Computational Design of Robotic Devices From High-Level Motion Specifications

Ha, Sehoon; Coros, Stelian; Alspach, Alexander; Bern, James M.; Kim, Joohyung; Yamane, Katsu

doi:10.1109/tro.2018.2830419

Cited by 57 publications

(41 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The most closely related methods for manipulator arrangement synthesis are best-first graph searches (Desai, Yuan, and Coros 2017;Desai et al 2018;Ha et al 2018). In these works, a heuristic was crafted that estimated the ability of each partially complete arrangement in fulfilling the task, and was used to guide a search over a tree of different arrangements.…”

Section: Related Workmentioning

confidence: 99%

“…We build on prior modular design synthesis methods (Desai, Yuan, and Coros 2017;Ha et al 2018) which incrementally construct and search a tree of different modular Figure 1: Our approach searches for modular manipulator designs by viewing the space of arrangements as a tree, where modules are sequentially added to the end of the robot. The arrangement at the root of the tree is a base mounting location.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Modular Robot Design Synthesis with Deep Reinforcement Learning

Whitman

Bhirangi

Travers

et al. 2020

AAAI

View full text Add to dashboard Cite

Modular robots hold the promise of versatility in that their components can be re-arranged to adapt the robot design to a task at deployment time. Even for the simplest designs, determining the optimal design is exponentially complex due to the number of permutations of ways the modules can be connected. Further, when selecting the design for a given task, there is an additional computational burden in evaluating the capability of each robot, e.g., whether it can reach certain points in the workspace. This work uses deep reinforcement learning to create a search heuristic that allows us to efficiently search the space of modular serial manipulator designs. We show that our algorithm is more computationally efficient in determining robot designs for given tasks in comparison to the current state-of-the-art.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Modular Robot Design Synthesis with Deep Reinforcement Learning

Whitman

Bhirangi

Travers

et al. 2020

AAAI

View full text Add to dashboard Cite

show abstract

“…Megaro et al present a design system that automates the tedious process of creating 3D-printable robotic creatures while allowing for customization for casual users [23]. Recent research focused on computational approaches for nonexpert users to design animated or robotic characters by highlevel motion specification such as motion curves [6,31,12]. LinkedIt enables end-user design and fabrication of robots, from creating linkages to exhibit specific motion path [2].…”

Section: Interactive Design Of Robotic Charactersmentioning

confidence: 99%

Romeo

Cui

Kim

et al. 2020

Proceedings of the 33rd Annual ACM Symposium on User Interface Software and Technology

View full text Add to dashboard Cite

show abstract

“…It is important to note that the use of a genetic algorithm is not the only possible approach to solve the presented issue. Patel [15] used a simulated annealing algorithm, while Ha [16] addressed the issue using a heuristic-guided tree search algorithm. Whitman [17] and Dogra [18] applied a deterministic quasi-Newton algorithm to find the local minimum of a cost function.…”

Section: Introductionmentioning

confidence: 99%

Genetic Optimization of a Manipulator: Comparison between Straight, Rounded, and Curved Mechanism Links

et al. 2021

View full text Add to dashboard Cite

There are several ubiquitous kinematic structures that are used in industrial robots, with the most prominent being a six-axis angular structure. However, researchers are experimenting with task-based mechanism synthesis that could provide higher efficiency with custom optimized manipulators. Many studies have focused on finding the most efficient optimization algorithm for task-based robot manipulators. These manipulators, however, are usually optimized from simple modular joints and links, without exploring more elaborate modules. Here, we show that link modules defined by small numbers of parameters have better performance than more complicated ones. We compare four different manipulator link types, namely basic predefined links with fixed dimensions, straight links that can be optimized for different lengths, rounded links, and links with a curvature defined by a Hermite spline. Manipulators are then built from these modules using a genetic algorithm and are optimized for three different tasks. The results demonstrate that manipulators built from simple links not only converge faster, which is expected given the fewer optimized parameters, but also converge on lower cost values.

show abstract

Computational Design of Robotic Devices From High-Level Motion Specifications

Cited by 57 publications

References 27 publications

Modular Robot Design Synthesis with Deep Reinforcement Learning

Modular Robot Design Synthesis with Deep Reinforcement Learning

Romeo

Genetic Optimization of a Manipulator: Comparison between Straight, Rounded, and Curved Mechanism Links

Contact Info

Product

Resources

About