Industrial applications increasingly require fleets of autonomous ground vehicles (AGVs). One of these is warehouse management, where autonomous machines such as forklifts are used to transport incoming items to storage facilities, and subsequently on to outgoing queues. Many of these settings require the deployment of very large fleets, in the order of hundreds of machines. At the same time, fleet management systems must be able to respond to changes very rapidly.
In intralogistics, for example, fleets of autonomous forklifts have to adapt to changing warehouse layouts, the addition/removal of forklifts, dynamic priorities, and new storage and shelving areas. Fleet behaviour must meet certain operational requirements, and it must be possible to specify these requirements at a high-level of abstraction.
Kollmorgen, a leading developer of autonomous systems for warehouse management, is contributing this problem to Semantic Robots’ research profile. Together, we are studying efficient online task allocation and scheduling algorithms for large fleets. Kollmorgen’s personnel and Semantic Robots’ researchers are engaged closely in a co-development process, with the former providing APIs for algorithm development, and the latter contributing realization of the algorithms.