Toward Autonomous Driving by Musculoskeletal Humanoids: A Study of Developed Hardware and Learning-Based Software

Kawaharazuka, Kento; Tsuzuki, K.; Koga, Yuya; Omura, Yusuke; Makabe, Tasuku; Shinjo, Koki; Onitsuka, Moritaka; Nagamatsu, Yuya; Asano, Yuki; Okada, Kei; Kawasaki, Koji; Inaba, Masayuki

doi:10.1109/mra.2020.2987805

Cited by 5 publications

(1 citation statement)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Subsequent to the DRC, research efforts aimed at enabling autonomous vehicle driving by humanoid robots have yielded several successful implementations [2,3]. Conversely, the ingress task-wherein a humanoid robot enters a ground vehicle-has been subject to less investigation.…”

Section: Introductionmentioning

confidence: 99%

Robust Humanoid Robot Vehicle Ingress with A Finite State Machine Integrated with Deep Reinforcement Learning

Wang,

Chen,

et al. 2023

Preprint

View full text Add to dashboard Cite

Ingress task is crucial in a humanoid robot's attempt to drive a land vehicle and reach its destination fast. Previous work is inefficient in granting robots the ability to enter a vehicle from random starting position and orientation or withstand elasticity in vehicles, which are both hard to model. Deep Reinforcement Learning (DRL) could be introduced to address these issues. Previous applications of DRL in humanoid control tend to use consistent reward terms for the whole control process, which is not suitable for the ingress task with many distinctive states. This letter proposes a novel Finite State Machine control method integrated with Deep Reinforcement Learning for the humanoid ingress task. It collects the robot's status at the end of each state and immediately adjusts its next move. It has 97 percent ingress success rate with random initial displacement and vehicle elasticity in simulation.

show abstract

Section: Introductionmentioning

confidence: 99%