Motion Planning for Multi-Mobile-Manipulator Payload Transport Systems
01 Aug 2019-pp 1469-1474
TL;DR: A hierarchical approach is introduced to compute realtime collision-free motion plans for a formation of mobile manipulator robots and a convex decentralized model-predictive controller is formulated to plan collision- free trajectories for the formation ofMobile manipulators.
Abstract: In this paper, a kinematic motion planning algorithm for cooperative spatial payload manipulation is presented. A hierarchical approach is introduced to compute realtime collision-free motion plans for a formation of mobile manipulator robots. Initially, collision-free configurations of a deformable 2-D virtual bounding box are identified, over a planning horizon, to determine a convex workspace for the entire system. Then, 3-D payload configurations whose projections lie within the convex workspace are computed. Finally, a convex decentralized model-predictive controller is formulated to plan collision-free trajectories for the formation of mobile manipulators. Our work facilitates real-time motion planning for the system and is scalable in the number of robots. The algorithm is validated in simulated dynamic environments.
Citations
More filters
[...]
TL;DR: In this article, a decentralized motion planning and collision avoidance algorithm for multi-robot payload transport systems (PTS) is presented, which is a formation of loosely coupled nonholonomic robots that cooperatively transport a deformable payload.
Abstract: We present a decentralized motion planning and collision avoidance algorithm for multi-robot payload transport systems (PTS). A PTS is a formation of loosely coupled non-holonomic robots that cooperatively transport a deformable payload. Each PTS is constrained to navigate safely in a dynamic environment by inter-formation, environmental, and intra-formation collision avoidance. Real-time collision avoidance for such systems is challenging due to the deformability of formations and high dimensional multi-robot non-convex workspace. We resolve the above challenges by embedding workspaces defined by a multi-robot collision avoidance algorithm and multi-scale repulsive potential fields as constraints within a decentralized convex optimization problem. Specifically, we present two main steps to plan the motion of each formation. First, we compute collision-free multi-scale convex workspaces over a planning horizon using a combination of ORCA and repulsive potential fields. Subsequently, we compute the motion plans of formation over a horizon by proposing a novel formulation for collision avoidance, and we leverage a model predictive controller (MPC) to solve the problem. The results validate that our solution facilitates real-time navigation of formations and computationally scales well with an increase in the number of robots and formations used. The algorithm is validated through extensive preliminary simulations, experiments in the gazebo simulator, and a proof of concept using real robots.
1 citations
Posted Content•
[...]
TL;DR: A deep reinforcement learning (DRL) based multi-robot formation controller for the task of autonomous aerial human motion capture (MoCap), where the objective is to estimate the trajectory of body pose, and shape of a single moving person using multiple micro aerial vehicles.
Abstract: In this letter, we introduce a deep reinforcement learning (RL) based multi-robot formation controller for the task of autonomous aerial human motion capture (MoCap). We focus on vision-based MoCap, where the objective is to estimate the trajectory of body pose and shape of a single moving person using multiple micro aerial vehicles. State-of-the-art solutions to this problem are based on classical control methods, which depend on hand-crafted system and observation models. Such models are difficult to derive and generalize across different systems. Moreover, the non-linearity and non-convexities of these models lead to sub-optimal controls. In our work, we formulate this problem as a sequential decision making task to achieve the vision-based motion capture objectives, and solve it using a deep neural network-based RL method. We leverage proximal policy optimization (PPO) to train a stochastic decentralized control policy for formation control. The neural network is trained in a parallelized setup in synthetic environments. We performed extensive simulation experiments to validate our approach. Finally, real-robot experiments demonstrate that our policies generalize to real world conditions. Video Link: this https URL Supplementary: this https URL
Cites methods from "Motion Planning for Multi-Mobile-Ma..."
[...]
[...]
[...]
TL;DR: In this article, a distributed position-force control framework for multiple mobile manipulators in charge of achieving a tightly cooperative transportation task is presented, where position control works in combination with force to ensure that the most important manipulator achieves cooperative transportation accurately and compliantly.
Abstract: This paper presents a distributed position-force control framework for multiple mobile manipulators in charge of achieving a tightly cooperative transportation task. Since the effect of each robot is different in the whole system, a three-layer control framework is designed. For the first layer, mobile bases run distributed observer which uses global states. At the second layer, the position deviation is adopted to improve the accuracy of general manipulators. Then, position control works in combination with force to ensure that the most important manipulator achieves cooperative transportation accurately and compliantly. The designed controller is extensible, which suits not only for pure transportation tasks but can also be exploited in those cases where a closed kinematic chain is generated by multi-robots manipulations. An analysis of the proposed controller is validated by simulation with three UR5 manipulators mounted on differential driven mobile bases separately.
References
More filters
[...]
TL;DR: This paper proposes individual robot control laws defined with respect to the payload that stabilize the payload along three-dimensional trajectories and detail the design of a gripping mechanism attached to each quadrotor that permits autonomous grasping of the payload.
Abstract: In this paper, we consider the problem of controlling multiple quadrotor robots that cooperatively grasp and transport a payload in three dimensions.We model the quadrotors both individually and as a group rigidly attached to a payload. We propose individual robot control laws defined with respect to the payload that stabilize the payload along three-dimensional trajectories. We detail the design of a gripping mechanism attached to each quadrotor that permits autonomous grasping of the payload. An experimental study with teams of quadrotors cooperatively grasping, stabilizing, and transporting payloads along desired three-dimensional trajectories is presented with performance analysis over many trials for different payload configurations.
487 citations
Additional excerpts
[...]
[...]
TL;DR: The first motion planning methodology applicable to articulated, nonpoint nonholonomic robots with guaranteed collision avoidance and convergence properties is presented, based on a new class of nonsmooth Lyapunov functions and a novel extension of the navigation function method to account for nonpoint articulated robots.
Abstract: This paper presents the first motion planning methodology applicable to articulated, nonpoint nonholonomic robots with guaranteed collision avoidance and convergence properties. It is based on a new class of nonsmooth Lyapunov functions and a novel extension of the navigation function method to account for nonpoint articulated robots. The dipolar inverse Lyapunov functions introduced are appropriate for nonholonomic control and offer superior performance characteristics compared to existing tools. The new potential field technique uses diffeomorphic transformations and exploits the resulting point-world topology. The combined approach is applied to the problem of handling deformable material by multiple nonholonomic mobile manipulators in an obstacle environment to yield a centralized coordinating control law. Simulation results verify asymptotic convergence of the robots, obstacle avoidance, boundedness of object deformations, and singularity avoidance for the manipulators.
302 citations
"Motion Planning for Multi-Mobile-Ma..." refers background in this paper
[...]
[...]
TL;DR: This work builds on four methodologies developed for fixed-base manipulation and presents the extension of these methodologies to mobile manipulation systems and proposes a new decentralized control structure for cooperative tasks.
Abstract: Mobile manipulation capabilities are key to many new applications of robotics in space, underwater, construction, and service environments. This article discusses the ongoing effort at Stanford University for the development of multiple mobile manipulation systems and presents the basic models and methodologies for their analysis and control. This work builds on four methodologies we have previously developed for fixed-base manipulation: the Operational Space Formulation for task-oriented robot motion and force control; the Dextrous Dynamic Coordination of Macro/Mini structures for increased mechanical bandwidth of robot systems; the Augmented Object Model for the manipulation of objects in a robot system with multiple arms; and the Virtual Linkage Model for the characterization and control of internal forces in a multi-arm system. We present the extension of these methodologies to mobile manipulation systems and propose a new decentralized control structure for cooperative tasks. The article also discusses experimental results obtained with two holonomic mobile manipulation platforms we have designed and constructed at Stanford University. © 1996 John Wiley & Sons, Inc.
207 citations
[...]
TL;DR: The implementation of the distributed auction algorithm and sequential convex programming using model predictive control produces the swarm assignment and trajectory optimization (SATO) algorithm that transfers a swarm of robots or vehicles to a desired shape in a distributed fashion.
Abstract: This paper presents a distributed, guidance and control algorithm for reconfiguring swarms composed of hundreds to thousands of agents with limited communication and computation capabilities. This algorithm solves both the optimal assignment and collision-free trajectory generation for robotic swarms, in an integrated manner, when given the desired shape of the swarm without pre-assigned terminal positions. The optimal assignment problem is solved using a distributed auction assignment that can vary the number of target positions in the assignment, and the collision-free trajectories are generated using sequential convex programming. Finally, model predictive control is used to solve the assignment and trajectory generation in real time using a receding horizon. The model predictive control formulation uses current state measurements to resolve for the optimal assignment and trajectory. The implementation of the distributed auction algorithm and sequential convex programming using model predictive control produces the swarm assignment and trajectory optimization SATO algorithm that transfers a swarm of robots or vehicles to a desired shape in a distributed fashion. Once the desired shape is uploaded to the swarm, the algorithm determines where each robot goes and how it should get there in a fuel-efficient, collision-free manner. Results of flight experiments using multiple quadcopters show the effectiveness of the proposed SATO algorithm.
116 citations
"Motion Planning for Multi-Mobile-Ma..." refers methods in this paper
[...]
[...]
TL;DR: A formalism that exploits deformability during manipulation of soft objects by robot teams, formulated as a convex optimization problem in velocity space and incorporating constraints for both collision avoidance and shape maintenance is presented.
Abstract: This paper presents a formalism that exploits deformability during manipulation of soft objects by robot teams. A hybrid centralized/distributed approach restricts centralized planning to high-level global guidance of the object for consensus. Low-level control is thus delegated to the individual manipulator robots, which retain manipulation and collision avoidance guarantees by passing forces to one another through the object. A distributed receding horizon planner provides local control, formulated as a convex optimization problem in velocity space and incorporating constraints for both collision avoidance and shape maintenance. We demonstrate teams of mobile manipulators autonomously carrying various deformable objects.
101 citations
Related Papers (5)
[...]
[...]
[...]
[...]
[...]