GIFTED-MRS · Generic fault-detection for multirobot systems
FP7 — People (Marie Curie Actions)
- Duration
- 2015-03-01 → 2017-02-28
- EU contribution
- €221,606
- Participants
- 1
- Scheme
- MC-IEF
Lines connect the coordinator with its partners.
Results in brief
Generic fault-detection for multirobot systems
Robot swarms are large-scale multirobot systems (robot teams) with decentralized control which means that each robot acts based only on local perception and on local coordination with neighboring robots. The decentralized approach to control confers number of potential benefits. In particular, inherent scalability and robustness are often highlighted as key distinguishing features of robot swarms. It has, however, been shown that swarm robotics systems are not always fault tolerant. To realize the robustness potential of robot swarms, it is thus essential to give systems the capacity to actively detect and accommodate faults. This is the motivation underlying our project In the GiFteD-MRS project, we have developed a generic fault-detection system for robot swarms, thus giving the swarm the capability to monitor itself and signal the presence of faulty robots. We demonstrate how robots with limited and imperfect sensing capabilities are able to observe and classify the behavior of one another. In order to achieve this, the underlying classifier is an immune system-inspired algorithm that learns to distinguish between normal behavior and abnormal behavior online. Through a series of experiments, we systematically assess the performance of our approach for a large swarm of robots. In particular, we analyze our system's capacity to correctly detect robots with faults, the time required to detect a fault since it first occurred in the robot, and the false-alarm rate. Our results show that our developed generic fault-detection system is robust, that it is able to detect faults in a timely manner (in less than 3 minutes for more than 90% of the tested conditions), and that it achieves a low false-alarm rate (mean less than 2% of the experiment duration). Robot fault detection and fault tolerance are two of the most important problems in the field of robotics. Robot swarms in many real world scenarios, operating in unstructured environments for instance, require a fault-detection system that can adapt to temporal variations in the robot's behavior and perturbations to the environment. The GiFteD-MRS fault-detection system demonstrates these capabilities. Therefore, our developed fault-detection system has the potential to assist in long-term autonomous operation with minimal human intervention, thus increasing the usefulness of robots for a diverse repertoire of upcoming applications in distributed intelligent automation, such as environmental monitoring, and agriculture automation.
Data: CORDIS, © European Union
Project objective
Multirobot systems (MRS) have recently been of great interest to environmental scientist, consequent to their ability to monitor large-scale atmospheric and aquatic environmental processes. Furthermore, the deployment of large numbers of relatively inexpensive robots for such monitoring purposes may soon be possible. However, considerable technological advances in platform reliability and endurance are required, to potentiate the wide-spread deployment and usage of MRS for environmental monitoring.Fault tolerance is one of the most prominent challenges in the field of MRS. Efficient and long term operation of a MRS requires an accurate and timely detection, and accommodation of abnormally behaving robots. This is particularly relevant in environmental monitoring scenarios, wherein an undetected faulty robot may interfere with, and possibly damage the very system being monitored. Most existing fault tolerant systems prescribe a characterization of the normal robot behaviors, and train the fault-detection model to recognize these behaviors. Behaviors not recognized by the model are consequently labelled abnormal or faulty. However, these models assume a priori knowledge of normal behavior. In addition, MRS employing these models do not transition well to scenarios involving temporal changes in behavior (e.g., robots change their behavior through learning, or in response to environment perturbations).The applicant proposes to develop a generic fault-detection system for a real-world environmental monitoring MRS. The developed system will be capable of robustly detecting faults, while adapting itself online to changes in the robot collectives behavior, thus avoiding the need to retrain the system for any new exhibited behavior. The developed system would also have a significant impact on long-term operations of MRS operating in other upcoming areas, such as the health careindustry (e.g., potential deployment of MRS in hospitals to interact with patients).
Original text from CORDIS.
Participants
- UNIVERSITY OF YORK · YORK NORTH YORKSHIRECoordinatorUnited Kingdom
Links
Data: CORDIS, © European Union