Synergistic Task and Motion Planning With Reinforcement Learning-Based Non-Prehensile Actions
نویسندگان
چکیده
Robotic manipulation in cluttered environments requires synergistic planning among prehensile and non-prehensile actions. Previous works on sampling-based Task Motion Planning (TAMP) algorithms, e.g. PDDLStream, provide a fast generalizable solution for multi-modal manipulation. However, they are likely to fail scenarios where no collision-free grasping approaches can be sampled without preliminary manipulations. To extend the ability of we integrate vision-based Reinforcement Learning (RL) procedure, pusher . The pushing actions generated by eliminate interlocked situations make problem solvable. Also, algorithm evaluates providing rewards training process, thus learn avoid leading irreversible failures. proposed hybrid method is validated bin-picking implemented both simulation real world. Results show that effectively improve success ratio previous algorithm, while help skills.
منابع مشابه
Decidability of Semi-Holonomic Prehensile Task and Motion Planning
In this paper, we define semi-holonomic controllability (SHC) and a general task and motion planning framework. We give a perturbation algorithm that can take a prehensile task and motion planning (PTAMP) domain and create a jointly-controllable-open (JC-open) variant with practically identical semantics. We then present a decomposition-based algorithm that computes the reachability set of a pr...
متن کاملTask Planning with Continuous Actions and Nondeterministic Motion Planning Queries
We present a new probabilistic tree-of-roadmaps (PTR) planner that integrates discrete task planning and continuous motion planning using a probabilistic forward search. PTR is explicitly designed with two characteristics in mind: 1) it allows tasks to take continuously variable parameters as input, such as the location of contact points for a robot manipulator executing an object pickup task; ...
متن کاملDiGrad: Multi-Task Reinforcement Learning with Shared Actions
Most reinforcement learning algorithms are inefficient for learning multiple tasks in complex robotic systems, where different tasks share a set of actions. In such environments a compound policy may be learnt with shared neural network parameters, which performs multiple tasks concurrently. However such compound policy may get biased towards a task or the gradients from different tasks negate ...
متن کاملRobust Reinforcement Learning in Motion Planning
While exploring to find better solutions, an agent performing online reinforcement learning (RL) can perform worse than is acceptable. In some cases, exploration might have unsafe, or even catastrophic, results, often modeled in terms of reaching 'failure' states of the agent's environment. This paper presents a method that uses domain knowledge to reduce the number of failures during explorati...
متن کاملCombining Planning with Reinforcement Learning for Multi-robot Task Allocation
We describe an approach to the multi-robot task allocation (MRTA) problem in which a group of robots must perform tasks that arise continuously, at arbitrary locations across a large space. A dynamic scheduling algorithm is derived in which proposed plans are evaluated using a combination of short-term lookahead and a value function acquired by reinforcement learning. We demonstrate that this d...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: IEEE robotics and automation letters
سال: 2023
ISSN: ['2377-3766']
DOI: https://doi.org/10.1109/lra.2023.3261708