动态窗口法引导的TD3无地图导航算法

柳佳乐,薛雅丽,崔闪,洪君

TD3 mapless navigation algorithm guided by dynamic window approach

Jiale LIU,Yali XUE,Shan CUI,Jun HONG

表 2 训练所得模型的平均奖励值

Tab.2 Average reward value of trained model

方法 成功率 步数 奖励 PPO 0.76 44.29 31.92 DDPG 0.87 43.88 67.33 TD3 0.90 52.27 55.71 DWA-TD3 0.87 35.93 63.96 LSTM-TD3 0.91 44.75 60.07 DWA-LSTM TD3 0.91 36.89 70.19