Anytime Multi-arm Task and Motion Planning for Pick-and-Place of Individual Objects via Handoffs

Abstract

Automation applications are pushing the deployment of many high DoF manipulators in warehouse and manufacturing environments. This has motivated many efforts on optimizing manipulation tasks involving a single arm. Coordinating multiple arms for manipulation, however, introduces additional computational challenges arising from the increased DoFs, as well as the combinatorial increase in the available operations that many manipulators can perform, including handoffs between arms. The focus here is on the case of pickand-place tasks, which require a sequence of handoffs to be executed, so as to achieve computational efficiency, asymptotic optimality and practical anytime performance. The paper leverages recent advances in multi-robot motion planning for high DoF systems to propose a novel multi-modal extension of the dRRT * algorithm. The key insight is that, instead of naively solving a sequence of motion planning problems, it is computationally advantageous to directly explore the composite space of the integrated multi-arm task and motion planning problem, given input sets of possible pick and handoff configurations. Asymptotic optimality guarantees are possible by sampling additional picks and handoffs over time. The evaluation shows that the approach finds initial solutions fast and improves their quality over time. It also succeeds in finding solutions to harder problem instances relative to alternatives and can scale effectively as the number of robots increases.

GRSTAPS: Graphically Recursive Simultaneous Task Allocation, Planning, and Scheduling

“…This work, however, is only demonstrated with a single robot and a single human. Shome and Bekris (2019) developed a framework for TAMP with multiple arms and anytime behavior. The framework is designed specifically for pick-handoff-place tasks and is not generalizable to other types of tasks or types of robots.…”

“…Prior techniques that address multi-agent TAMP problems are limited in multiple areas. First, all prior work is either limited to a specific problem domain (Motes et al, 2020;Toussaint and Lopes, 2017) or a particular type of robot (Akbari et al, 2019;Shome and Bekris, 2019;Umay et al, 2019). In addition, none of these approaches can handle tightly coupled problems with joint actions.…”

“…Although the four questions of coordination have been studied extensively within the context of homogeneous robots (e.g., Bredstro ¨m and Ro ¨nnqvist, 2008;Irnich et al, 2014;Shome and Bekris, 2019;Torren ˜o et al, 2017), these questions require further investigation within the context of heterogeneous teams. Perhaps owing to the complexity of the questions and, more importantly, their interactions, existing approaches to heterogeneous multi-robot coordination have resorted to defining and solving problems that focus on some subset of the four questions.…”

V-MAO: Generative Modeling for Multi-Arm Manipulation of Articulated Objects

“…Previous works have investigated using decentralized motion planner to avoid collision in multi-arm robot system [1]. Previous works have also investigated using multiple robots in the application of construction assembly [2], pick-and-place [21], and tabletop rearrangement [22]. Our work is an application of using multiple robot arms to manipulate articulated objects where we focus on learning the contact point distribution on object parts.…”

Path Planning for Multi-Arm Manipulators Using Deep Reinforcement Learning: Soft Actor–Critic with Hindsight Experience Replay

“…Hence, if the start and goal locations of the end-effector vary owing to task change, this procedure has to be repeated. Moreover, in the case of multi-arm manipulators, such a procedure becomes much more difficult [ 3 , 4 ]. Hence, it is important to make such a procedure carried out automatically.…”