This study investigates the distributed finite-time consensus problem of networked agents with second-order integrators. First of all, to achieve finite-time consensus, a saturated protocol is proposed based on both relative position and relative velocity measurements. In light of the tools from homogeneous theory, it is theoretically shown that the proposed protocol will guarantee finite-time consensus under bounded control inputs. Then, based only on the relative position measurements, an observer-type protocol is developed and employed for achieving finite-time consensus. Theoretical studies indicate that under the proposed algorithms, the states of the agents will achieve finite-time consensus without velocity measurements. Finally, the effectiveness of the theoretical results is illustrated by numerical simulations.

This article studies a synchronised output regulation (SOR) problem exchanging information at sampled time. Each node has a continuous-time internal model controller and a local exosystem providing a reference output signal. At sampled time, each node sends the exosystem state to its neighbouring nodes, as well as updates the exosystem dynamics by the received sampled data. A necessary and sufficient condition that the sampled period should satisfy for SOR with fixed digraph and time delay is presented. A numerical simulation shows an efficacy of analytic results.

This article studies the problem of H _∞ dynamic output-feedback control of non-linear networked discrete-time systems with data packet dropouts. Specifically, the data missing happens randomly from the sensor to the controller and the controller to the actuator. A sequence of independent variables obeying the Bernoulli random binary distribution is introduced to account for the measurement missing. The Takagi–Sugeno fuzzy model is used to describe the non-linear plant. In addition, the strategy of parallel distributed compensation is adopted to design the fuzzy dynamic output-feedback controller. Here, attention is focused on obtaining the sufficient conditions to guarantee the closed-loop system to be stochastically stable as well as the prescribed H _∞ performance. An approach based on fuzzy Lyapunov function is developed to solve it. To tackle the non-convex problem, the cone-complementarity linearisation procedure is adopted. Finally, a numerical example is employed to show the usefulness of the proposed result.

This study investigates the consensus problem of second-order multi-agent systems (MASs) via impulsive control using position-only information with communication delays. The communication delays between any two distinct agents are different which can be larger than one impulsive period. A distributed impulsive consensus protocol is designed, in which only the delayed sampled relative positions to neighbours and the relative position to the last sampling state are utilised. By introducing some virtual subgraphs and performing three steps of model transformation, the consensus problem of original continuous-time system is converted to the stability problem of a discrete-time expanded error system. Some necessary and sufficient criteria are derived to grantee the dynamic average-consensus of the MAS. Furthermore, a special case, that is, the delays less than one impulsive period, is discussed, and the bounds of impulsive periods for the dynamic average-consensus are obtained in explicit expressions for both the undirected and directed communication graphs, respectively. Two numerical simulation examples are given to illustrate the effectiveness of theoretical results.

Consider a system of autonomous interacting agents moving in space, adjusting each own velocity as a weighted mean of the relative velocities of the other agents. In order to test the robustness of the model, the authors assume that each pair of agents, at each time step, can fail to connect with certain probability, the failure rate. This is a modification of the (deterministic) Flocking model introduced by Cucker and Smale (2007). They prove that, if this random failures are independent in time and space, and have linear or sub-linear distance dependent rate of decay, the characteristic behaviour of flocking exhibited by the original deterministic model, also holds true under random failures, for all failure rates.

Dynamic behaviour of multi-agents raises many novel problems in control theory and robotics. In particular, the flocking problem for a group of mobile agents with uncertain parameters that can always achieve the desired coordination is a great new challenge problem. In this study, a adaptive flocking of non-linear multi-agents systems with uncertain parameters is investigated. Firstly, adaptive controller and an update law are designed for multi-agents system with uncertain parameters. Second, based on the Lyapunov stability theory, it is proved that velocities of each agent come to be the same asymptotically, the collisions among agents can be avoided, and the uncertain parameters can also be identified. Finally, an illustrative example is given to verify the analytical results.

In this study, the multi-consensus problem is investigated for second-order multi-agent systems with only sampled position information. By matrix theory and algebraic graph, the multi-consensus and multi-tracking can be achieved if the gains and sampling period satisfy a range. Simulations are provided to verify the effectiveness of the results.

Guaranteed cost consensus for second-order multi-agent systems with fixed topologies are investigated. Firstly, a cost function is constructed based on state errors among neighbouring agents and control inputs of all the agents, which is to find a tradeoff between the consensus regulation performance and the control energy consumption. Secondly, by the state-space decomposition approach and the Lyapunov method, a sufficient condition for the guaranteed cost consensus is presented and an upper bound of the cost function is given. It should be pointed out that these criteria are related to the second smallest and the maximum eigenvalues of the Laplacian matrix associated with the interaction topology. Thirdly, an approach is presented to obtain the consensus function when second-order multi-agent systems achieve guaranteed cost consensus. Finally, numerical simulations are given to demonstrate theoretical results.

In this study, the distributed averaging of high-dimensional first-order agents is investigated with relative-state-dependent measurement noises. Each agent can measure or receive its neighbours’ state information with random noises, whose intensity is a non-linear matrix function of agents’ relative states. By the tools of stochastic differential equations and algebraic graph theory, the authors give sufficient conditions to ensure mean square and almost sure average consensus and the convergence rate and the steady-state error for average consensus are quantified. Especially, if the noise intensity function depends linearly on the relative distance of agents’ states, then a sufficient condition is given in terms of the control gain, the noise intensity coefficient constant, the number of agents and the dimension of agents’ dynamics.

This study investigates aperiodic sampled-data synchronisation of complex dynamical network (CDN) incorporate dynamics of actuators saturation. The sampling intervals considered here are time-varying which allowed to be variable within the lower and upper bounds. In order to fully captures the sawtooth structure characteristic of the sampled-data systems, a novel time-dependent continuous Lyapunov functional is employed to obtain less conservative criteria. Furthermore, the generalised sector bound condition is utilised for the estimate about the basin of the attraction. Based on the local stability condition, the sampled-data controller is constructed to guarantee the synchronisation of CDN in the presence of actuators saturation. The obtained sufficient conditions can be cast in two optimisation cases to maximise the admissible upper bound of sampling instants or enlarge estimate about the domain of attraction for the closed-loop systems. Subsequently, the validity and applicability of the criteria are verified through numerical examples.

In most existing works on the flocking problem, results were obtained for the case where a double integrator dynamic is used to describe motion of agents and virtual leader. These results, however, may not be generalised to agents with general linear dynamic. In this study, it is shown that in the flocking of agents with a linear dynamic, the velocity convergence of agents does not generally lead to the position convergence of informed agents and the position convergence is possible only for a specified group of linear dynamics. In this study, a control protocol is proposed to flock agents with specified linear dynamics and guarantee the velocity convergence of all agents as well as average position convergence of informed agents to those of the virtual leader.

This study aims to investigate containment control of linear multi-agent systems with input saturation on switching topologies. For such a multi-agent system, both state feedback and output feedback containment control protocols are proposed via a novel low-gain feedback approach. Consequently, the authors show that, under the assumption that the switching networks are jointly connected, semi-global state feedback and output feedback containment control of the multi-agent system can be attained. The effectiveness of the proposed algorithms is illustrated via numerical examples.

In this work, distributed filtering over wireless sensor networks with limited energy is considered, where each sensor sends its local state estimate to its adjacent sensors through an unreliable wireless channel, which introduces random data packet drops. The packet drop rate depends on the power allocated by the sensor under an energy constraint. Several offline power scheduling strategies are introduced to distribute the power of sensors. A sufficient condition is provided to guarantee the convergence of network estimation error covariance. Further, an online power scheduling strategy is proposed, where each sensor utilises its real-time information to distribute energy for communications. The filtering performance of different power scheduling strategies are compared to show the influence of power distributed scheme on the expected state estimation error covariance.

This study proposes an improved particle filter by incorporating semi-supervised machine learning for location estimation in mobile sensor networks (MSNs). A time-varying prior model is learned online as the likelihood of particle filter in order to adapt to dynamic characteristics of state and observation. Thanks to semi-supervised learning, the proposed particle filter can improve efficiency and accuracy, where the amount of available labelled training data is limited. The authors compare the proposed algorithm with the particle filter based on supervised learning. The algorithms are evaluated for received signal strength indicator (RSSI)-based distributed location estimation for MSN in which communication bandwidth and accuracy of the range measurement are limited. First, experimental results show that the semi-supervised algorithm can learn suddenly-changed RSSI characteristics while the supervised learning cannot. Second, the proposed particle filter is more accurate and robust against variations of the environment such as new obstacle configurations. Furthermore, the suggested particle filter shows low statistical variability during repeated experiments, confirmed by much smaller error deviation than the compared particle filter.

This study considers the decentralised periodic event-triggered consensus problem for continuous-time linear multi-agent systems (MASs) under general directed graphs. The authors discretise the continuous-time MASs based on zero-order hold, and propose an event-triggered consensus control (ETCC) algorithm rendering the states of agents reach consensus. They prove that under the proposed ETCC implemented in multiple steps, there is no Zeno-like behaviour exhibited in the sense of sampling. To relax the requirement of the periodic monitoring in the ETCC, it is further proposed a self-triggered consensus control algorithm, where the next triggering instant is predetermined by each agent at the previous triggering instant. For the case where the states of MASs are unavailable, the authors further propose an observer-based ETCC algorithm. Simulation results are given to illustrate the theoretical results.

This study addresses the cooperative control problem of multi-missile systems. It proposed a two-stage control strategy, aiming at simultaneous attack from a group of missiles at a static target. The first stage adopts a special distributed consensus protocol in order for all missiles to asymptotically achieve a consensus of states. During the second stage, the local sightline control law allows the missiles to independently reach the target. The dynamic equation of the missile agent is normalised to a quasi-double-integrator model which is convenient for designing the consensus protocol. The proposed strategy is suitable for missiles of different speeds that have been self-organised without air operations centres. Two convincing simulation results are given to illustrate the efficiency of the proposed method.

This study investigates the emergency decision-making problem in a multi-agent system. Departments are modeled as agents to perform coordinated planning to obtain a global action plan with a short execution time constrained by a prescribed deadline. A novel multi-agent planning approach is proposed to coordinate for the solution among agents in the system. The approach consists of two stages. For the first stage, a coordinated planning and scheduling method is designed to generate a candidate global action plan. In this plan, durative actions are assigned to proper time intervals with the consideration of temporal relationships among agents. For the second stage, a coordination mechanism based on local heuristics is proposed to reduce the global time cost of the candidate plan to satisfy the deadline. An experimental study of emergency evacuation problem is conducted to demonstrate the effectiveness and efficiency of the proposed approach.

This study presents a novel observer-based electronic line-shafting control strategy for permanent magnet synchronous motor systems. Adopting the sliding-mode variable structure technology, the proportional–integral speed controller for every servo unit is designed, followed by design of two cascades connected observers, in order to improve the system synchronisation performance when a large load disturbance occurs. The load torque observer transmits the observed load torque value instead of the calculated one back to the virtual shafting, which reflects the dynamic relationships of all the virtual or slave shafts more accurately. Another load torque derivative observer feeds the observed value forward to the corresponding sliding-mode controller for reducing the switching gain. The proposed control system is proved stable by Lyapunov stability theory, and simulation results show the strategy is effective on reducing sliding-mode chattering and achieving higher synchronisation precision.

This study examines equilibrium behaviour and negotiation protocol design for a class of systems composed of multiple, non-cooperative, agents. The agents modelled as finite-state transition systems, are autonomous, and are interacting ‘concurrently’ aiming at achieving individual tasks specified in temporal logic. Each agent has its own preferences over outcomes of its interaction with others. The agents’ goals and preferences are neither perfectly aligned nor necessarily opposing. The authors reason about agent behaviours in such a system, by formulating a concurrent multi-agent game with infinitely many stages. To enable the synthesis of strategies, they develop a negotiation protocol which ensures that under a proper design of preferences and tasks, the mutually accepted plan is a Pareto optimal pure Nash equilibrium.

Urban traffic networks are large-scale systems, consisting of many intersections controlled by traffic lights and interacting connected links. For efficiently regulating the traffic flows and mitigating the traffic congestion in cities, a network-wide control strategy should be implemented. Control of large-scale traffic networks is often infeasible by only using a single controller, that is, in a centralised way, because of the high dimension, complicated dynamics and uncertainties of the system. In this study, the authors propose a multi-agent control approach using a congestion-degree-based serial scheme. Each agent employs a model-based predictive control approach and communicates with its neighbours. The congestion-degree-based serial scheme helps the agents to reach an agreement on their decisions regarding traffic control actions as soon as possible. A simulation study is carried out on a hypothetical large-scale urban traffic network based on the presented control strategy. The results illustrate that this approach has a better performance with regard to computation time compared with the centralised control method and a faster convergence speed compared with the classical parallel scheme.

This study studies the opinion dynamics with group polarisation. The authors say a process of opinion formation is polarising if the final opinions after group discussion are more divergent than individuals’ initial opinions. To capture this phenomenon, they develop a multi-agent model of opinion formation in which each agent forms their opinion according to all arguments available to them, and during interacting with other agents, they assimilate the arguments of neighbours in a biased manner which can be represented by a modified Polya's urn process. They find that biased assimilation of arguments can result in group polarisation effectively, especially under the condition when the agents are relatively isolated in the network.

This study is concerned with integrated target search, tasking and tracking using multiple fixed-wing unmanned air vehicles (UAVs). The problem is to design control logic and optimise flight paths for UAVs. The fixed-wing UAVs are required to cooperatively search the potential targets in the area of operation (AO) and keep monitoring and tracking the found targets according to a certain predefined minimal revisit time. Each UAV can only communicate with its neighbours and also flight autonomy is designed for individual UAVs. High level control logic based on finite state automaton model, integrating the four modes of operations, that is, take-off mode, fly-to-AO mode, search mode and tracking mode, is developed. For the target search, an improved decentralised coverage based algorithm is designed for fixed-wing UAVs. Task assignment is developed based on contract net protocol. For the target tracking, the optimal paths are derived for the UAVs to continuously monitoring multiple static targets. The developed integrated multi-agent system is evaluated by a hybrid three-dimensional simulation system using a real miniature fixed-wing UAV dynamic model.

In this study, the authors consider the large population dynamic games where each agent evolves according to a dynamic equation containing the input average of all agents. The long time average (LTA) cost that each agent aims to minimise is coupled with other agents’ states via a population state average (PSA), which is also known as the mean field term. In order to design decentralised controls, the Nash certainty equivalence is introduced. It is shown that the resulting decentralised mean field control laws lead the system to mean-consensus asymptotically as time goes to infinity. The stability property of the mass behaviour and the almost sure asymptotic Nash equilibrium property of the optimal controls are also guaranteed, and the case with non-linear system dynamics is also discussed. In addition, the influence of inaccurate mean field information on individual agent is analysed. Finally, they investigate the socially cooperative formulation where the objective is to minimise the social cost as the sum of all individual LTA costs containing the PSA. In this case, they show that the decentralised mean field social controls are the same as the mean field Nash controls for infinite population systems.