Food quality and safety issues occurred frequently in recent years, which have attracted more and more attention of social and international organizations. Considering the increased quality risk in the food supply chain, many researchers have applied various information technologies to develop real‐time risk identification and traceability systems (RITSs) for preferable food safety guarantee. This paper presents an innovative approach by utilizing the deep‐stacking network method for hazardous risk identification, which relies on massive multisource data monitored by the Internet of Things timely in the whole food supply chain. The aim of the proposed method is to help managers and operators in food enterprises to find accurate risk levels of food security in advance and to provide regulatory authorities and consumers with potential rules for better decision‐making, thereby maintaining the safety and sustainability of food product supply. The verification experiments show that the proposed method has the best performance in terms of prediction accuracy up to 97.62%, meanwhile achieves the appropriate model parameters only up to 211.26 megabytes. Moreover, the case analysis is implemented to illustrate the outperforming performance of the proposed method in risk level identification. It can effectively enhance the RITS ability for assuring food supply chain security and attaining multiple cooperation between regulators, enterprises, and consumers.

In this paper, a novel semi-supervised segmentation framework based on a spot-divergence supervoxelization of multi-sensor fusion data is proposed for autonomous forest machine (AFMs) applications in complex environments. Given the multi-sensor measuring system, our framework addresses three successive steps: firstly, the relationship of multi-sensor coordinates is jointly calibrated to form higher-dimensional fusion data. Then, spot-divergence supervoxels representing the size-change property are given to produce feature vectors covering comprehensive information of multi-sensors at a time. Finally, the Gaussian density peak clustering is proposed to segment supervoxels into sematic objects in the semi-supervised way, which non-requires parameters preset in manual. It is demonstrated that the proposed framework achieves a balancing act both for supervoxel generation and sematic segmentation. Comparative experiments show that the well performance of segmenting various objects in terms of segmentation accuracy (F-score up to 95.6%) and operation time, which would improve intelligent capability of AFMs.

The hydraulic system is a key component of intelligent forestry harvesters. In the testing of a forestry harvester, researchers need to analyze the operating efficiency and energy consumption of the forestry harvester based on the pressure and flow rate data in the hydraulic system of the forestry harvester and formulate energy-efficient control strategies. In order to enable researchers to monitor and extract the parameters of the hydraulic system of an intelligent forestry harvester in real time, this paper designs a hydraulic online monitoring system for forestry harvesters based on the LabVIEW 2019 software platform. This system realizes the following functions by reading the CAN (controller area network, a serial communication protocol for multi-host localized networks) bus of the harvester: (1) it collects and stores hydraulic system pressure, flow, and other data and displays the value curve in the system interface in real time, and (2) it monitors the control signals received by the hydraulic system and displays the control signal status and value received by the system interface in real time. This paper used this system to carry out on-site testing of an actual machine, and compared and analyzed the online monitoring data of the hydraulic system and the theoretical pressure data of the main hydraulic valve manifold. The average value of the relative error between the two was 1.65%, and the maximum value of the relative error was 2.75%. The results show that the designed system has good accuracy and stability, and can effectively realize online monitoring of the working status of the hydraulic system of forestry harvesters during their operation.

To improve the stability of forestry chassis on the slope, a chassis-installed barycenter adjustable mechanism (BAM) is designed, and the control method of the counterweight is proposed to make the chassis barycenter move suitably to achieve the design purpose. The kinematic analysis of BAM is carried out, and the relationship between the translation, rotation, and vertical displacement of counterweight and the chassis barycenter is calculated. Furthermore, the variation curves obtained in Matlab show the barycenter can translate 100 mm, rotate from 0 to 360 degrees, and lower about 180 mm in the vertical direction. Adams is adopted to complete the kinematics simulation of the chassis, indicating that the control method can effectively adjust the barycenter position. Finally, experiments are carried out under slope conditions to analyze chassis stability by testing plantar pressure. The results show that forest chassis using the barycenter control method helps keep stable on the slope of 15 degrees, much better than standard normal chassis.

As an underactuated, heavy-load attachment mounted at the end of an external crane boom, a forest harvester head exhibits coupled active motion and passive swing, making its attitude difficult to observe directly during operation. This study considers an underactuated harvester-head system with an undamped suspension link and proposes a hydraulic–inertial multi-source information fusion cubature Kalman filter (MCKF) for monitoring the head’s active attitude and passive swing. First, a multibody dynamics model of the harvester head, including friction terms, is established from the Lagrange equations. Second, a hydraulic observation model is constructed from the fluid-continuity equations to estimate the active joint variables from pressure and flow-rate measurements. On this basis, the resulting MCKF incorporates segmented geometric-constraint mapping to combine hydraulic and dual-IMU information, thereby providing an integrated observation of active attitude and passive swing. Tilt-up and rotation tests are then conducted with the chassis and external crane boom held stationary. For the tested sequences, the mean absolute errors were 1.03° for active tilt, 0.99° for body motion expressed in the base frame, 0.679° for passive swing, and 1.60° for active rotation; the corresponding maximum absolute errors were 2.07°, 2.22°, 1.469°, and 3.84°. The results show that the complete observer can track the active and passive motion components under the tested conditions and provide a methodological basis for subsequent harvesting-head attitude monitoring and control studies.