Tingrui Han

A fully distributed approach to resource allocation problem under directed and switching topologies

This paper addresses the distributed resource allocation problem for a network of multiple agents with directed and time-varying communication topologies. Suppose that the total amount of resources is a constant, represented by an equality constraint, and that the amount of resources allocated to each agent is subject to an inequality constraint, called the state constraint. We then aim to solve it in a fully distributed manner. By introducing a surplus variable to store the residue at each step due to the state constraint on each agent, a distributed iteration algorithm is proposed to solve the distributed resource allocation problem subject to the state constraints. It is shown that the algorithm converges globally provided that the communication graph is jointly strongly connected. The most promising characteristic of the algorithm is that the parameters used in the iteration by each agent depend only on local knowledge of the in-degree and out-degree of itself, yet the algorithm converges globally for a time-varying communication network.

Energy-Efficient Time Synchronization in Wireless Sensor Networks via Temperature-Aware Compensation

Time synchronization is critical for wireless sensor networks (WSNs) because data fusion and duty cycling schemes all rely on synchronized schedules. Traditional synchronization protocols assume that wireless channels are available around the clock. However, this assumption is not true for WSNs deployed in intertidal zones. In this article, we present TACO, a synchronization scheme for WSNs with intermittent wireless channels and volatile environmental temperatures. TACO estimates the correlation of clock skews and temperatures by solving a constrained least squares problem and continuously adjusts the local time with the predicted clock skews according to temperatures. Our experiment conducted in an intertidal zone shows that TACO can greatly reduce the clock drift and prolong the resynchronization intervals.

A Distributed Algorithm for Resource Allocation Over Dynamic Digraphs

This paper studies a distributed resource allocation problem for a multiagent network with a time-varying digraph. Each agent in the network is associated with a local variable (resource) and a convex cost function. The goal is to collectively minimize the total cost in a distributed fashion, subject to individual resource constraints, and collective equality constraints. The main challenge of the problem is due to the local information structure imposed by the time-varying digraph that should be considered as part of the problem formulation. This paper develops a nonnegative surplus-based distributed optimization algorithm. It is shown that the proposed distributed algorithm converges to the global minimizer provided that the time-varying digraph is jointly strongly connected. Also, all the parameters used in the proposed algorithm rely only on local knowledge.

Three-dimensional formation merging control under directed and switching topologies

This paper concentrates on the formation merging control problem for a leader-follower network. The objective is to control a team of agents called followers such that they are merged with another team of agents called leaders to form a single globally rigid formation. A method based on graph Laplacian is introduced to address this problem. Each follower selects its interaction neighbors and interaction weights according to the given target configuration. The graph modeling the interaction topology of all the agents is directed and time-varying. First, by assuming that the synchronized velocity of the leaders is known to all the followers, a necessary and sufficient condition is derived to ensure uniform asymptotic formation merging. Second, we relax this assumption and consider that the velocity of the leaders is known to only a subset of followers, for which the same necessary and sufficient condition is obtained with the help of an internal model for velocity synchronization.

A barycentric coordinate based approach to formation control of multi-agent systems under directed and switching topologies

This paper studies the formation control problem for a leader-follower network in 3D. The objective is to control the agents to form a globally rigid formation, for which the sensing graph is directed and switching while the communication graph is undirected and switching. Under such a setup, a barycentric coordinate based approach is proposed for the design of formation control laws ensuring global convergence. A necessary and sufficient graphical condition is obtained to guarantee that the followers converge to form a globally rigid formation together with the leaders. By this approach, the formation of the whole group, namely, the orientation, translation and formation scale, can be reconfigured by the leaders.

Formation Merging Control in 3D under Directed and Switching Topologies

Abstract The paper studies the formation merging problem for a leader-follower network. That is, how to control a team of agents called followers so that they are merged with a team of agents called leaders to form a larger globally rigid formation. Under the premise that a group of leaders move in a globally rigid formation with their synchronized velocity known to the followers, we show that the followers can asymptotically merge themselves to the formation for arbitrarily initial configurations. Each follower selects its neighbors and also its control law according to the target formation they aim to achieve and thus it allows directed and time-varying switching topologies. It is shown that a globally rigid formation can be merged asymptotically for the leader-follower network in a setup with directed and time-varying graphs if and only if every follower frequently has a joint path from at least a leader.

Three-dimensional formation merging control of second-order agents under directed and switching topologies

This paper investigates the formation merging problem for a leader-follower network. The objective is to control a group of agents called followers and modeled by double-integrator dynamics so that they are merged with another group of agents called leaders to form a single globally rigid formation. With the assumption that leaders move in a globally rigid formation with their synchronized velocity known to the followers, we show that a globally rigid formation can be merged. Each follower selects its neighbors and control law according to the target configuration and thus it allows directed and time-varying switching topologies. We provide a necessary and sufficient condition such that a globally rigid formation can be merged asymptotically for the leader-follower network in a setup with directed and time-varying graphs.

Distributed localization with mixed measurements under switching topologies

Abstract This paper investigates the distributed localization problem of sensor networks with mixed measurements. Each node holds a local coordinate system without a common orientation and is capable of measuring only one type of information (either distance, bearing, or relative position) to near-by nodes. Thus, three types of measurements are mixed in the sensor networks. Moreover, the communication topologies in the sensor networks may be time-varying due to unreliable communications. This paper develops a fully distributed algorithm called BCDL (Barycentric Coordinate based Distributed Localization) where each node starts from a random initial guess about its true coordinate and converges to the true coordinate via only local node interactions. The key idea in BCDL is to establish a unified linear equation constraints for the sensor coordinates by using the barycentric coordinates of each node with respect to its neighbors though the sensor nodes may have different types of measurements. Then a distributed iterative algorithm is proposed to solve the linear equations under time-varying communication networks. A necessary and sufficient graphical condition is obtained to ensure global convergence of the distributed algorithm.

A barycentric coordinate based approach to three-dimensional distributed localization for wireless sensor networks

This paper introduces a new range-based distributed localization algorithm for wireless sensor networks in the three-dimensional space. To overcome the difficulty in computing the barycentric coordinates of a node with respect to its neighbors in 3D, a new scheme is developed for this purpose by using the fact that a congruent framework of the subnetwork consisting of the node and its neighbors has the same barycentric coordinates. A multidimensional scaling (MDS) method is used to calculate the coordinates of the congruent framework, which can provide a solution for each subnetwork with not only exact range measurements but also noisy range measurements. Thus, a linear equation related to a signed Laplacian is obtained to describe the geometric constraints of the whole network. We then show a globally convergent and linear iterative algorithm called MDS-DILOC for each node to locate itself in the three-dimensional space by solving the linear equation. Simulation is carried out to demonstrate the validity of MDS-DILOC.

A survey on concurrent network localization for autonomous multi-vehicle systems

Network localization plays the key role in position based operations and services of autonomous multi-vehicle systems such as unmanned aerial vehicles and satellite formation, which aims to determine the Euclidean coordinates of all the vehicles given a collection of inter-vehicle measurements and the Euclidean coordinates of a small number of vehicles. Concurrent network localization approaches allow each vehicle solve its own coordinate in a distributed manner via local information exchange with its neighbors and thus exhibit many desirable advantages in applications. This paper reviews recent development on concurrent network localization algorithms and aims to provide a comprehensive insight for these novel ideas. In terms of the types of inter-vehicle information, distance-based, bearing-based, and relative-position-based concurrent network localization are discussed in a unified framework.