Notion of Control-Law Module and Modular Framework of Cooperative Transportation Using Multiple Nonholonomic Robotic Agents With Physical Rigid-Formation-Motion Constraints

Design and control of a multi-mobile-robot cooperative transport system based on a novel six degree-of-freedom connector

Multi-robot cooperative object transport on uneven roads is challenging. The key barrier is dealing with nonholonomic and rigid-formation motion constraints. In this study, to alleviate the influence of these constraints on a multi-robot cooperative transport system (MRCTS), a six degree-of-freedom connector capable of sensing three-axial displacements, three-axial forces, and three-axial angular displacements is designed and employed. Based on the local displacements derived from each connector, we develop a position calibration method to calculate the relative position of each robot and achieve a centralized control strategy. Based on the forces sensed by each connector, we design a decentralized control strategy to accomplish cooperative transport in which a leader robot guides the follower robots toward a destination by applying forces, instead of centralized information broadcasting. The experimental results show that the MRCTS works well on an uneven surface, and the tracking errors are within the design stroke of the connectors, demonstrating the effectiveness of the design and control methods of the MRCTS.

Distributed Hierarchical Shared Control for Flexible Multirobot Maneuver Through Dense Undetectable Obstacles

When teleoperating a multirobot system (MRS) in outdoor environments, human operators can often detect obstacles that are not detected by robots and spot emergencies faster than robots do. However, the lack of efficient methods for operators to manipulate an MRS has limited the number of robots in a human-robot team. To handle this problem, a distributed hierarchical shared control scheme is proposed, aiming to provide a safe and flexible control interface for a few human operators to interact with a large MRS. The proposed hierarchical control scheme employs a two-layered structure. In the upper layer, intention field networks are designed to generate virtual human control signals. Two functionalities for human teleoperation, called: 1) group management and 2) motion intervention, are realized using intention fields, allowing the operators to split the robot formation into different groups and steer individual robots away from immediate danger. In parallel, a blending-based shared control algorithm is designed in the lower layer to resolve the conflict between human intervention inputs and autonomous formation control signals. The input-to-output stability (IOS) of the proposed distributed hierarchical shared control scheme is proved by exploiting the properties of weighting functions. Results from a usability testing experiment and a physical experiment are also presented to validate the effectiveness and practicability of the proposed method.

Distributed Cooperative Control of Redundant Mobile Manipulators With Safety Constraints

In this article, the distributed cooperative control problem of redundant mobile manipulators is investigated. A novel method is proposed to solve the problem by integrating formation control with constrained optimization, which not only transports the object along a reference trajectory in a distributed manner but also obtains the dexterous joint postures and end-effector displacements under safety constraints for collision avoidance. For the constrained optimization, the cost function and safety constraints are designed to quantify the mobility and manipulability of mobile manipulators, and collision-free working ranges with the object and obstacles, respectively. A discontinuous projected primal-dual algorithm with damping terms is proposed to solve the constrained optimization problem, providing the joint postures and end-effector displacements, which minimize the cost function and satisfy safety constraints. For the formation control, a finite-time control law, guided by end-effector displacements from the primal-dual algorithm, is developed in order to transport the object by establishing a prescribed formation and moving its centroid to track the reference trajectory. The cooperative manipulation is therefore achieved by the proposed method, which is further validated through numerical simulations.

Roto-Translation Invariant Formation of Multiple Underactuated Planar Rigid Bodies

This article investigates the roto-translation invariant (RTI) formation of multiple underactuated planar rigid bodies, which are established under the framework of matrix Lie groups. The main contribution is that we define the RTI and pseudo RTI (P-RTI) formation of planar rigid bodies. Different from the common formation given in the earth-fixed frame, the RTI formation is defined in the body-fixed frame so that it possesses a rigid-body motion obtained by composing rotation and translation simultaneously. Moreover, regarding fully actuated planar rigid bodies, we propose the velocity and force requirements to maintain the RTI formation, which are derived based on the kinematic and dynamic model, respectively. Another contribution of this article is that the RTI formation feasibility is investigated for underactuated planar rigid bodies subject to nonholonomic constraints on velocities and accelerations. To be more specific, we study the occasions when wheeled mobile robots and underactuated surface vessels can maintain the RTI or P-RTI formation. Finally, the results of the simulation and experiment are presented so as to exhibit the RTI and P-RTI formation intuitively.

Attentive Relational State Representation in Decentralized Multiagent Reinforcement Learning

In multiagent reinforcement learning (MARL), it is crucial for each agent to model the relation with its neighbors. Existing approaches usually resort to concatenate the features of multiple neighbors, fixing the size and the identity of the inputs. But these settings are inflexible and unscalable. In this article, we propose an attentive relational encoder (ARE), which is a novel scalable feedforward neural module, to attentionally aggregate an arbitrary-sized neighboring feature set for state representation in the decentralized MARL. The ARE actively selects the relevant information from the neighboring agents and is permutation invariant, computationally efficient, and flexible to interactive multiagent systems. Our method consistently outperforms the latest competing decentralized MARL methods in several multiagent tasks. In particular, it shows strong cooperative performance in challenging StarCraft micromanagement tasks and achieves over a 96% winning rate against the most difficult noncheating built-in artificial intelligence bots.

Fully Adaptive Practical Time-Varying Output Formation Tracking for High-Order Nonlinear Stochastic Multiagent System With Multiple Leaders

Fully adaptive practical time-varying output formation tracking issues of high-order nonlinear stochastic multiagent systems with multiple leaders are researched, where the adaptive fuzzy-logic system (FLS) is introduced for estimating the mismatched integrated uncertain items. Distinctive with former results, stochastic noise is considered in the dynamics, and the followers are required for achieving the time-varying output formation tracking in probability of the convex combination of the leaders' outputs. First, a fully adaptive practical time-varying output formation tracking protocol is put forward, which only utilizes the neighboring relative information, and the global interaction topology information is not used. Besides, the designed protocol employs the adaptive FLSs to estimate the mismatched uncertainties of the followers and the leaders, and the uncertain boundary functions of the stochastic noise. Then, the design process of control protocol and parameter adaptive update law is summarized within four steps in an algorithm. Third, the stability and the properties of the proposed protocol and algorithm are analyzed by employing the Lyapunov theories and stochastic stability theories. Finally, numerical simulation results illustrate the effectiveness of achieved protocol and algorithm.

Event-Triggered Coordination for Formation Tracking Control in Constrained Space With Limited Communication

In this paper, the formation tracking control is studied for a multiagent system (MAS) with communication limitations. The objective is to control a group of agents to track a desired trajectory while maintaining a given formation in nonomniscient constrained space. The role switching triggered by the detection of unexpected spatial constraints facilitates efficiency of event-triggered control in communication bandwidth, energy consumption, and processor usage. A coordination mechanism is proposed based on a novel role "coordinator" to indirectly spread environmental information among the whole communication network and form a feedback link from followers to the leader to guarantee the formation keeping. A formation scaling factor is introduced to scale up or scale down the given formation size in the case that the region is impassable for MAS with the original formation size. Controllers for the leader and followers are designed and the adaptation law is developed for the formation scaling factor. The conditions for asymptotic stability of MAS are discussed based on the Lyapunov theory. Simulation results are presented to illustrate the performance of proposed approaches.

A New Conflict Resolution Method for Multiple Mobile Robots in Cluttered Environments With Motion-Liveness

This paper presents a new conflict resolution methodology for multiple mobile robots while ensuring their motion-liveness, especially for cluttered and dynamic environments. Our method constructs a mathematical formulation in a form of an optimization problem by minimizing the overall travel times of the robots subject to resolving all the conflicts in their motion. This optimization problem can be easily solved through coordinating only the robots' speeds. To overcome the computational cost in executing the algorithm for very cluttered environments, we develop an innovative method through clustering the environment into independent subproblems that can be solved using parallel programming techniques. We demonstrate the scalability of our approach through performing extensive simulations. Simulation results showed that our proposed method is capable of resolving the conflicts of 100 robots in less than 1.23 s in a cluttered environment that has 4357 intersections in the paths of the robots. We also developed an experimental testbed and demonstrated that our approach can be implemented in real time. We finally compared our approach with other existing methods in the literature both quantitatively and qualitatively. This comparison shows while our approach is mathematically sound, it is more computationally efficient, scalable for very large number of robots, and guarantees the live and smooth motion of robots.

Predictive Control of Networked Multiagent Systems via Cloud Computing

This paper studies the design and analysis of networked multiagent predictive control systems via cloud computing. A cloud predictive control scheme for networked multiagent systems (NMASs) is proposed to achieve consensus and stability simultaneously and to compensate for network delays actively. The design of the cloud predictive controller for NMASs is detailed. The analysis of the cloud predictive control scheme gives the necessary and sufficient conditions of stability and consensus of closed-loop networked multiagent control systems. The proposed scheme is verified to characterize the dynamical behavior and control performance of NMASs through simulations. The outcome provides a foundation for the development of cooperative and coordinative control of NMASs and its applications.

Escape Analysis on the Confinement-Escape Problem of a Defender Against an Evader Escaping From a Circular Region

In this paper, we investigate some mathematical properties of the confinement-escape problem of a defender and an evader with respect to a circular region, which was proposed in the author's previous work. Initially, the evader is located inside the circle, the defender patrols on the circle and tries to seal it to prevent the evader' escape; while the evader attempts to escape with avoidance of the defender. Here, we adopt the same control laws of the agents and consider particularly the successful-escape conditions which ensure a monotone-increasing distance (MID) between the defender and the evader as the system evolves, for abbreviation, we call it the escape with the MID to the defender, or simply the MID escape. Then, we: 1) provide some sufficient conditions for the MID escape under different situations; 2) provide the corresponding upper-limit estimations of the escape time; and 3) discuss the characteristics of the analytical results.

Unified Generic Geometric-Decompositions for Consensus or Flocking Systems of Cooperative Agents and Fast Recalculations of Decomposed Subsystems Under Topology-Adjustments

This paper considers a unified geometric projection approach for: 1) decomposing a general system of cooperative agents coupled via Laplacian matrices or stochastic matrices and 2) deriving a centroid-subsystem and many shape-subsystems, where each shape-subsystem has the distinct properties (e.g., preservation of formation and stability of the original system, sufficiently simple structures and explicit formation evolution of agents, and decoupling from the centroid-subsystem) which will facilitate subsequent analyses. Particularly, this paper provides an additional merit of the approach: considering adjustments of coupling topologies of agents which frequently occur in system design (e.g., to add or remove an edge, to move an edge to a new place, and to change the weight of an edge), the corresponding new shape-subsystems can be derived by a few simple computations merely from the old shape-subsystems and without referring to the original system, which will provide further convenience for analysis and flexibility of choice. Finally, such fast recalculations of new subsystems under topology adjustments are provided with examples.

The Confinement-Escape Problem of a Defender Against an Evader Escaping from a Circular Region

In this paper, we first formulate the confinement-escape problem of a defender and an evader who attempts escaping from a circular region, which differs from the traditional pursuit-evasion problems. In our setting of the confinement-escape problem, the defender is restricted to move and patrol on the circle, trying to prevent possible escape of the evader who is initially located inside the circle. We describe and characterize some general properties of the problem, and then design two bio-inspired control strategies for the evader and the defender, respectively. In addition, we illustrate some possible motion patterns of the system, investigate the escaping time as a function of the relative-initial-positions of the agents, as well as the winning sets of the two players, respectively, under different system-parameters. To that end, we characterize the contour lines of the winning sets with their gradient properties. Finally, we indicate the abrupt phase transitions between successful confinement and escaping, revealing the strong sensitivity and nonlinearities of the system under critical conditions.