In campus security management,traditional regular fixed point patrols and camera monitoring are difficult to cover the entire campus,while drone monitoring can make up for the aforementioned shortcomings.Due to the difficulty in ensuring the timeli-ness of information in current unmanned aerial vehicle path planning algorithms,a path planning algorithm based on deep Q-network has been proposed.The experimental results show that the success rate of the deep Q-network increases with the increase of testing times,and ultimately stabilizes at around 0.79,which is higher than the trajectory planning algorithm and Q-learning algorithm based on information age.At the same time,the number of path inflection points in deep Q network planning is only 16,with an average in-formation age of 19 seconds,which is lower than other algorithms.In free space and densely built space,the success rates of the deep Q-network ultimately stabilized at around 0.99 and 0.86,respectively,with an average number of steps not exceeding 100.The a-bove results indicate that the unmanned aerial vehicle path planning algorithm based on deep Q-network can efficiently and stably a-chieve optimal path planning,and achieve real-time monitoring of campus security without dead spots.
关键词
校园安全/无人机/路径规划/深度Q网络/信道模型
Key words
campus security/drones/path planning/deep q-network/channel model
维普
万方数据