Memory Clustering Using Persistent Homology for Multimodality- and Discontinuity-Sensitive Learning of Optimal Control Warm-Starts

Merkt, Wolfgang; Ivan, Vladimir; Dinev, Traiko; Havoutis, Ioannis; Vijayakumar, Sethu

doi:10.1109/tro.2021.3069132

Cited by 10 publications

(4 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While this did not result in collisions in our experiments, the phenomenon is evident in trajectories such as Figures 11c and 12c rather than planning to travel on the other side of the static obstacle to the human, the re-planned trajectory avoids the human but stays within the same homotopy class. To address this, one could consider maintaining and optimising multiple trajectories at a time in different homotopy classes, such as work by Kolur et al (2019) and Merkt et al (2021), however, this is likely to increase the planning time and limit the robot's ability to react.…”

Section: Discussionmentioning

confidence: 99%

Motion Planning in Dynamic Environments Using Context-Aware Human Trajectory Prediction

Finean¹,

Petrović²,

Merkt³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Over the years, the separate fields of motion planning, mapping, and human trajectory prediction have advanced considerably. However, the literature is still sparse in providing practical frameworks that enable mobile manipulators to perform whole-body movements and account for the predicted motion of moving obstacles. Previous optimisationbased motion planning approaches that use distance fields have suffered from the high computational cost required to update the environment representation. We demonstrate that GPU-accelerated predicted composite distance fields significantly reduce the computation time compared to calculating distance fields from scratch. We integrate this technique with a complete motion planning and perception framework that accounts for the predicted motion of humans in dynamic environments, enabling reactive and pre-emptive motion planning that incorporates predicted motions. To achieve this, we propose and implement a novel human trajectory prediction method that combines intention recognition with trajectory optimisation-based motion planning. We validate our resultant framework on a real-world Toyota Human Support Robot (HSR) using live RGB-D sensor data from the onboard camera. In addition to providing analysis on a publicly available dataset, we release the Oxford Indoor Human Motion (Oxford-IHM) dataset and demonstrate stateof-the-art performance in human trajectory prediction. The Oxford-IHM dataset is a human trajectory prediction dataset in which people walk between regions of interest in an indoor environment. Both static and robot-mounted RGB-D cameras observe the people while tracked with a motion-capture system.

show abstract

Section: Discussionmentioning

confidence: 99%

Motion Planning in Dynamic Environments Using Context-Aware Human Trajectory Prediction

Finean¹,

Petrović²,

Merkt³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…using active learning techniques [209]. Another promising approach that has been recently investigated is to learn good initial guesses to speed up the convergence of OCP solvers [164], [210]. Finally, another method to help OCP solvers is to learn an approximation of the Value function, which can then be used as a terminal cost to guide the solver [106], [165].…”

Section: A Main Trendsmentioning

confidence: 99%

Optimization-Based Control for Dynamic Legged Robots

Wensing,

Posa,

et al. 2024

IEEE Trans. Robot.

View full text Add to dashboard Cite

In a world designed for legs, quadrupeds, bipeds, and humanoids have the opportunity to impact emerging robotics applications from logistics, to agriculture, to home assistance. The goal of this survey is to cover the recent progress toward these applications that has been driven by model-based optimization for the real-time generation and control of movement. The majority of the research community has converged on the idea of generating locomotion control laws by solving an optimal control problem (OCP) in either a model-based or data-driven manner. However, solving the most general of these problems online remains intractable due to complexities from intermittent unidirectional contacts with the environment, and from the many degrees of freedom of legged robots. This survey covers methods that have been pursued to make these OCPs computationally tractable, with specific focus on how environmental contacts are treated, how the model can be simplified, and how these choices affect the numerical solution methods employed. The survey focuses on model-based optimization while paving its way for broader combination with learning-based formulations to accelerate progress in this growing field.

show abstract

“…Overall, the main contributions of this work lie in 1) defining suitable local representations of motion planning problems, 2) learning a similarity function over them, and 3) applying it in the motion planning problem through our new framework. Although FIRE is tailored to retrieval frameworks that use local features and biased sampling distributions [11], [12] we believe it could be easily adapted to work with other retrieval-based methods [13]- [15].…”

Section: Similar Dissimilarmentioning

confidence: 99%

“…Motion planning is sensitive to input; small changes in W, x START , or x GOAL can drastically alter the resulting solution [12], [14], [32]. Furthermore, this mapping is usually multi-modal, since a motion planning problem may have multiple solution paths or multiple disjoint "challenging regions" [15], [33]. For these reasons, some approaches have adopted retrievalbased methods, also known as library- [34] or memorybased [35] methods.…”

Section: Related Workmentioning

confidence: 99%

Learning to Retrieve Relevant Experiences for Motion Planning

Chamzas¹,

Cullen²,

Shrivastava³

et al. 2022

Preprint

View full text Add to dashboard Cite

Recent work has demonstrated that motion planners' performance can be significantly improved by retrieving past experiences from a database. Typically, the experience database is queried for past similar problems using a similarity function defined over the motion planning problems. However, to date, most works rely on simple hand-crafted similarity functions and fail to generalize outside their corresponding training dataset. To address this limitation, we propose (FIRE), a framework that extracts local representations of planning problems and learns a similarity function over them. To generate the training data we introduce a novel self-supervised method that identifies similar and dissimilar pairs of local primitives from past solution paths. With these pairs, a Siamese network is trained with the contrastive loss and the similarity function is realized in the network's latent space. We evaluate FIRE on an 8-DOF manipulator in five categories of motion planning problems with sensed environments. Our experiments show that FIRE retrieves relevant experiences which can informatively guide sampling-based planners even in problems outside its training distribution, outperforming other baselines.

show abstract

Memory Clustering Using Persistent Homology for Multimodality- and Discontinuity-Sensitive Learning of Optimal Control Warm-Starts

Cited by 10 publications

References 36 publications

Motion Planning in Dynamic Environments Using Context-Aware Human Trajectory Prediction

Motion Planning in Dynamic Environments Using Context-Aware Human Trajectory Prediction

Optimization-Based Control for Dynamic Legged Robots

Learning to Retrieve Relevant Experiences for Motion Planning

Contact Info

Product

Resources

About