基于卷积神经网络的自动反击的研发设计  

Automatic Counterattack Based on Convolution Neural Network

作　　者：卫超奇 李威 Wei Chaoqi;Li Wei(School of Mechanical Engineering,Taiyuan University of Science and Technology,Taiyuan City,Shanxi Province 030024,China)

机构地区：[1]太原科技大学机械工程学院,山西省太原市030024

出　　处：《农业装备与车辆工程》2020年第5期139-141,共3页Agricultural Equipment & Vehicle Engineering

摘　　要：为满足ROBOMASTER机甲大师机器人比赛的需要,针对在复杂环境下对具有不同特征的物体进行快速的识别与定位,从总体的设计思路、识别模块、运动控制、通信模块等几个主要部分详细介绍了设计原理与具体的实现方法,设计了一套可以实现机器人自动反击的应用程序。该程序能够满足提高机器人射击精准度的性能需求。In order to meet the needs of robot competition,aiming at the rapid recognition and positioning of objects with different characteristics in complex environments,the team designed a set of application programs,which can meet higher performance requirements.This paper introduces the principle and concrete implementation method of this design in detail in the aspects of overall design ideas,recognition module,motion control,communication module and other major parts.

关 键 词：卷积神经网络 深度学习 串口通信 

分 类 号：TP303[自动化与计算机技术—计算机系统结构]