D2CoPlan: A Differentiable Decentralized Planner for Multi-Robot Coverage

Centralized approaches for multi-robot coverage planning problems suffer from the lack of scalability. Learning-based distributed algorithms provide a scalable avenue in addition to bringing data-oriented feature generation capabilities to the table, allowing integration with other learning-based approaches. To this end, we present a learning-based, differentiable distributed coverage planner (D2CoP LAN ) which scales efficiently in runtime and number of agents compared to the expert algorithm, and performs on par with the classical distributed algorithm. In addition, we show that D2CoP LAN can be seamlessly combined with other learning methods to learn end-to-end, resulting in a better solution than the individually trained modules, opening doors to further research for tasks that remain elusive with classical methods.

Authors

• Vishnu Dutt Sharma
• Lifeng Zhou 0001
• Pratap Tokekar

Keywords

• Training
• Learning systems
• Runtime
• Scalability
• Robot kinematics
• Reinforcement learning
• Prediction algorithms
• Multi-robot Coverage
• Learning-based Approaches
• Centralized Approach
• Distributed Algorithm
• Neural Network
• Time Step
• Convolutional Neural Network
• Grid Size
• Local Map
• Motion Model
• Differential Modulation
• Feature Aggregation
• Formation Control
• Graph Neural Networks
• Target Coverage
• Communication Range
• Target Density
• Mapping Coverage
• Coverage Problem
• Classical Counterpart
• Coverage Performance