LiDAR-based simultaneous localization and mapping (SLAM) methods often degrade in environments lacking geometric features (such as tunnels and long corridors), where planar features are abundant but linear features are sparse. To address this issue, we propose RID-LIO, an intensity-assisted LiDAR-inertial SLAM framework that integrates adaptive intensity feature extraction and intensity-based loop detection, overcoming the reliance on geometric features in existing methods. First, 3D point clouds are cylindrically projected to generate intensity images, from which intensity line features are extracted to enhance constraints in degraded directions. A weighting function is also incorporated to optimize the quality of the pose estimation, while an efficient intensity edge context descriptor improves loop detection efficiency and reduces trajectory drift. Evaluations on the VECtor dataset show an average improvement of 63.59% in trajectory accuracy. Tests on a private dataset demonstrate that RID-LIO outperforms other state-of-the-art methods in terms of end-to-end error and map consistency.

As unmanned vehicle technology advances rapidly, obstacle recognition and target detection are crucial links, which directly affect the driving safety and efficiency of unmanned vehicles. In response to the inaccurate localization of small targets such as pedestrians in current object detection tasks and the problem of losing local features in the PointPillars, this paper proposes a three-dimensional object detection method based on improved PointPillars. Firstly, addressing the issue of lost spatial and local information in the PointPillars, the feature encoding part of the PointPillars is improved, and a new pillar feature enhancement extraction module, CSM-Module, is proposed. Channel encoding and spatial encoding are introduced in the new pillar feature enhancement extraction module, fully considering the spatial information and local detailed geometric information of each pillar, thereby enhancing the feature representation capability of each pillar. Secondly, based on the fusion of CSPDarknet and SENet, a new backbone network CSE-Net is designed in this paper, enabling the extraction of rich contextual semantic information and multi-scale global features, thereby enhancing the feature extraction capability. Our method achieves higher detection accuracy when validated on the KITTI dataset. Compared to the original network, the improved algorithm’s average detection accuracy is increased by 3.42%, it shows that the method is reasonable and valuable.

Soft robotics represents a rapidly advancing and significant subfield within modern robotics. However, existing soft actuators often face challenges including unwanted deformation modes, limited functional diversity, and a lack of versatility. This paper presents a four-chamber multimodal soft actuator with a centrally symmetric layout and independent pneumatic control. While building on existing multi-chamber concepts, the design incorporates a cruciform constraint layer and inter-chamber gaps to improve directional bending and reduce passive chamber deformation. An empirical model based on the vector superposition of single- and dual-chamber inflations is developed to describe the bending behavior. Experimental results show that the actuator can achieve omnidirectional bending with errors below 5% compared to model predictions. To demonstrate versatility, the actuator is implemented in two distinct applications: a three-finger soft gripper that can grasp objects of various shapes and perform in-hand twisting maneuvers, and a steerable crawling robot that mimics inchworm locomotion. These results highlight the actuator’s potential as a reusable and adaptable driving unit for diverse soft robotic tasks.

Background Colonoscopy is the best choice for detecting bowel cancer; its effectiveness and early screening can significantly reduce the incidence rate of bowel cancer. However, the existing colonoscopy procedure is very technically demanding for doctors and painful for patients due to the difficult maneuver of long and slender colonoscope inside human colon. This paper will be focused on designing flexible robots with high dexterity to improve the performance of current colonoscopy. Methods This article optimizes and simulates the shape and layout of the driving cavity of the soft robot, as well as the multi-stage structure. This paper presents a design scheme of the soft robot with variable cross-section. The robot comprises two cross-sections in serial connection, with the outer diameter of the upper section 14mm and the outer diameter of the lower part 16mm. The bending angle and direction of the soft robot can be controlled by adjusting the pressure of the cavity. The soft robot is placed at the end of the traditional colonoscopy, replacing the standard manual operation for guidance during colonoscopy surgery. The relationship between the end position of the actuator and air pressure is calculated by the piecewise constant curvature (PCC) method and verified by subsequent experiments. At the same time, the motion trajectory of the soft robot is further simulated by finite element analysis. Results Experimental verification shows that the bending performance of the newly designed soft robot has significantly improved compared to the old design. Conclusion An improvement was made to address the design shortcomings of traditional actuators, and a new actuator was obtained. The performance of the latest and old actuators was compared through experiments, and it was ultimately known that the new actuator had significantly improved performance compared to the old actuator.

To address the challenge that existing soft grippers have difficulty achieving fine manipulation comparable to the human finger’s “circular twisting” motion, this paper proposes a four-chamber spatial bending soft actuator based on the principle of virtual work. The actuator incorporates an internal cross-shaped restricting layer that divides its cross-section into four independent pneumatic chambers. Through independent regulation of the pressure in each chamber, continuous and controllable bending in arbitrary spatial directions is achieved, replicating the bending and abduction/adduction degrees of freedom (DoFs) of a human finger and their composite motions on a single actuator. Based on the Yeoh hyperelastic constitutive model and the principle of virtual work, a static deformation model of the actuator is established. By introducing an engineering assumption of “deformation vector superposition” and correction coefficients fitted from experimental data, high-precision prediction from multi-chamber pressure input to spatial bending output is realized. Furthermore, a three-finger soft gripper is constructed based on this actuator, successfully demonstrating fingertip pinching and enveloping grasping. Through open-loop programmed control, the fine “circular twisting” manipulation is demonstrated (exemplified by light bulb installation). This study provides an effective structural design and modeling method for soft actuators to achieve decoupled multi-DoF motion control, showcasing their application potential in adaptability and dexterous manipulation.