Decentralized Decision Making in Adaptive Multi-robot Teams

(De Gruyter) – Robots have entered into many areas of our lives with the future outlooks leaning towards teams of robots showing up in ever more tasks. A team of robots is able to perform a task that exceeds the capabilities of a single robot.

Middleware frameworks have been proposed to address the requirements of robot communication when working as a team. However, the design and implementation of cooperative multi-robot behaviors is a complex and time-consuming task.

Previously, the authors had designed a modeling framework called “A Language for Interactive Cooperative Agents (ALICA)”. ALICA is geared towards highly dynamic domains, such as robot soccer and autonomous driving, where the robots must cope with potentially unreliable communication and noisy sensors. In such environments only a fully decentralized decision process makes sense to avoid a single point of failure and degraded decision time performance. Each ALICA agent makes its own observations and takes decisions according to the common team plan, based on its own observations and the observations of its team colleagues that share their current view of the world via wireless communication. Thus, the overall team behavior results from individual “informed” decisions of the autonomous team members.

The potential diversity of applications for multi-robot systems has shown that the ALICA framework is insufficient for a number of scenarios. Therefore, the middleware PROViDE extends ALICA and adds flexibility of choosing an appropriate proposal sharing and decision making strategy in order to support application-specific team decision making in potentially adverse communication environments.

Within their article the authors discuss the middleware requirements of such a multi-robot decision making scenario in adverse environments and present their own specific middleware solution called PROViDE (Proposal Replication for Value Determination) that offers a choice of protocols and decision methods for decentralized decision making.

The authors presents a short overview of ALICA and discuss specific middleware requirements and why conventional middleware is not appropriate, the specific features of PROViDE, and the evaluation of the middleware and design decisions from different perspectives.

Since application requirements in terms of performance and safety may be quite diverse and application specific, PROViDE offers a choice of protocols and strategies that application designers can select to adapt the middleware to the application requirements. PROViDE was designed with small to medium sized robot teams in mind, i. e. not more than 15 robots per team. In swarm scenarios, the coordination overhead using PROViDE would increase beyond acceptable limits and one would need to consider other kinds of coordination algorithms, protocols and consistency guarantees.

PROViDE middleware and ALICA modeling framework together form a strong team that facilitates substantially the development of robust multi-robot solutions.

***

Edited for Content and Length by Dr. Matthew A. Hood.

The full article can be found at De Gruyter in the journal of it – Information Technology.

DOI: 10.1515/itit-2017-0029

Pin It on Pinterest