Person:

Egerstedt, Magnus B.

Permanent Link

https://hdl.handle.net/1853/72204

Associated Organization(s)

Organizational Unit

School of Electrical and Computer Engineering

Full item page

Publication Search Results

Now showing 1 - 10 of 171

Safety Barrier Certificates for Heterogeneous Multi-Robot Systems

(Georgia Institute of Technology, 2016-07) Wang, Li ; Ames, Aaron ; Egerstedt, Magnus B.

This paper presents a formal framework for collision avoidance in multi-robot systems, wherein an existing controller is modified in a minimally invasive fashion to ensure safety. We build this framework through the use of control barrier functions (CBFs) which guarantee forward invariance of a safe set; these yield safety barrier certificates in the context of heterogeneous robot dynamics subject to acceleration bounds. Moreover, safety barrier certificates are extended to a distributed control framework, wherein neighboring agent dynamics are unknown, through local parameter identification. The end result is an optimization-based controller that formally guarantees collision free behavior in heterogeneous multi-agent systems by minimally modifying the desired controller via safety barrier constraints. This formal result is verified in simulation on a multi-robot system consisting of both “sluggish” and “agile” robots.
Controllability of Prosumer-Based Networks in the Presence of Communication Failures

(Georgia Institute of Technology, 2015-12) Ramachandran, Thiagarajan ; Nazari, Masoud ; Egerstedt, Magnus B.

Typically, interconnected dynamical systems rely on communication in order to coordinate and compute appropriate control actions. Loss of communication links can exclude key decision makers from providing input and can even alter the system properties. This paper investigates the impact of communication loss on the controllability of a specific networked system, a homogeneous power-grid populated by producer-consumer hybrids. The notion of muteness is introduced in order to characterize the control policy adopted by the nodes which are isolated due to communication loss. We provide results which relates the controllability of such a system with mute nodes to the topology of the underlying electrical network and show that under certain topological conditions, controllability is preserved.
Temporal Heterogeneity and the Value of Slowness in Robotic Systems

(Georgia Institute of Technology, 2015-12) Arkin, Ronald C. ; Egerstedt, Magnus B.

Robot teaming is a well-studied area, but little research to date has been conducted on the fundamental benefits of heterogeneous teams and virtually none on temporal heterogeneity, where timescales of the various platforms are radically different. This paper explores this aspect of robot ecosystems consisting of fast and slow robots (SlowBots) working together, including the bio-inspiration for such systems.
Cloud-Based Centralized/Decentralized Multi-Agent Optimization with Communication Delays

(Georgia Institute of Technology, 2015-12) Hale, Matthew T. ; Nedić, Angelia ; Egerstedt, Magnus B.

We present and analyze a hybrid computational architecture for performing multi-agent optimization. The optimization problems under consideration have convex objective and constraint functions with mild smoothness conditions imposed on them. For such problems, we provide a primaldual algorithm implemented in the hybrid architecture, which consists of a decentralized network of agents into which an updated dual vector is occasionally injected, and we establish its convergence properties. In this setting, a central cloud computer is responsible for aggregating information, computing dual variable updates, and distributing these updates to the agents. The agents update their (primal) state variables and also communicate among themselves with each agent sharing and receiving state information with some number of its neighbors. Throughout, communications with the cloud are not assumed to be synchronous or instantaneous, and communication delays are explicitly accounted for in the modeling and analysis of the system. Experimental results for a team of robots are presented to support the theoretical developments made.
A Game-theoretic Formulation of the Homogeneous Self-Reconfiguration Problem

(Georgia Institute of Technology, 2015-12) Pickem, Daniel ; Egerstedt, Magnus B. ; Shamma, Jeff

In this paper we formulate the homogeneous two- and three-dimensional self-reconfiguration problem over discrete grids as a constrained potential game. We develop a game-theoretic learning algorithm based on the Metropolis-Hastings algorithm that solves the self-reconfiguration problem in a globally optimal fashion. Both a centralized and a fully decentralized algorithm are presented and we show that the only stochastically stable state is the potential function maximizer, i.e. the desired target configuration. These algorithms compute transition probabilities in such a way that even though each agent acts in a self-interested way, the overall collective goal of self-reconfiguration is achieved. Simulation results confirm the feasibility of our approach and show convergence to desired target configurations.
Correct-by-Construction Control Synthesis for Multi-Robot Mixing

(Georgia Institute of Technology, 2015-12) Diaz-Mercado, Yancy ; Jones, Austin ; Belta, Calin ; Egerstedt, Magnus B.

This paper considers the problem of controlling a team of heterogeneous agents to conform to high- level interaction (coordination, sensing, and communication) missions. We consider interactions that can be specified via symbolic inputs from the braid group. We define a novel specification language, called Braid Temporal Logic (BTL), that allows us to specify rich, temporally-layered tasks involving agents’ locations in an environment, their relative positions to each other, and frequency of location swaps and information exchanges between agents. We use techniques from formal methods to generate symbolic inputs that conform to a given BTL specification and use recently developed hybrid optimal control synthesis techniques to enact the synthesized pattern. The generated trajectories are provably guaranteed to be collision-free, respect physical boundaries of the agents’ mission space, and to satisfy the high-level mission. Results are validated via implementation on a team of wheeled robots.
Control Barrier Certificates for Safe Swarm Behavior

(Georgia Institute of Technology, 2015-10) Borrmann, Urs ; Wang, Li ; Ames, Aaron D. ; Egerstedt, Magnus B.

Multi-agent robotics involves the coordination of large numbers of robots, which leads to significant challenges in terms of collision avoidance. This paper generates provably collision free swarm behaviours by constructing swarm safety control barrier certificates. The safety barrier, implemented via an optimization-based controller, serves as a low level safety controller formally ensuring the forward invariance of the safe operating set. In addition, the proposed method naturally combines the goals of collision avoidance and interference with the coordination laws in a uni ed and computationally efficient manner. The centralized version of safety barrier certificate is designed on double integrator dynamic model, and then a decentralized formulation is proposed as a less computationally intensive and more scalable solution. The safety barrier certificate is validated in simulation and implemented experimentally on multiple mobile robots; the proposed optimization-based controller successfully generates collision free control commands with minimal overall impact on the coordination control laws.
Altering UAV Flight Path by Threatening Collision

(Georgia Institute of Technology, 2015-09) Pierpaoli, Pietro ; Egerstedt, Magnus B. ; Rahmani, Amir

The ongoing transformation of air traffic control towards decentralized decision making based on ADS-B information shared by neighboring traffic will allow all aircraft and UAS in particular, to automatically detect and resolve collisions. In this work we highlight the importance of trustworthiness in such distributed systems, showing that autonomous aircraft can be forced into predetermined trajectories when their precise position and velocity are available to a potentially malicious craft. In other words, malicious pursuer players (real or hoaxed) taking advantage of shared data and collision avoidance properties, can dictate evader agent trajectory, which might not realize the threat at all. As shown by numerical simulations and ground robot experiments, combination of arcs and straight paths can be achieved and be used to arbitrarily control the evader.
A Control Lyapunov Function Approach to Human-Swarm Interactions

(Georgia Institute of Technology, 2015-07) de la Croix, Jean-Pierre ; Egerstedt, Magnus B.

In this paper, we seek to establish formal guarantees for whether or not a given human-swarm interaction (HSI) is appropriate for achieving multi-robot tasks. Examples of such tasks include guiding a swarm of robots into a particular geometric configuration. In doing so, we define what it means to impose a HSI control structure on a multi-robot system. Control Lyapunov functions (CLFs) are used to prove that it is feasible for a user to achieve a particular geometric configuration with a multi-robot system under some selected HSI control structure. Several examples of multi-robot systems with unique HSI control structures are provided to illustrated the use of CLFs to establish feasibility.
Differentially Private Cloud-Based Multi-Agent Optimization with Constraints

(Georgia Institute of Technology, 2015-07) Hale, M. T. ; Egerstedt, Magnus B.

We present an optimization framework that solves constrained multi-agent optimization problems while keeping each agent’s state differentially private. The agents in the network seek to optimize a local objective function in the presence of global constraints. Agents communicate only through a trusted cloud computer and the cloud also performs computations based on global information. The cloud computer modifies the results of such computations before they are sent to the agents in order to guarantee that the agents’ states are kept private. We show that under mild conditions each agent’s optimization problem converges in mean-square to its unique solution while each agent’s state is kept differentially private. A numerical simulation is provided to demonstrate the viability of this approach.