Towards High Efficient Long-Horizon Planning With Expert-Guided Motion-Encoding Tree Search

Zhou，Tong1; Lyu，Erli2; Cen，Guangdu3; Zha，Ziqi3; Qi，Senmao3; Wang，Jiaole3; Meng，Max Q.H.4,5,6

2024-07-01

10.1109/LRA.2024.3401683

IEEE Robotics and Automation Letters   影响因子和分区

2377-3766

Autonomous driving holds promise for increased safety, optimized traffic management, and a new level of convenience in transportation. While model-based reinforcement learning approaches such as MuZero enables long-term planning, the exponentially increase of the number of search nodes as the tree goes deeper significantly effect the searching efficiency. To deal with this problem, in this letter we proposed the expert-guided motion-encoding tree search (EMTS) algorithm. EMTS extends the MuZero algorithm by representing possible motions with a comprehensive motion primitives latent space and incorporating expert policies to improve the searching efficiency. The comprehensive motion primitives latent space enables EMTS to sample arbitrary trajectories instead of raw action to reduce the depth of the search tree. And the incorporation of expert policies guided the search and training phases the EMTS algorithm to enable early convergence. In the experiment section, the EMTS algorithm is compared with other four algorithms in three challenging scenarios. The experiment result verifies the effectiveness and the searching efficiency of the proposed EMTS algorithm.

Autonomous driving expert guided learning Monte Carlo tree search motion primitive

SCI ; EI

英语

其他

20242116123858

Efficiency ; Encoding (symbols) ; Job analysis ; Signal encoding ; Trees (mathematics)

Information Theory and Signal Processing:716.1 ; Data Processing and Image Processing:723.2 ; Artificial Intelligence:723.4 ; Production Engineering:913.1 ; Combinatorial Mathematics, Includes Graph Theory, Set Theory:921.4

2-s2.0-85193476565

Scopus

1.The Chinese University of Hong Kong,Department of Electronic Engineering,Hong Kong,999077,Hong Kong
2.Macao Polytechnic University,Faculty of Applied Science,999078,Macao
3.Harbin Institute of Technology (Shenzhen),School of Mechanical Engineering and Automation,Shenzhen,518055,China
4.Southern University of Science and Technology,Department of Electronic and Electrical Engineering,Shenzhen,518055,China
5.The Chinese University of Hong Kong,Department of Electronic Engineering,Hong Kong
6.The Chinese University of Hong Kong,Shenzhen Research Institute,Shenzhen,518057,China

Zhou，Tong,Lyu，Erli,Cen，Guangdu,et al. Towards High Efficient Long-Horizon Planning With Expert-Guided Motion-Encoding Tree Search[J]. IEEE Robotics and Automation Letters,2024,9(7):6272-6279.

Zhou，Tong.,Lyu，Erli.,Cen，Guangdu.,Zha，Ziqi.,Qi，Senmao.,...&Meng，Max Q.H..(2024).Towards High Efficient Long-Horizon Planning With Expert-Guided Motion-Encoding Tree Search.IEEE Robotics and Automation Letters,9(7),6272-6279.

Zhou，Tong,et al."Towards High Efficient Long-Horizon Planning With Expert-Guided Motion-Encoding Tree Search".IEEE Robotics and Automation Letters 9.7(2024):6272-6279.