Modular Robot Design Optimization with Generative Adversarial Networks

Hu, Jiaheng; Whitman, Julian; Travers, Matthew; Choset, Howie

doi:10.1109/icra46639.2022.9812091

Cited by 11 publications

(9 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To this end, more and more works propose novel techniques in pursuit of diversity (Gupta et al 2021;Medvet et al 2021;Hu et al 2022). Specifically, Hu et al (2022) verified the intrinsic potential of probabilistic generative models in proposing diverse robots, which we further dig into in this work.…”

Section: Diversity In Robot Design Automationmentioning

confidence: 52%

“…However, traditional EAs do not guarantee diversity; on the contrary, they tend to end up with similar solutions for a given task (Miras, Ferrante, and Eiben 2020). To this end, more and more works propose novel techniques in pursuit of diversity (Gupta et al 2021;Medvet et al 2021;Hu et al 2022). Specifically, Hu et al (2022) verified the intrinsic potential of probabilistic generative models in proposing diverse robots, which we further dig into in this work.…”

Section: Diversity In Robot Design Automationmentioning

confidence: 84%

“…Nonetheless, their integration into robot design automation remains relatively limited. Existing efforts have primarily concentrated on rigid robots, optimizing fixedtopology design parameters through inter-generation optimization (as opposed to free-form optimization), or mapping robot designs into continuous latent spaces for subsequent optimization using traditional EAs (Spielberg et al 2019(Spielberg et al , 2021Hu et al 2022;Hu, Whitman, and Choset 2023). Consequently, we contend that the potential of probabilistic generative models in the design automation of soft robot has yet to be fully harnessed.…”

Section: Robot Co-designmentioning

confidence: 99%

“…A considerable number of works have studied the design automation of VSRs, mostly with traditional Evolutionary Algorithms (EAs) that feature biologically inspired meta-heuristics (Cheney, Bongard, and Lipson 2015;Bhatia et al 2021;Medvet et al 2021). However, the substantial combinatorial search space of VSRs poses a big challenge to EAs and limits their performance; Generative encoding, on the other hand, has demonstrated promise as a solution Corucci et al 2018;Hu et al 2022). Nonetheless, the potential of probabilistic generative models, extensively applied in modeling complex ge-ometries (Gómez-Bombarelli et al 2018;Joy et al 2020;Xu et al 2023), remains largely untapped in the realm of soft robot design.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

MorphVAE: Advancing Morphological Design of Voxel-Based Soft Robots with Variational Autoencoders

Song,

Yang,

Peng

et al. 2024

AAAI

View full text Add to dashboard Cite

Soft robot design is an intricate field with unique challenges due to its complex and vast search space. In the past literature, evolutionary computation algorithms, including novel probabilistic generative models (PGMs), have shown potential in this realm. However, these methods are sample inefficient and predominantly focus on rigid robots in locomotion tasks, which limit their performance and application in robot design automation. In this work, we propose MorphVAE, an innovative PGM that incorporates a multi-task training scheme and a meticulously crafted sampling technique termed ``continuous natural selection'', aimed at bolstering sample efficiency. This method empowers us to gain insights from assessed samples across diverse tasks and temporal evolutionary stages, while simultaneously maintaining a delicate balance between optimization efficiency and biodiversity. Through extensive experiments in various locomotion and manipulation tasks, we substantiate the efficiency of MorphVAE in generating high-performing and diverse designs, surpassing the performance of competitive baselines.

show abstract

Section: Diversity In Robot Design Automationmentioning

confidence: 52%

Section: Diversity In Robot Design Automationmentioning

confidence: 84%

Section: Robot Co-designmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

MorphVAE: Advancing Morphological Design of Voxel-Based Soft Robots with Variational Autoencoders

Song,

Yang,

Peng

et al. 2024

AAAI

View full text Add to dashboard Cite

show abstract

“…Given task objectives, a robot configuration can be searched to complete the task with the highest performance. Generative Adversarial Networks (GANs) are used to learn to map one task to a distribution of configurations [40]. Motivated by the graph-like kinematics of robots, a global control policy of modular robots can be learned using Graph Neural Networks (GNNs) in which knowledge is shared among different configurations [41].…”

Section: E Work-life-flexibility Between Industry and Societymentioning

confidence: 99%

Metarobotics for Industry and Society: Vision, Technologies, and Opportunities

Kaigom

2024

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

Metarobotics aims to combine next generation wireless communication, multi-sense immersion, and collective intelligence to provide a pervasive, itinerant, and non-invasive access and interaction with distant robotized applications. Industry and society are expected to benefit from these functionalities. For instance, robot programmers will no longer travel worldwide to plan and test robot motions, even collaboratively. Instead, they will have a personalized access to robots and their environments from anywhere, thus spending more time with family and friends. Students enrolled in robotics courses will be taught under authentic industrial conditions in real-time. This paper describes objectives of Metarobotics in society, industry, and in-between. It identifies and surveys technologies likely to enable their completion and provides an architecture to put forward the interplay of key components of Metarobotics. Potentials for self-determination, self-efficacy, and work-life-flexibility in robotics-related applications in Society 5.0, Industry 4.0, and Industry 5.0 are outlined.

show abstract