SafeSteps: Learning Safer Footstep Planning Policies for Legged Robots via Model-Based Priors

被引:2
|
作者
Omar, Shafeef [1 ]
Amatucci, Lorenzo [1 ]
Barasuol, Victor [1 ]
Turrisi, Giulio [1 ]
Semini, Claudio [1 ]
机构
[1] Ist Italiano Tecnol IIT, Dynam Legged Syst Lab, Genoa, Italy
来源
2023 IEEE-RAS 22ND INTERNATIONAL CONFERENCE ON HUMANOID ROBOTS, HUMANOIDS | 2023年
关键词
PREDICTIVE CONTROL; LOCOMOTION; TIME;
D O I
10.1109/HUMANOIDS57100.2023.10375218
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
We present a footstep planning policy for quadrupedal locomotion that is able to directly take into consideration a-priori safety information in its decisions. At its core, a learning process analyzes terrain patches, classifying each landing location by its kinematic feasibility, shin collision, and terrain roughness. This information is then encoded into a small vector representation and passed as an additional state to the footstep planning policy, which furthermore proposes only safe footstep location by applying a masked variant of the Proximal Policy Optimization algorithm. The performance of the proposed approach is shown by comparative simulations and experiments on an electric quadruped robot walking in different rough terrain scenarios. We show that violations of the above safety conditions are greatly reduced both during training and the successive deployment of the policy, resulting in an inherently safer footstep planner. Furthermore, we show how, as a byproduct, fewer reward terms are needed to shape the behavior of the policy, which in return is able to achieve both better final performances and sample efficiency.
引用
收藏
页数:8
相关论文
共 7 条
  • [1] Constrained footstep planning using model-based reinforcement learning in virtual constraint-based walking
    Jin, Takanori
    Kobayashi, Taisuke
    Matsubara, Takamitsu
    ADVANCED ROBOTICS, 2024, 38 (08) : 525 - 545
  • [2] Concepts of model-based control and trajectory planning for parallel robots
    Belda, Kvetoslav
    Boehm, Josef
    Pisa, Pavel
    PROCEEDINGS OF THE 13TH IASTED INTERNATIONAL CONFERENCE ON ROBOTICS AND APPLICATIONS/PROCEEDINGS OF THE IASTED INTERNATIONAL CONFERENCE ON TELEMATICS, 2007, : 15 - +
  • [3] Using First Principles for Deep Learning and Model-Based Control of Soft Robots
    Johnson, Curtis C.
    Quackenbush, Tyler
    Sorensen, Taylor
    Wingate, David
    Killpack, Marc D.
    FRONTIERS IN ROBOTICS AND AI, 2021, 8
  • [4] Mode-constrained Model-based Reinforcement Learning via Gaussian Processes
    Scannell, Aidan
    Ek, Carl Henrik
    Richards, Arthur
    INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE AND STATISTICS, VOL 206, 2023, 206
  • [5] Model-Based Chance-Constrained Reinforcement Learning via Separated Proportional-Integral Lagrangian
    Peng, Baiyu
    Duan, Jingliang
    Chen, Jianyu
    Li, Shengbo Eben
    Xie, Genjin
    Zhang, Congsheng
    Guan, Yang
    Mu, Yao
    Sun, Enxin
    IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2024, 35 (01) : 466 - 478
  • [6] Fast Human-in-the-Loop Control for HVAC Systems via Meta-Learning and Model-Based Offline Reinforcement Learning
    Chen, Liangliang
    Meng, Fei
    Zhang, Ying
    IEEE TRANSACTIONS ON SUSTAINABLE COMPUTING, 2023, 8 (03): : 504 - 521
  • [7] Data-driven distributed formation control of under-actuated unmanned surface vehicles with collision avoidance via model-based deep reinforcement learning
    Pan, Chao
    Peng, Zhouhua
    Liu, Lu
    Wang, Dan
    OCEAN ENGINEERING, 2023, 267