K-means++ algorithm is used to cluster the priori boxes for obtaining the more appropriate size parameters of the priori boxes. The CBAM attention mechanism and SPP (spatial pyramid pooling layer) structure are added to YOLOv4 model to improve the efficiency of network learning. The dataset for network training is designed for the human–computer interaction in the aerospace space. To reduce the redundant features of the captured images and enhance the effect of model training, the Wiener filter and bilateral filter are superimposed on the blurred images in the dataset to simply remove the motion blur. The augmentation of the model is executed by imitating different environments. A YOLOv4-gesture model is built, which collaborates with K-means++ algorithm, the CBAM and SPP mechanism. A DeblurGanv2 model is built to process the input images of the YOLOv4 target recognition. The YOLOv4-motion-blur-gesture model is composed of the YOLOv4-gesture and the DeblurGanv2. The augmented and enhanced gesture data set is used to simulate the model training. The experimental results demonstrate that the YOLOv4-motion-blur-gesture model has relatively better performance. The proposed model has the high inclusiveness and accuracy recognition effect in the real-time interaction of motion blur gestures, it improves the network training speed by 30%, the target detection accuracy by 10%, and the value of mAP by about 10%. The constructed YOLOv4-motion-blur-gesture model has a stable performance. It can not only meet the real-time human–computer interaction in aerospace space under real-time complex conditions, but also can be applied to other application environments under complex backgrounds requiring real-time detection.