面向轨道智能交通大学生创新训练实践教学平台设计

Design of Practical Teaching Platform for Innovation and Entrepreneurship Training for College Students in Rail Intelligent Transportation

以轨道智能交通中的入侵异物智能检测为对象,设计了一种新型的大学生创新训练实践教学平台,提出了一种基于深度学习的轨道异物入侵检测模型,并结合人工智能和机械设计等交叉学科专业知识,自主研发了一种小型轨道异物入侵智能检测系统物理样机,实现了复杂环境下的高效轨道异物入侵检测。实验和实践结果表明,该模型很好地平衡了检测速度和精度,在NVIDIA GTX1080Ti平台上对自建轨道异物入侵数据集的平均检测精度为96.1%,检测速度为209 FPS。通过上述大创实践教学平台的设计、实施和考核,培养和激发了大学生的创新实践和团队协作能力。

In response to the requirements of experimental teaching reform in the innovation and entrepreneurship training practice project for college students,taking the intelligent detection of invading foreign objects in rail intelligent transportation as an opportunity,a new practical teaching platform for college students’innovative training is designed,and a deep learning-based detection of foreign object intrusion on rails is proposed.By combining interdisciplinary knowledge such as artificial intelligence and mechanical design,a prototype is independently developed to detect a small foreign object intrusion.The prototype achieves efficient detection for foreign object intrusion on rails in complex environments.Through the design,implementation and assessment of the above-mentioned practical teaching platform,the innovative practical ability of college students is cultivated and stimulated,and has a significant effect on improving the innovative practical ability and teamwork ability of college students.