Tracked-vehicle terramechanics aims to investigate the interaction between tracked running gear and deformable terrain in hilly and mountainous areas. Previous reviews have focused on individual models, numerical simulations, or application scenarios. It is often required to integrate with knowledge structure, research-hotspot evolution, engineering applications, and patent evolution. This study presented research progress and trends in tracked-vehicle terramechanics using bibliometric and knowledge-mapping analysis. Literature was also retrieved from the China National Knowledge Infrastructure and the Web of Science Core Collection in 2009-2025. Furthermore, 1 287 CNKI records and 907 WoS records were retained after deduplication and screening. Patent documents were retrieved from Google Patents in 2005-2025. Document-level records were used to determine annual trends in published invention applications and granted invention patents. Whereas simple patent families were used for global technology-topic analysis. As such, a systematic investigation was combined with bibliometric analysis, knowledge mapping, patent analysis, and mechanism-oriented synthesis. VOSviewer and CiteSpace were used to examine annual publication output, country/region and institutional collaboration, author networks, keyword co-occurrence, keyword bursts, and thematic evolution. The results show that research output increased overall, with marked acceleration after 2015. International research was organized around terramechanics, traction, sinkage, vehicle dynamics, parameter identification, and machine learning, whereas CNKI-indexed studies placed greater emphasis on tracked vehicles, tracked chassis, steering, simulation, agricultural machinery, and hilly terrain. The trajectory evolved from the Bekker pressure–sinkage relation, the Janosi–Hanamoto shear stress–displacement relation, and Wong–Reece semi-empirical formulations, according to the discrete element method, multibody dynamics, the finite element method, and their coupled high-fidelity frameworks. Emerging topics included dynamic sinkage prediction, terrain-parameter calibration, multisource sensing, online slip-ratio and sinkage estimation, as well as data-driven control. Patent activity closely followed academic hotspots. But much emphasis was placed on track and grouser structures, ground-pressure measurement and regulation, anti-slip devices, state sensing, and implementable control systems. The field was shifting from isolated performance assessment to integrated modelling, sensing, and adaptive decision-making. Semi-empirical models remained valuable for preliminary design and real-time computation, but their homogeneous-terrain, steady-state, and simplified-contact assumptions constrain prediction accuracy under spatially variable moisture, straw-covered surfaces, repeated compaction, slopes, obstacles, and transient operation. Future research should develop multiscale representations of heterogeneous terrain, transferable macro–micro parameter-calibration, physics-informed machine learning, and coordinated optimization of track structure, vehicle configuration, sensing, and control. Tracked mobile platforms can be expected for agricultural mechanization in hilly and mountainous areas, even planetary and underwater extreme environments. The findings can also provide the mechanical basis to evaluate the trafficability, tractive performance, and stability of agricultural machinery.

The feeding and conveying device for combine harvester is often prone to blockage during field harvesting operations, which is why it is very important to study the working load of its drive shaft. To obtain the true load of the drive shaft of the feeding and conveying device for combine harvester under typical field operating conditions, the domestic 12 kg/s wheeled combine harvester was taken as the test object and a field load test system was constructed based on the perception of strain torque sensor and speed sensor. Through the field load tests, the results showed that the maximum power consumption of the drive shaft was 4.06 kW under the harvesting operation conditions and the average power consumption was 1.56 kW. Based on the rainflow counting method, the two-dimensional extrapolation load spectrum of the key power input section of the drive shaft under typical operating conditions was compiled by using the matrix cyclic period extrapolation method, which can provide loading conditions for the fatigue life test of the drive shaft of the feeding and conveying device for combine harvester. This method will provide a basis for the power matching of the working device of the harvester and a reference for the anti-fatigue design of the transmission system of the combine harvester.