There is a broad range of applications for mobile robots exploring unknown environments. Robot teams can explore faster than a single robot but require coordination and information exchange. Solutions for mapping, wireless networking, and coordinated recharging have been developed.
Mobile robots are more and more used in the exploration of unknown environments, in particular, for planetary exploration and support of first responders after a disaster. Robots have to learn their environment in an autonomous manner and create a map. Some applications have strict constraints on the duration and reliability of the exploration mission. In such critical applications, a team of robots rather than a single robot should be used. Multiple robots have the potential to speed up the mission by parallelizing tasks and are robust against failures of single robots but in turn require coordination and information exchange between the robots.
Software packages for autonomous exploration suited for robot teams have been developed and implemented by researchers and interns in the team of Christian Bettstetter. A first building block required is communications between robots. Our ad hoc communications package for the Robot Operating System (ROS) enables robots to wirelessly communicate directly with each other, without relying on any preexisting network infrastructure, such as base stations or access points. A second building block for exploration is to create a map common to all robots, as a basis for a common understanding for coordination. This task is achieved by our map merging package. “Our communication modules are used to exchange local maps produced by robots and to merge these local maps to a common global map,” predoctoral researcher Torsten Andre says. Together with a novel package for exploration, a tightly integrated software solution for coordinated multi-robot exploration is offered. “We tested the software on Turtlebots and Pioneer3-DX robots and now invite other researchers to use and further improve the implementation,” Andre concludes. A prototype with four robots was built that demonstrates its functionality in an indoor environment. The first video below shows this prototype in operation.
Another aspect of multi-robot exploration is the recharging of batteries. “The idea is that multiple robots acting as a team can explore areas beyond the reach of an individual robot,” predoctoral researcher Micha Rappaport explains the key idea behind his research.
He developed an energy-aware exploration strategy enabling robots to decide where to explore and when to recharge. A limited number of docking stations is available for a larger number of robots. Rappaport explains the key challenges: “Firstly, the robots need to be aware of their energy and decide autonomously when to seek a docking station. Secondly, exploring robots need to coordinate for deciding which robot is allowed to recharge and where it should recharge. Thirdly, robots need to form teams and coordinate task assignment.” It is impossible to precompute an optimized schedule in this setting due to the limited time horizon in the exploration of unknown environments. For this reason, an energy-aware planner is employed to adaptively decide the time and location to recharge a certain robot. The coordination between robots is based on market economy, giving robots the flexibility to incorporate multiple objectives into a single cost function. Different policies are evaluated through simulations with different robot team sizes on different maps. Results demonstrate the applicability for multi-robot exploration and show the strengths and weaknesses of each policy.
Ideally, the robots collaborate to push the energetic horizon to longer distances. This is achieved by forming teams where some robots are exploring and other robots assist as mobile rechargers. In other words, robots recharge not only at fixed docking stations, but some mobile robots act as energy transporters to form supply chains to deliver energy to explorer robots.
Selected Publications
- Collaboration in multi-robot exploration
Journal of Intelligent and Robotic Systems, May 2016
Additional Publications
- Coordinated recharging of mobile robots during exploration
- Coordinated multi-robot exploration: Out of the box packages for ROS
- Assessing the value of coordination in mobile robot exploration …
- On path planning strategies for networked unmanned aerial vehicles
- A discrete stochastic process for coverage analysis of autonomous UAV networks
- Area coverage with unmanned vehicles: A belief-based approach
Videos