Multi-robot Formation Research Articles

Methods of Formation Control for a Group of Mobile Robots (a Review)

The multi-robot formation control is an essential issue in robotics. This review focuses on important lines of research on current contr ol is s ues and s tr ateg ies on a g r oup of unmanned autonomous vehicles/r obots for mation. In this paper, we pr ovide a br ief description of each method characterizing its key benefits and drawbacks. A multilayered classification of both centralized and decentralized formation control methods is proposed. We consider the classification of robot communication topologies in terms of centralized control. Seminal works dedicated to the practical application of centralized approach are briefly discussed. The majority of centralized methodsare represented by a " leader-follower" approach, taking into account the robot’s dynamics models. Furthermore, the most common models of vehicle dynamics are mentioned. In the framework of decentralized approach, behaviour-based algorithms, as well as swarm algorithms, are discussed. Then, we present an outlook of both centralized and decentralized virtual structure methods used in robot formation control. The described modifications of these methods allow tracing the evolution of the virtual structure approach to hybrid algorithms used for cooperative movement of a group of robots. This paper deals with formation control approach considering communication delays and low carrying capacity in an inter-vehicular communication network as very few works discussed this issue despite its relevance. We pointed out the main development trends of formation control approaches. The most effective approach is the integration of various methods of the formation control so that their disadvantages are nullified. As the same time, the most common disadvantage of discussed formation control methods is their weak conceptual framework in terms of kinematic and dynamic constraints of robots.

Null-space-based modulated reference trajectory generator for multi-robots formation in obstacle environment

This paper devotes to the three-dimensional formation problem of multi-robots in obstacle environment. Given the desired formation pattern and the group trajectory, it is formulated as obtaining the control inputs of robots so that the formation errors converge to zero gradually with obstacle/collision avoidance subject to state and input constraints. The well-known nonlinear model predictive control (NMPC) can be utilized as the solution framework due to its stability and robustness according with the reference state vector. Particularly, the null-space-based modulated reference trajectory generator is proposed to modulate the reference state vector of each robot. The original reference velocity, obtained from the Lyapunov stability theory, will be modulated quantitatively in the presence of each obstacle, and then the modulated velocities are integrated effectively on the basis of null space. From the perspective of trajectory generator, it is proven that the robots will avoid obstacles or collision without violating the stability of formation system. Finally the simulation results demonstrate the high efficiency and strong robustness of our method.

A Path-Following Controller for Guiding a Single Robot or a Multi-robot System

A solution to the problem of path-following by a team of terrestrial mobile robots is proposed in this paper. The proposal, in this case, is a formation controller dealing with three robots navigating in a coordinate way (as a formation). Based on the controller proposed to guide the formation to follow a prescribed path, an extension to the case of path-following with a single terrestrial mobile robot is also proposed. When regarding a formation of mobile robots, the proposed solution consists in applying a path-following controller to the center of mass of the formation, dealt with as a single virtual robot, and trajectory-tracking controllers to the individual robots in the formation. The advantage of such approach is that it allows planning the motion of the desired formation without specifying how each robot should move. The movement is specified for the formation as a whole, using a representation called cluster-space, and the movement of the individual robots is derived from the specification of the formation movement directly, using transformations from the cluster-space to the space of the robots and vice versa. In the sequel, the path-following controller designed for the virtual robot is analyzed in detail, now dealing with the possibility of being also used as a path-following controller applied to a single real robot. Theoretical stability analysis is presented, as well as some experimental results, whose analysis allows claiming that the proposed controller is suitable to guide either a single robot or a multi-robot formation when following a path.

Adaptive Finite-Time Tracking Control of Nonholonomic Multirobot Formation Systems With Limited Field-of-View Sensors.

This article studies the vision-based tracking control problem for a nonholonomic multirobot formation system with uncertain dynamic models and visibility constraints. A fixed onboard vision sensor that provides the relative distance and bearing angle is subject to limited range and angle of view due to limited sensing capability. The constraint resulting from collision avoidance is also taken into account for safe operations of the formation system. Furthermore, the preselected specifications on transient and steady-state performance are provided by considering the time-varying and asymmetric constraint requirements on formation tracking errors for each robot. To address the constraint problems, we incorporate a novel barrier Lyapunov function into controller design and analysis. Based on the recursive adaptive backstepping procedure and neural-network approximation, we develop a vision-based formation tracking control protocol such that formation tracking errors can converge into a small neighborhood of the origin in finite time while meeting the requirements of visibility and performance constraints. The proposed protocol is decentralized in the sense that the control action on each robot only depends on the local relative information, without the need for explicit network communication. Moreover, the control protocol could extend to an unconstrained multirobot system. Both simulation and experimental results show the effectiveness of the control protocol.

Online Nash Solution in Networked Multirobot Formation Using Stochastic Near-Optimal Control Under Dynamic Events.

This article proposes an online stochastic dynamic event-based near-optimal controller for formation in the networked multirobot system. The system is prone to network uncertainties, such as packet loss and transmission delay, that introduce stochasticity in the system. The multirobot formation problem poses a nonzero-sum game scenario. The near-optimal control inputs/policies based on proposed event-based methodology attain a Nash equilibrium achieving the desired formation in the system. These policies are generated online only at events using actor-critic neural network architecture whose weights are updated too at the same instants. The approach ensures system stability by deriving the ultimate boundedness of estimation errors of actor-critic weights and the event-based closed-loop formation error. The efficacy of the proposed approach has been validated in real-time using three Pioneer P3-Dx mobile robots in a multirobot framework. The control update instants are minimized to as low as 20% and 18% for the two follower robots.

NPQ‐RRT ∗: An Improved RRT ∗ Approach to Hybrid Path Planning

In recent years, the path planning of robot has been a hot research direction, and multirobot formation has practical application prospect in our life. This article proposes a hybrid path planning algorithm applied to robot formation. The improved Rapidly Exploring Random Trees algorithm PQ‐RRT ∗ with new distance evaluation function is used as a global planning algorithm to generate the initial global path. The determined parent nodes and child nodes are used as the starting points and target points of the local planning algorithm, respectively. The dynamic window approach is used as the local planning algorithm to avoid dynamic obstacles. At the same time, the algorithm restricts the movement of robots inside the formation to avoid internal collisions. The local optimal path is selected by the evaluation function containing the possibility of formation collision. Therefore, multiple mobile robots can quickly and safely reach the global target point in a complex environment with dynamic and static obstacles through the hybrid path planning algorithm. Numerical simulations are given to verify the effectiveness and superiority of the proposed hybrid path planning algorithm.

Cooperative Load Transportation With Two Quadrotors Using Adaptive Control

The problem of carrying a bar-shaped payload suspended by flexible cables attached to two quadrotors is analyzed in this work. The aerial vehicles and the load are dealt with as a single system, whose kinematics is described as a multi-robot formation using the virtual structure approach. The dynamic effects caused by the tethered load over the quadrotors, as well as those caused by each quadrotor over the other, are treated by an adaptive dynamic compensator. To validate the proposal, experiments were run testing the system in adverse conditions: transportation far from quasi-static motion, high payload-to-quadrotor weight ratio, 20% of error in the robot model parameters, transportation under wind disturbances, and payload weight changes during flight. The good performance of the proposed control system in all these tests allows concluding that the proposed system is able to accomplish payload positioning, orientation, and trajectory tracking under adverse conditions, with accelerations up to 1.6 m/s2.

Three controllers via 2nd-order sliding mode for leader-following formation control of multi-robot systems

The multi-robot formation has always been an attractive field for cooperative control of robots. This paper uses the improved second-order sliding mode method to control the leader-following formation system. The traditional second-order sliding mode control is not perfect because the disturbances and uncertainties of the system may cause the compensation control to overestimate. The adaptive gain super-twisting sliding mode control suppresses the output oscillation and speeds up the system response by improving the gain function and adding a differential link. However, this method is generally applied to single-input single-output systems. This paper attempts to apply this method to a multi-input-multi-output leader-following formation system and improve its switching function and compensation control part. We use the estimators, i.e., the extreme learning machine and the nonlinear disturbance observer. These estimators can observe disturbance changes to replace the adaptive gain function to realise disturbance estimation and compensation control. Then combine the extreme learning machine and the nonlinear disturbance observer with the super-twisting sliding mode control to produce two new sliding mode methods. We use the Lyapunov method to prove the system stability as well. Finally, we illustrate the feasibility of the three control methods.

A UHF RFID-Based Dynamic Object Following Method for a Mobile Robot Using Phase Difference Information

Dynamic object following is an essential ability for mobile robots in various application scenarios such as auxiliary cargo transportation, disabled assistant caring, and multirobot formation. A dynamic object following method for a mobile robot based on ultrahigh-frequency (UHF) radio frequency identification (RFID), which uses a phase difference information-based dynamic following model, is presented in this article. Through this method, a mobile robot equipped with an RFID reader can follow a dynamic target object attached with a passive RFID tag. A cost function is constructed to indicate the changes in relative distance and bearing between the target object and the following robot, allowing the real-time calculation of the translational and rotational velocities of the following robot using an algorithm base on proportional integral differential (PID)-based controllers. Compared with the existing RFID-based following methods, the proposed method can achieve a satisfactory distance and bearing following accuracy by using RFID phase difference information without any tedious and time-consuming phase drift or antenna installation calibration. Real experiments were conducted using commercial off-the-shelf (COTS) UHF RFID, and results show that the proposed method is effective and robust, the respective average distance and bearing following errors were approximately 1.65 cm and 4.94° during the robot-following tests and 4.45 cm and 6.03° during the human-following tests. The proposed method has promising applications in mobile robot following of dynamic object to provide service or assistance in large storage and manufacturing industries.

Improved Q‐Learning Method for Multirobot Formation and Path Planning with Concave Obstacles

Aiming at the formation and path planning of multirobot systems in an unknown environment, a path planning method for multirobot formation based on improved Q‐learning is proposed. Based on the leader‐following approach, the leader robot uses an improved Q‐learning algorithm to plan the path and the follower robot achieves a tracking strategy of gravitational potential field (GPF) by designing a cost function to select actions. Specifically, to improve the Q‐learning, Q‐value is initialized by environmental guidance of the target’s GPF. Then, the virtual obstacle‐filling avoidance strategy is presented to fill non‐obstacles which is judged to tend to concave obstacles with virtual obstacles. Besides, the simulated annealing (SA) algorithm whose controlling temperature is adjusted in real time according to the learning situation of the Q‐learning is applied to improve the action selection strategy. The experimental results show that the improved Q‐learning algorithm reduces the convergence time by 89.9% and the number of convergence rounds by 63.4% compared with the traditional algorithm. With the help of the method, multiple robots have a clear division of labor and quickly plan a globally optimized formation path in a completely unknown environment.

AirCapRL: Autonomous Aerial Human Motion Capture Using Deep Reinforcement Learning

In this letter, we introduce a deep reinforcement learning (DRL) 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-linearities, 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.

Formation Control of Multi-Robot using Virtual Structures with a Linear Algebra Approach

The separation bearing controller (SBC) and separation-separation controller (SSC) approaches can be used to maintain multi-robot formation, while the Computed Torque Control (CTC) method is used to solve the dynamic problems of the robot. However, the SBC and SSC approaches combined with CTC cannot maintain multi-robot formation, if the leader or follower is disturbed. In this paper, the virtual structure method and the linear algebra approach added to the CTC method are used to maintain the triangular robot formation to follow the given trajectory. The simulation results show that the proposed control system design of multi-robots capable of following the trajectory and can maintain multi-robot formation with/without disturbance and noise.

Multi-robot formation control based on high-order bilateral consensus

A high-order bilateral consensus robot formation control protocol for multi-agent systems is proposed in this paper. Considering the relationship between the state of the information exchange topology and derivatives, a third-order bilateral consistency protocol is presented and is extended it to a higher order bilateral consensus protocol. First, sufficient conditions for the third-order multi-agent system are given to achieve the bilateral consensus control protocol, and the system’s asymptotical stability is also achieved by adjusting the feedback system gain parameters. Then, by further studying the cohesive relationship between each state variable of the third-order protocol and the gauge transformation, the sufficient conditions of the higher order system are also provided. Finally, by applying the third-order control protocol to the control of multi-robot formation, the general control scheme of robot formation is given and the control of robot formation is successfully achieved.

Multi-robot formation control: a comparison between model-based and learning-based methods

Formation control of multi-robot systems has been extensively studied by model-based methods, where analytic control inputs are constructed based on the kinematics and/or dynamics model and the communication graphs of the multi-robot system. Recently, driven by remarkable advances of robotic learning techniques, emerging studies on learning-based methods for formation control have been developed for adaptive and intelligent control of multi-robot systems. This paper aims to provide a brief overview of our recent development of learning-based formation control, and compare it with a model-based method for a case study of three-robot formation control. Fundamental principles, experimental results and technical challenges are presented, comparing the two different methodologies.

Relative Docking and Formation Control via Range and Odometry Measurements

This article studies the problem of distance-based relative docking of a single robot and formation control of multirobot systems. In particular, an integrated localization and navigation scheme is proposed for a robot to navigate itself to a desired relative position with respect to a fixed landmark at an unknown position, where only range and odometry measurements are used. By carefully embedding historical measurements into equilibrium conditions, we design an integrated estimation-control scheme to achieve the relative docking asymptotically. It is rigorously proved that the robot will converge to the desired docking position asymptotically provided that control gains are chosen to satisfy certain conditions. This scheme is further extended to multirobot systems to consider an integrated relative localization and formation control problem. Unlike widely used spatial cooperation in the existing literature, we propose to exploit both spatial and temporal cooperations for achieving formation control. It is proved that multirobot formation can be achieved with zero error for directed acyclic graphs. Several simulation examples are provided to validate our theoretical results.

Coalition Formation for Multi-Robot Task Allocation via Correlation Clustering

The complexity of a vast number of real world tasks provides a great challenge for the currently available robots due to their limited capabilities. Thus, multiple robots would need to form coalitions for the completion of such tasks. In this paper, we examine the multi-robot coalition formation problem for task allocation where a group of robots needs to be allocated to a set of tasks. Our approach for this problem is to use a correlation clustering technique enabling similar robots to form coalitions. The algorithm presented in this paper is fast and scales better in comparison to two existing algorithms.

Adaptive Multirobot Formation Planning to Enclose and Track a Target With Motion and Visibility Constraints

Addressing the problem of enclosing and tracking a target requires multiple agents with adequate motion strategies. We consider a team of unicycle robots with a standard camera on board. The robots must maintain the desired enclosing formation while dealing with their nonholonomic motion constraints. The reference formation trajectories must also guarantee permanent visibility of the target by overcoming the limited field of view of the cameras. In this article, we present a novel approach to characterize the conditions on the robots’ trajectories taking into account the motion and visual constraints. We also propose online and offline motion planning strategies to address the constraints involved in the task of enclosing and tracking the target. These strategies are based on maintaining the formation shape with variable size or, alternatively, on maintaining the size of the formation with flexible shape.

Mapless Collaborative Navigation for a Multi-Robot System Based on the Deep Reinforcement Learning

Compared with the single robot system, a multi-robot system has higher efficiency and fault tolerance. The multi-robot system has great potential in some application scenarios, such as the robot search, rescue and escort tasks, and so on. Deep reinforcement learning provides a potential framework for multi-robot formation and collaborative navigation. This paper mainly studies the collaborative formation and navigation of multi-robots by using the deep reinforcement learning algorithm. The proposed method improves the classical Deep Deterministic Policy Gradient (DDPG) to address the single robot mapless navigation task. We also extend the single-robot Deep Deterministic Policy Gradient algorithm to the multi-robot system, and obtain the Parallel Deep Deterministic Policy Gradient (PDDPG). By utilizing the 2D lidar sensor, the group of robots can accomplish the formation construction task and the collaborative formation navigation task. The experiment results in a Gazebo simulation platform illustrates that our method is capable of guiding mobile robots to construct the formation and keep the formation during group navigation, directly through raw lidar data inputs.

A Stable Control Algorithm for Multi Robot Formation

This paper presents the developed trajectory tracking controller for a formation of nonholonomic robots, which combines features from the leader-follower and virtual-structure approaches. The implemented decentralized control strategy allows the robots to be relatively independent and to switch easily between the executed individual tasks and the collective tasks. Convergence is thoroughly analyzed and guarantied using the Lyapunov approach.

A robust localization system for multi-robot formations based on an extension of a Gaussian mixture probability hypothesis density filter

This paper presents a strategy for providing reliable state estimates that allow a group of robots to realize a formation even when communication fails and the tracking data alone is insufficient for maintaining a stable formation. Furthermore, the tracking information does not provide the identity of the robot, therefore a simple fusion of tracking and communication data is not possible. We extend a Gaussian mixture probability hypothesis density filter to incorporate, firstly, absolute poses exchanged by the robots, and secondly, the geometry of the desired formation. Our method of combining communicated data, information about the formation and sensory detections is capable of maintaining the state estimates even when long-duration occlusions occur, and improves awareness of the situation when the communication is sporadic or suffers from short-term outage. The proposed method is validated using a high-fidelity simulator in scenarios with a formation of up to five robots. The results show that the proposed tracking strategy allows for sustaining formations in cluttered environments, with high measurement uncertainty and low quality communication.