Search for: All records

Abstract In various applications of multi-robotics in disaster response, warehouse management, and manufacturing, tasks that are known a priori and tasks added during run time need to be assigned efficiently and without conflicts to robots in the team. This multi-robot task allocation (MRTA) process presents itself as a combinatorial optimization (CO) problem that is usually challenging to be solved in meaningful timescales using typical (mixed)integer (non)linear programming tools. Building on a growing body of work in using graph reinforcement learning to learn search heuristics for such complex CO problems, this paper presents a new graph neural network architecture called the covariant attention mechanism (CAM). CAM can not only generalize but also scale to larger problems than that encountered in training, and handle dynamic tasks. This architecture combines the concept of covariant compositional networks used here to embed the local structures in task graphs, with a context module that encodes the robots’ states. The encoded information is passed onto a decoder designed using multi-head attention mechanism. When applied to a class of MRTA problems with time deadlines, robot ferry range constraints, and multi-trip settings, CAM surpasses a state-of-the-art graph learning approach based on the attention mechanism, as well as a feasible random-walk baseline across various generalizability and scalability tests. Performance of CAM is also found to be at par with a high-performing non-learning baseline called BiG-MRTA, while noting up to a 70-fold improvement in decision-making efficiency over this baseline. 

Abstract For a wide variety of envisioned humanitarian and commercial applications that involve a human user commanding a swarm of robotic systems, developing human-swarm interaction (HSI) principles and interfaces calls for systematic virtual environments to study such HSI implementations. Specifically, such studies are fundamental to achieving HSI that is operationally efficient and can facilitate trust calibration through the collection-use-modeling of cognitive information. However, there is a lack of such virtual environments, especially in the context of studying HSI in different operationally relevant contexts. Building on our previous work in swarm simulation and computer game-based HSI, this paper develops a comprehensive virtual environment to study HSI under varying swarm size, swarm compliance, and swarm-to-human feedback. This paper demonstrates how this simulation environment informs the development of an indoor physical (experimentation) environment to evaluate the human cognitive model. New approaches are presented to simulate physical assets based on physical experiment-based calibration and the effects that this presents on the human users. Key features of the simulation environment include medium fidelity simulation of large teams of small aerial and ground vehicles (based on the Pybullet engine), a graphical user interface to receive human command and provide feedback (from swarm assets) to human in the case of non-compliance with commands, and a lab-streaming layer to synchronize physiological data collection (e.g., related to brain activity and eye gaze) with swarm state and human commands. 

Abstract This paper introduces a new graph neural network architecture for learning solutions of Capacitated Vehicle Routing Problems (CVRP) as policies over graphs. CVRP serves as an important benchmark for a wide range of combinatorial planning problems, which can be adapted to manufacturing, robotics and fleet planning applications. Here, the specific aim is to demonstrate the significant real-time executability and (beyond training) scalability advantages of the new graph learning approach over existing solution methods. While partly drawing motivation from recent graph learning methods that learn to solve CO problems such as multi-Traveling Salesman Problem (mTSP) and VRP, the proposed neural architecture presents a novel encoder-decoder architecture. Here the encoder is based on Capsule networks, which enables better representation of local and global information with permutation invariant node embeddings; and the decoder is based on the Multi-head attention (MHA) mechanism allowing sequential decisions. This architecture is trained using a policy gradient Reinforcement Learning process. The performance of our approach is favorably compared with state-of-the-art learning and non-learning methods for a benchmark suite of Capacitated-VRP (CVRP) problems. A further study on the CVRP with demand uncertainties is conducted to explore how this Capsule-Attention Mechanism architecture can be extended to handle real-world uncertainties by embedding them through the encoder.