This paper describes a small range six-axis accelerometer (the measurement range of the sensor is ±g) with high sensitivity DCB (Double Cantilever Beam) elastic element. This sensor is developed based on a parallel mechanism because of the reliability. The accuracy of sensors is affected by its sensitivity characteristics. To improve the sensitivity, a DCB structure is applied as the elastic element. Through dynamic analysis, the dynamic model of the accelerometer is established using the Lagrange equation, and the mass matrix and stiffness matrix are obtained by a partial derivative calculation and a conservative congruence transformation, respectively. By simplifying the structure of the accelerometer, a model of the free vibration is achieved, and the parameters of the sensor are designed based on the model. Through stiffness analysis of the DCB structure, the deflection curve of the beam is calculated. Compared with the result obtained using a finite element analysis simulation in ANSYS Workbench, the coincidence rate of the maximum deflection is 89.0% along the x-axis, 88.3% along the y-axis and 87.5% along the z-axis. Through strain analysis of the DCB elastic element, the sensitivity of the beam is obtained. According to the experimental result, the accuracy of the theoretical analysis is found to be 90.4% along the x-axis, 74.9% along the y-axis and 78.9% along the z-axis. The measurement errors of linear accelerations a_x, a_y and a_z in the experiments are 2.6%, 0.6% and 1.31%, respectively. The experiments prove that accelerometer with DCB elastic element performs great sensitive and precision characteristics.

Background. With the advancement of “digital forestry” and “intelligent forestry”, point cloud data have emerged as a powerful tool for accurately capturing three-dimensional forest scenes. It enables the creation and presentation of digital forest systems, facilitates the monitoring of dynamic changes such as forest growth and logging processes, and facilitates the evaluation of forest resource fluctuations. However, forestry point cloud data are characterized by its large volume and the need for time-consuming and labor-intensive manual processing. Deep learning, with its exceptional learning capabilities, holds tremendous potential for processing forestry environment point cloud data. This potential is attributed to the availability of accurately annotated forestry point cloud data and the development of deep learning models specifically designed for forestry applications. Nonetheless, in practical scenarios, conventional direct annotation methods prove to be inefficient and time-consuming due to the complex terrain, dense foliage occlusion, and uneven sparsity of forestry point clouds. Furthermore, directly applying deep learning frameworks to forestry point clouds results in subpar accuracy and performance due to the large size, occlusion, sparsity, and unstructured nature of these scenes. Therefore, the proposal of accurately annotated forestry point cloud datasets and the establishment of semantic segmentation methods tailored for forestry environments hold paramount importance. Methods. A point cloud data annotation method based on single-tree positioning to enhance annotation efficiency was proposed and challenges such as occlusions and sparse distribution in forestry environments were addressed. This method facilitated the construction of a forestry point cloud semantic segmentation dataset, consisting of 1259 scenes and 214.4 billion points, encompassing four distinct categories. The pointDMM framework was introduced, a semantic segmentation framework specifically designed for forestry point clouds. The proposed method first integrates tree features using the DMM module and constructs key segmentation graphs utilizing energy segmentation functions. Subsequently, the cutpursuit algorithm is employed to solve the graph and achieve the pre-segmentation of semantics. The locally extracted forestry point cloud features from the pre-segmentation are comprehensively inputted into the network. Feature fusion is performed using the MLP method of multi-layer features, and ultimately, the point cloud is segmented using the lightweight PointNet. Result. Remarkable segmentation results are demonstrated on the DMM dataset, achieving an accuracy rate of 93% on a large-scale forest environment point cloud dataset known as DMM-3. Compared to other algorithms, the proposed method improves the accuracy of standing tree recognition by 21%. This method exhibits significant advantages in extracting feature information from artificially planted forest point clouds obtained from TLS. It establishes a solid foundation for the automation, intelligence, and informatization of forestry, thereby possessing substantial scientific significance.

Inertial navigation technology composed of inertial sensors is widely used in foot-mounted pedestrian positioning. However, inertial sensors are susceptible to noise, which affects the performance of the system. Zero-velocity update (ZUPT) is a traditional method utilized to suppress the cumulative error. Unfortunately, the walking distance calculated by a Kalman filter still has positional error. To improve the positional accuracy, in this work we propose a nonlinear Kalman filter with a spatial distance inequality constraint for a single foot. Since the stride distance between adjacent stance phases has an upper bound in plane and height, an inertial navigation system established by one inertial measurement unit is adopted to constrain the stride process. Eventually, the performance of the proposed method is verified by experiments. Compared with the single foot-mounted ZUPT method, the proposed method suppresses the plane error and the height error by 46.04% and 65.48%, respectively. For the dual-foot constraint method, the proposed constraint method can reduce the number of sensors while ensuring positioning accuracy. Moreover, the height error is reduced by 59.98% on average by optimizing the constraint algorithm. The experimental results show that the trajectory estimated by the proposed method is closer to the actual path.

This paper presents a theoretical and experimental study conducted on the rollover warning of wheeled off-road operating vehicles. The time to rollover warning algorithm was studied with real-time vehicle roll angle and roll angle velocity as the input variables, and lateral load transfer ratio was used as the rollover determination index. Subsequently, a vehicle dynamics model was built using CarSim software, and a warning algorithm was established in the MATLAB/Simulink environment. The rollover joint simulation in CarSim and MATLAB/Simulink was conducted under typical working conditions. Finally, combined with inertial measurements, a rollover warning system was independently developed. In addition, the rollover warning system was installed on a light forest firefighting truck to verify the feasibility of the system via a real vehicle experiment, and the law of vehicle rollover motion was also studied. The serpentine experiment and steady-state rotation experiment were conducted. The experimental results showed that at identical front-wheel steering angles, the roll angle and lateral acceleration increased with an increase in the vehicle speed. Furthermore, for identical vehicle speeds, the roll angle and lateral acceleration of the vehicle increased with an increase in the front-wheel steering angle. The dangerous vehicle speed was 50 km h ⁻¹ in the serpentine condition and 40 km h ⁻¹ in the steady-state rotation condition. The risk trend and alarm signal obtained by the rollover warning system were consistent with the actual situation. Thus, this can assist drivers in judging the rollover risk and effectively improve the active safety of special vehicles. Furthermore, it also provides a reference for further research on active rollover control technology of special vehicles.