Agricultural machinery intelligence is the inevitable direction of agricultural machinery design, and the systems in these designs are important tools. In this paper, to address the problem of low processing power of traditional agricultural machinery design systems in analyzing data, such as fit, tolerance, interchangeability, and the assembly process, as well as to overcome the disadvantages of the high cost of intelligent design modules, lack of data compatibility, and inconsistency between modules, a novel agricultural machinery intelligent design system integrating image processing and knowledge reasoning is constructed. An image-processing algorithm and trigger are used to detect the feature parameters of key parts of agricultural machinery and build a virtual prototype. At the same time, a special knowledge base of agricultural machinery is constructed to analyze the test data of the virtual prototype. The results of practical application and software evaluation of third-party institutions show that the system improves the efficiency of intelligent design in key parts of agricultural machinery by approximately 20%, reduces the operation error rate of personnel by approximately 40% and the consumption of computer resources by approximately 30%, and greatly reduces the purchase cost of intelligent design systems to provide a reference for intelligent design to guide actual production.
The agricultural machinery experiment is restricted by the crop production season. Missing the crop growth cycle will extend the machine development period. The use of virtual reality technology to complete preassembly and preliminary experiments can reduce the loss caused by this problem. To improve the intelligence and stability of virtual assembly, this paper proposed a more stable dynamic gesture cognition framework: the TCP/IP protocol constituted the network communication terminal, the leap motion-based vision system constituted the gesture data collection terminal, and the CNN-LSTM network constituted the dynamic gesture recognition classification terminal. The dynamic gesture recognition framework and the harvester virtual assembly platform formed a virtual assembly system to achieve gesture interaction. Through experimental analysis, the improved CNN-LSTM network had a small volume and could quickly establish a stable and accurate gesture recognition model with an average accuracy of 98.0% (±0.894). The assembly efficiency of the virtual assembly system with the framework was improved by approximately 15%. The results showed that the accuracy and stability of this model met the requirements, the corresponding assembly parts were robust in the virtual simulation environment of the whole machine, and the harvesting behaviour in the virtual reality scene was close to the real scene. The virtual assembly system under this framework provided technical support for unmanned farms and virtual experiments on agricultural machinery.
To address the difficulty of obtaining the optimal driving strategy under the condition of a complex environment and changeable tasks of vehicle autonomous driving, this paper proposes an end-to-end autonomous driving strategy learning method based on deep reinforcement learning. The ideas of target attraction and obstacle rejection of the artificial potential field method are introduced into the distributed proximal policy optimization algorithm, and the APF-DPPO learning model is established. To solve the range repulsion problem of the artificial potential field method, which affects the optimal driving strategy, this paper proposes a directional penalty function method that combines collision penalty and yaw penalty to convert the range penalty of obstacles into a single directional penalty, and establishes the vehicle motion collision model. Finally, the APF-DPPO learning model is selected to train the driving strategy for the virtual vehicle, and the transfer learning method is selected to verify the comparison experiment. The simulation results show that the completion rate of the virtual vehicle in the obstacle environment that generates penalty feedback is as high as 96.3%, which is 3.8% higher than the completion rate in the environment that does not generate penalty feedback. Under different reward functions, the method in this paper obtains the highest cumulative reward value within 500 s, which improves 69 points compared with the reward function method based on the artificial potential field method, and has higher adaptability and robustness in different environments. The experimental results show that this method can effectively improve the efficiency of autonomous driving strategy learning and control the virtual vehicle for autonomous driving behavior decisions, and provide reliable theoretical and technical support for real vehicles in autonomous driving decision-making.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.