本研究利用機器學習找出能即時避開移動物體的路徑規劃演算法，並導入優先度參數，讓複數台不同層級的機器人能完成各自的任務。藉由深度狀態價值函數學習(Deep V learning)，學習並產生用來協助機器人選擇動作的狀態價值深度神經網路。利用機器人可以獲得的參數和到達終點的時間，建立深度神經網路。接著設定獎勵條件和移動規則，並使系統在模擬中反覆執行，同時蒐集路徑資料，學習並更新深度神經網路。最後利用訓練完成的深度神經網路，來協助機器人判斷該選擇執行何種動作。利用本研究的方法能達成多台機器人同時執行不同層級任務的路徑規劃，達成比一般的演算法更佳的效能。

This study uses machine learning methods to find a trajectory planning algorithm for the multiple-robot system with hierarchy. By introducing priority parameters during the learning process, multiple robots can perform their respective tasks according to the hierarchy system. The deep neural network is designed using the information obtained from sensors on the robot. To reduce the computation time, the neural network is trained initially using the data generated by A* algorithm. By setting proper rewards and rules and run the simulations with different boundary conditions, the network is able to evolve itself with the data. The learned state-value deep neural network is then used to determine the control action for each of the robot so that the tasks can be accomplished in a prioritized manner. Both simulations and experiments verify that the proposed approach can make multiple robots with hierarchy move in a more efficient way.

[1] [Online].Available:http://planning.cs.uiuc.edu/node659.html[Accessed 13 November 2018]
[2] [Online].Available: http://wiki.ros.org/amcl [Accessed 30 November 2018]
[3] Dijkstra, E. W. (1959). A note on two problems in connexion with graphs. Numerische mathematik, 1(1), 269-271.
[4] Hart, P. E., Nilsson, N. J., & Raphael, B. (1968). A formal basis for the heuristic determination of minimum cost paths. IEEE transactions on Systems Science and Cybernetics, 4(2), 100-107.
[5] Kuswadi, S., Santoso, J. W., Tamara, M. N., & Nuh, M. (2018, October). Application SLAM and Path Planning using A-Star Algorithm for Mobile Robot in Indoor Disaster Area. In 2018 International Electronics Symposium on Engineering Technology and Applications (IES-ETA) (pp. 270-274). IEEE.
[6] Chen, T., Zhang, G., Hu, X., & Xiao, J. (2018, May). Unmanned aerial vehicle route planning method based on a star algorithm. In 2018 13th IEEE Conference on Industrial Electronics and Applications (ICIEA) (pp. 1510-1514). IEEE.
[7] Guruji, A. K., Agarwal, H., & Parsediya, D. K. (2016). Time-efficient A* algorithm for robot path planning. Procedia Technology, 23, 144-149.
[8] Zheng, Y., Wang, L., & Xi, P. (2018, August). Improved Ant Colony Algorithm for Multi-Agent Path Planning in Dynamic Environment. In 2018 International Conference on Sensing, Diagnostics, Prognostics, and Control (SDPC) (pp. 732-737). IEEE.
[9] Borenstein, J., & Koren, Y. (1991). The vector field histogram-fast obstacle avoidance for mobile robots. IEEE transactions on robotics and automation, 7(3), 278-288.
[10] Fox, D., Burgard, W., & Thrun, S. (1997). The dynamic window approach to collision avoidance. IEEE Robotics & Automation Magazine, 4(1), 23-33.
[11] Van Den Berg, J., Guy, S. J., Lin, M., & Manocha, D. (2011). Reciprocal n-body collision avoidance. In Robotics research (pp. 3-19). Springer, Berlin, Heidelberg.
[12] Yan, C., & Xiang, X. (2018, June). A Path Planning Algorithm for UAV Based on Improved Q-Learning. In 2018 2nd International Conference on Robotics and Automation Sciences (ICRAS) (pp. 1-5). IEEE.
[13] Chen, T., Zhang, G., Hu, X., & Xiao, J. (2018, May). Unmanned aerial vehicle route planning method based on a star algorithm. In 2018 13th IEEE Conference on Industrial Electronics and Applications (ICIEA) (pp. 1510-1514). IEEE.
[14] Chen, Y. F., Liu, M., Everett, M., & How, J. P. (2017, May). Decentralized non-communicating multiagent collision avoidance with deep reinforcement learning. In 2017 IEEE international conference on robotics and automation (ICRA) (pp. 285-292). IEEE.
[15] Chen, Y. F., Everett, M., Liu, M., & How, J. P. (2017, September). Socially aware motion planning with deep reinforcement learning. In 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (pp. 1343-1350). IEEE.
[16] Sui, Z., Pu, Z., Yi, J., & Tan, X. (2018, July). Path Planning of Multiagent Constrained Formation through Deep Reinforcement Learning. In 2018 International Joint Conference on Neural Networks (IJCNN) (pp. 1-8). IEEE.
[17] Sartoretti, G., Kerr, J., Shi, Y., Wagner, G., Kumar, T. S., Koenig, S., & Choset, H. (2019). PRIMAL: Pathfinding via reinforcement and imitation multi-agent learning. IEEE Robotics and Automation Letters, 4(3), 2378-2385.
[18] Xin, J., Zhao, H., Liu, D., & Li, M. (2017, October). Application of deep reinforcement learning in mobile robot path planning. In 2017 Chinese Automation Congress (CAC) (pp. 7112-7116). IEEE.
[19] Long, P., Fanl, T., Liao, X., Liu, W., Zhang, H., & Pan, J. (2018, May). Towards optimally decentralized multi-robot collision avoidance via deep reinforcement learning. In 2018 IEEE International Conference on Robotics and Automation (ICRA) (pp. 6252-6259). IEEE.
[20] Brock, O., & Khatib, O. (1999, May). High-speed navigation using the global dynamic window approach. In Proceedings 1999 IEEE International Conference on Robotics and Automation (Cat. No. 99CH36288C) (Vol. 1, pp. 341-346). IEEE.
[21] Siegwart, R., Nourbakhsh, I. R., & Scaramuzza, D. (2011). Introduction to autonomous mobile robots (pp. 81-88). MIT press.
[22] [Online].Available: http://wiki.ros.org/turtlesim [Accessed 30 November 2018]
[23] Menzies, T., & Hu, Y. (2003). Data mining for very busy people. Computer, 36(11), 22-29.
[24] Clatworthy, J., Buick, D., Hankins, M., Weinman, J., & Horne, R. (2005). The use and reporting of cluster analysis in health psychology: A review. British journal of health psychology, 10(3), 329-358.
[25] Bengio, Y. (2009). Learning deep architectures for AI. Foundations and trends® in Machine Learning, 2(1), 1-127.
[26] Hoshino, Y., & Kamei, K. (2003, August). A proposal of reinforcement learning system to use knowledge effectively. In SICE 2003 Annual Conference (IEEE Cat. No. 03TH8734) (Vol. 2, pp. 1582-1585). IEEE.
[27] Rennie, J. D., Shih, L., Teevan, J., & Karger, D. R. (2003). Tackling the poor assumptions of naive bayes text classifiers. In Proceedings of the 20th international conference on machine learning (ICML-03) (pp. 616-623).
[28] Zhang, W., Hongwu, Y. A. N. G., & Pengpeng, Z. H. I. (2018, November). Emotional speech synthesis based on DNN and PAD emotional state model. In 2018 11th International Symposium on Chinese Spoken Language Processing (ISCSLP) (pp. 41-45). IEEE.
[29] Huang, Q., Bao, C., Wang, X., & Xiang, Y. (2018, September). DNN-Based Speech Enhancement Using MBE Model. In 2018 16th International Workshop on Acoustic Signal Enhancement (IWAENC) (pp. 196-200). IEEE.
[30] Yanagisawa, H., Yamashita, T., & Watanabe, H. (2018, January). A study on object detection method from manga images using CNN. In 2018 International Workshop on Advanced Image Technology (IWAIT) (pp. 1-4). IEEE.
[31] Kido, S., Hirano, Y., & Hashimoto, N. (2018, January). Detection and classification of lung abnormalities by use of convolutional neural network (CNN) and regions with CNN features (R-CNN). In 2018 International Workshop on Advanced Image Technology (IWAIT) (pp. 1-4). IEEE.
[32] Zhao, C., Song, W., Liu, X., Liu, L., & Zhao, X. (2018, November). Research on Authorship Attribution of Article Fragments via RNNs. In 2018 IEEE 9th International Conference on Software Engineering and Service Science (ICSESS) (pp. 156-159). IEEE.
[33] Dutta, K., & Sarma, K. K. (2012, December). Multiple feature extraction for RNN-based assamese speech recognition for speech to text conversion application. In 2012 International Conference on Communications, Devices and Intelligent Systems (CODIS) (pp. 600-603). IEEE.
[34] [Online].Available: http://wiki.ros.org/gmapping [Accessed 30 November 2018]
[35] [Online].https://www.slamtec.com/cn/Lidar/A2[Accessed 11 August 2019]
[36] [Online].https://www.nvidia.com/zh-tw/autonomous-machines/embedded-systems/jetson-tx2/ [Accessed 11 August 2019]
[37] [Online].https://www.leadtek.com/cht/products/AI_HPC(37)/NVIDIA_Jetson_TX2(10782)/detail [Accessed 11 August 2019]
[38] [Online].https://www.st.com/en/evaluation-tools/nucleo-f446re.html#overview [Accessed 11 August 2019]