| 基于深度学习改进的机器人轨迹规划算法 |
| |
| 引用本文： | 刘 玉,邓 琛,李文帅,韩宝磊.基于深度学习改进的机器人轨迹规划算法[J].教育技术导刊,2020,19(6):15-18. |
| |
| 作者姓名： | 刘 玉 邓 琛 李文帅 韩宝磊 |
| |
| 作者单位： | 1. 上海工程技术大学 电子电气工程学院，上海 201620；2.浙江慧勤医疗器械有限公司，浙江 德清 313200 |
| |
| 基金项目： | 国家自然科学基金项目（61701295） |
| |
| 摘 要： | 在公路环境巡逻机器人轨迹规划问题中，实时准确的交通流量预测对机器人轨迹规划尤为重要。然而由于车流量的随机非线性，使得机器人轨迹规划任务仍然充满挑战。提出一种深度神经网络与轨迹规划算法相结合的融合算法。通过深度学习预测短期交通流量，优化交通网络图并运用轨迹规划算法完成路径规划。实验表明，改进的机器人能够更快、更安全地完成道路巡逻任务。
|
| 关 键 词： | 机器人 深度学习 融合算法 优化网络图 轨迹规划 |
| 收稿时间： | 2019-08-06 |
Robot Trajectory Planning Optimization Algorithm Based on Deep Learning |
| |
| LIU Yu,DENG Chen,LI Wen-shuai,HAN Bao-lei.Robot Trajectory Planning Optimization Algorithm Based on Deep Learning[J].Introduction of Educational Technology,2020,19(6):15-18. |
| |
| Authors: | LIU Yu DENG Chen LI Wen-shuai HAN Bao-lei |
| |
| Institution: | 1. School of Electronic and Electrical Engineering，Shanghai University of Engineering Science，Shanghai 201620，China； 2. Zhejiang Huiqin Medical Devices Ltd，Deqing 313200，China |
| |
| Abstract: | In highway environment，real-time and accurate traffic flow prediction is particularly important for the trajectory planning of patrol robots. However，due to the stochastic nonlinearity of traffic flow，the task of trajectory planning of patrol robots is still full of challenges. In this paper，a fusion algorithm based on depth neural network and traditional trajectory planning algorithm is proposed. The short-term traffic flow is forecasted by in-depth learning to optimize the traffic network graph，and the trajectory planning algorithm is used to complete the path planning in the optimized network graph. Experiments show that the robot can complete the road patrol task faster and safer. |
| |
| Keywords: | robot deep learning fusion algorithm optimized network graph trajectory planning algorithm |
|
| 点击此处可从《教育技术导刊》浏览原始摘要信息 |
| 点击此处可从《教育技术导刊》下载免费的PDF全文 |
