|
Exponential Synchronization of Heterogeneous Multi-agent Systems via Linear-signum-type Diffusive Coupling: Blended Dynamics Approach |
|
Jeong Mo Seong, Donggil Lee, Hyungbo Shim(Seoul National University, Korea) |
| This paper investigates the behavior of heterogeneous agents that interact through diffusive coupling, resulting in emergent blended dynamics that may not be observed in the dynamics of individual agents. In particular, we propose a linear-signum-type diffusive coupling to achieve exponential synchronization among heterogeneous agents. We provide conditions of both linear and signum coupling strengths for achieving the exponential synchronization and analyze the associated convergence rate. |
|
|
MILP and CBBA-based Dynamic Assignment for Multi-USV Defensive Pursuit |
|
Min Kyu Shin, Han-Lim Choi(KAIST, Korea) |
| In this paper, we propose a novel approach to formulate a multi-unmanned surface vehicles(USVs) defensive pursuit problem that can be solved using Mixed Integer Linear Programming (MILP) and Consensus-Based Bundle Algorithm (CBBA). Our objective is to enhance the safety and efficiency of the defense process, which is achieved through a two-step assignment strategy involving position assignments and target assignments.The simulation results show that the formulation successfully defends the targets away from the asset. Also, both results using MILP and CBBA demonstrate the formation of cohesive teams among the USVs, effectively defending against the incoming USVs. |
|
|
Safety-Critical Optimal Control Using Control Barrier Function with Selective Relaxation Method |
|
Yeongho Song, Ngo Phong Nguyen, Hyondong Oh(UNIST, Korea) |
| In this study, we propose a distributed safety-critical optimal flocking control algorithm for a multi-agent
system. In general, safety(i.e., collision avoidance) could be guaranteed by the use of control barrier function (CBF). However, the existence of conflicting safety constraints from multiple agents could lead to a feasibility problem (i.e., no feasible control command). To deal with this issue, we propose effective feasibility enhancement approaches, such as the development of the CBF associated with the collision risk, to alleviate the burden of distributed safety constraints on agents with a high risk of collision. Numerical simulations show the validation of the proposed algorithm. |
|
|
Decentralized Trajectory Planning for Quadrotor Swarm with Goal Convergence Guarantee |
|
Jungwon Park, H. Jin Kim(Seoul National University, Korea) |
| This paper presents a decentralized multi-agent trajectory planning (MATP) algorithm that guarantees the agents reach the desired goal in obstacle environments. The proposed approach adopts the subgoal optimization method that allows the agent to reach the subgoal without deadlock or livelock, and it utilizes the waypoint from a grid-based multi-agent path planning (MAPP) algorithm to guide the subgoal to the desired goal. We prove the proposed algorithm ensures goal convergence if all agents can communicate with each other. We validate that the proposed approach does not cause a deadlock or livelock in cluttered environments through the simulation. |
|