Learning Decentralized Controllers with Graph Neural Networks

A recent paper by members of the DCIST alliance develops a method for distributed control of large networks of mobile robots with interacting dynamics and sparsely available communications. Their approach is to learn local controllers that require only local information and communications at test time by imitating the policy of centralized controllers using global information […]

Heterogeneity and Uncertainty in Perimeter Defense

Surveillance of perimeters and securing perimeters are important tasks in civilian and military defense applications, and it has become practical to deploy a large number of autonomous agents to address these problems using multi-robot systems.   A recent paper by members of the DCIST alliance formulates this scenario as a variant of multi-player pursuit-evasion games, where […]

Human Information Processing in Complex Networks

In this work, we study the structure of real-world communication systems to understand how information can be rapidly and efficiently communicated to humans, for example from swarms of drones or other agents. Humans constantly receive information from systems of interconnected stimuli or concepts — from language and music to literature and science — yet it […]

Resilient Active Information Acquisition with Mobile Robots

In the future, teams of heterogeneous robot teams will be operating in unknown and adversarial environments.   In failure prone or adversarial environments, the capability of resilience is crucial to ensuring the robots can complete their mission. Mission resilience to robot failures, sensor attacks or communication disruptions is currently an afterthought leading to optimal over-provisioned designs.   […]

Finite-Time Performance of Distributed Temporal Difference Learning with Linear Function Approximation

While many distributed reinforcement learning (RL) has emerged as one of the important paradigms in distributed control, we are only beginning to understand the fundamental behavior of these algorithms.  Two recent papers from the DCIST alliance provide important progress in this direction. In the multi-agent policy evaluation problem, a group of agents operate in a […]

Learning to Learn with Probabilistic Task Embeddings

To operate successfully in a complex and changing environment, learning agents must be able to acquire new skills quickly. Humans display remarkable skill in this area — we can learn to recognize a new object from one example, adapt to driving a different car in a matter of minutes, and add a new slang word […]

Localization and Mapping using Instance-specific Mesh Models

A recent paper by members of the DCIST alliance proposes an approach for building semantic maps, containing object poses and shapes, in real time, onboard an autonomous robot equippend with a monocular camera. Rich understanding of the geometry and context of a robot’s surroundigs is important for specification and safe, efficient execution of complex missions. This […]

Aerial Robot Prototype

CLEAR: A Consistent Lifting, Embedding, and Alignment Rectification Algorithm for Multi-Agent Data Association

Composable Autonomy in Heterogeneous Groups

In multi-robot Simultaneous Localization and Mapping (SLAM), a group of robots explore and map an unknown area. The group can benefit from its size by combining the robots’ maps to improve coverage and by each robot using shared information to improve its own localization. Most approaches to multi-robot SLAM consider homogeneous groups, in which all […]