Task Allocation and Coalition Formation in Heterogeneous Multi-Robot-Systems
More and more robotic systems emerge, designed for a growing diversity of applications, with superior abilities and capable to perform increasingly challenging tasks. Human desire for interaction and cooperation of these systems leads to the demand for sophisticated solutions in the domain of multi-robot systems. As a consequence of the rising variety in robotic systems, future multi-robot-systems will be characterized by increasing robotic heterogeneity and dynamic changes of its robotic configuration and/or the assigned tasks.
The key challenges with respect to system architecture and task allocation are to provide dynamic (re-)allocation of tasks and dynamic system scalability, while keeping fault tolerance and a manageable complexity.
Within the CoTeSys demonstrator Multi-Robot-Lab various robotic platforms with different capabilities cooperate with humans and also with each other. One of the challenges within this scenario is how perceived tasks can be distributed efficiently among the robots. The key issues we focus on are a capability- and situation-aware task allocation and coalition formation - i.e. the dynamic formation of subgroups of robots for joint tasks - in heterogeneous Multi-Robot-Systems.
- Illustration of the market-based task allocation: The auctioneer forwards perceived tasks to the brokers, which request respective bids from their bidders. Then these bids are returned to the auctioneer, which assigns the task to the most efficient broker or coalition (team). In the example case here, Robots 1 and 2 form a coalition for one task while Robot n executes a further task alone.