Autonomous navigation of robots has entered an era of robustness. Therefore, the demand for stable navigation in extreme environments is increasing. 4D millimeter-wave radar has a strong ability to resist environmental interference and can maintain accurate perception even in extreme weather conditions. It is an excellent solution. However, current research on navigation mainly focuses on vision-based and LiDAR-based methods, and the attention paid to 4D millimeter-wave radar (x, y, z, Doppler) radar systems is very limited. The current 4D millimeter wave SLAM system treats 4D millimeter wave data as LiDAR data and directly processes point cloud data. At the same time, the 4D millimeter wave radar point cloud has large noise, strong sparsity, and difficult to extract geometric features such as edges and planes, which poses a challenge to the establishment of a high-precision SLAM system. In addition, the data volume and cost of 4D millimeter wave radar are directly related. The cost of tens of thousands of dollars for high-end 4D millimeter wave radar directly restricts its application. Therefore, the construction of a SLAM system using low-cost and low-data-volume 4D millimeter wave radar has become an urgent problem to be solved. To address these problems, this paper proposes a data-enhanced low-cost 4D millimeter-wave radar SLAM system. The system introduces: (1) a point cloud SLAM quality assessment framework to establish the relationship between data quality and SLAM performance; (2) a data enhancement technique based on quality assessment to generate high-quality data from sparse, low-quality 4D radar input to meet positioning and mapping requirements; (3) a SLAM system enhanced by quality characteristics, incorporating velocity calibration into scan matching to improve accuracy, and fusing IMU pre-integration with Doppler velocity pre-integration to obtain a robust integration factor. In addition, a velocity segment-based loop closure mechanism is constructed in the SLAM system. Verification experiments conducted in real environments with dataset comparisons covering distances from 180 meters to 4.8 kilometers have demonstrated the scientificity and effectiveness of our approach, making a significant contribution to the promotion of low-cost 4D millimeter-wave radar applications.

With the rapid development of smart cities and IoT technologies, the application of IoT sensing in smart firefighting and disaster response has gradually become a research and practical focus. As the primary sensing method for IoT sensors, LiDAR's performance under the influence of smoke has become particularly important. So far, no dataset with smoke characteristics is available. The actual LiDAR dataset for recording large-scale fires is difficult to operate, dangerous, and costly. In this regard, this paper uses a smoke physics diffusion model to simulate the smoke environment at the fire scene on existing datasets. The resulting enhanced dataset is used to train the network to reuse these datasets, aiming to achieve a more robust fire smoke neural network. The smoke physics diffusion model is parameterized, and wind direction and wind speed are added. Finally, a smoke diffusion model under the intensity and time of the diffusion source is obtained. The Hausdorff distance, Chamfer Distance (CD) and Earth Mover’s Distance (EMD), of the smoke point cloud are used as evaluation criteria, and the actual smoke data and simulated smoke data are compared to verify the effectiveness and accuracy of the above model. Combined with the LiDAR sensor model, the mechanism of smoke affecting LiDAR sensor imaging is studied. Based on this mechanism, a LiDAR smoke impact model with the Rayleigh scattering model as the core is established, the smoke diffusion model is combined with LiDAR data. This model modifies the smoke-free dataset data to obtain a smoke dataset enhanced by the smoke model that conforms to the sensor characteristics and actual smoke distribution. In order to verify the effectiveness and universality of the enhanced dataset, data training is performed on point cloud deep learning networks such as point-based RandLA-Net, grid-based Point-Voxel-KD, and projection-based Range-Net++ to verify the effectiveness of the above work. Finally, the training results show that the data enhanced by smoke conforms to the general law of the influence of smoke on object detection in the actual objective world. The presence of smoke will reduce the accuracy of network recognition, which has a greater impact on the projection-based point cloud deep learning method.

When a wheeled mobile robot (WMR) runs on loose soil (such as the planetary rover on surface of the planet), its wheels generally slip or skip. Since the slippage of the wheel directly affects the motion control and safety, it becomes urgent to effectively estimate the slippage. In this paper, an intuitive slippage estimation method is proposed based on wheel-ruts images, which aims to reduce the number of extra sensors and take advantages of the visual information effectively. Since the image samples sometimes are difficult to collect, a few-shot learning method is employed using distribution propagation graph network with dilated causal convolution layer (DCC-DPGN). The dilated causal convolution (DCC) layer is adopted in ResNet block to expand the receptive field and obtain the sequence information of wheel-ruts, which makes the model training more efficient. The proposed model is verified in the test set of images collected in real scene, which shows the potential of the proposed algorithm in slippage estimation.

Most existing vision-based simultaneous localization and mapping (SLAM) systems and their variants still assume that the observation is absolutely static and cannot work well in dynamic environments. Here, we present the Dynam-SLAM (Dynam), a stereo visual-inertial SLAM system capable of robust, accurate, and continuous work in high dynamic environments. Our approach is devoted to loosely coupling the stereo scene flow with an inertial measurement unit (IMU) for dynamic feature detection and tightly coupling the dynamic and static features with the IMU measurements for nonlinear optimization. First, the scene flow uncertainty caused by measurement noise is modeled to derive the accurate motion likelihood of landmarks. Meanwhile, to cope with highly dynamic environments, we additionally construct the virtual landmarks based on the detected dynamic features. Then, we build a tightly coupled, nonlinear optimization-based SLAM system to estimate the camera state by fusing IMU measurements and feature observations. Finally, we evaluate the proposed dynamic feature detection module (DFM) and the overall SLAM system in various benchmark datasets. Experimental results show that the Dynam is almost unaffected by DFM and performs well in static EuRoC datasets. Dynam outperforms the current state-of-the-art visual and visual-inertial SLAM implementations in terms of accuracy and robustness in self-collected dynamic datasets. The average absolute trajectory error of Dynam in the dynamic benchmark datasets is$\sim$90% lower than that of VINS-Fusion,$\sim$84% lower than that of ORB-SLAM3, and$\sim$88% lower than that of Kimera.

Free-space-oriented roadmaps typically generate a series of convex geometric primitives, which constitute the safe region for motion planning. However, a static environment is assumed for this kind of roadmap. This assumption makes it unable to deal with dynamic obstacles and limits its applications. In this paper, we present a dynamic free-space roadmap, which provides feasible spaces and a navigation graph for safe quadrotor motion planning. Our roadmap is constructed by continuously seeding and extracting free regions in the environment. In order to adapt our map to environments with dynamic obstacles, we incrementally decompose the polyhedra intersecting with obstacles into obstacle-free regions, while the graph is also updated by our well-designed mechanism. Extensive simulations and real-world experiments demonstrate that our method is practically applicable and efficient.

Simulation of planetary rovers moving on complex terrains is critical for Mars exploration. Equivalent stiffness is proposed and used to characterize the pressure-sinkage property of terrain, while friction angle to characterize the shearing property. Terramechanics model for calculating forces between rigid wheel and soil is proved to be the same with that contact model for calculating forces between rigid wheel and rock. A Digital Elevation Map with Physical Properties is developed and applied to simulate terrain physical properties along with its geometry information. The established methods are validated using simulation and experimental tests with a three-wheel-rover.

The wheels of planetary rovers will slip when they climbs up deformable slopes. On the contrary, the wheels will skid in the longitudinal direction in order to generate resistance force to balance the gravity component when a rover moves down the slopes. The wheel-terrain interaction principles of slip versus skid are quite different, but there is little research about the longitudinal skid mechanics and the relationship of it with the slip mechanics. This paper analyzes the problem of longitudinal slip and skid that occur to a wheel on the slopes with the knowledge of terramechanics. The slip and skid mechanics are compared based on experimental results measured by a single wheel testbed. The piece wise linear function is proposed to predict the drawbar pull and resistance moment under both slip and skid conditions. A semi-empirical equation of predicting the skid mechanics according to the slip mechanics is also provided. The models are verified using the experimental data.