Abstract This paper proposes an approach to pattern-based modeling and Uppaal-based verification for ROS 2 applications. The proposed verification focuses on callback execution latencies and buffer overflow. We propose formal model templates to model the execution of ROS 2 system components, created using a pattern-based approach. The model templates simplify the formal modeling of an ROS 2 application. Using Uppaal, we model in Uppaal timed automata, allowing the description of computation chains of ROS 2-based applications. Our focus is on execution behavior, including two versions of the mainline single-threaded executor of ROS 2. System traces generated using the formal models are validated in multiple experiments. Furthermore, we compare two approaches to modeling the execution of nodes that are typically the core units of computation of ROS 2. The first approach is a holistic approach to model ROS 2 applications, including communication and execution in computation chains. The second is an approach for individual nodes only, at a higher abstraction level. Additionally, we show the application of the verification by model checking in two ROS 2 system scenarios where we compare generated model traces to actual system executions. Overall, through formal modeling and verification, we showcase the potential for uncovering errors in the execution of distributed robotic systems.

Robot operating system 2 (ROS 2) is increasingly popular both in research and commercial robotic systems. ROS 2 is designed to allow real-time execution and data communication, enabling rapid prototyping and deployment of robotic systems. In order to predict and calculate execution times in ROS 2, one needs to analyze its internal scheduler, called executor. The executor has been updated in various distributions of ROS 2, which is shown to impact significantly the periodic execution invoked by the underlying operating system’s timers, potentially causing unexpected latencies. To expose the mentioned impact due to executor differences, in this paper, we present an experimental evaluation of the execution behavior of ROS 2’s schedulable entities, namely callbacks, among the existing versions of the executor. We visualize the differences of callback execution order via simulation, and we create design-level scenarios that impact the execution of periodically scheduled callbacks, negatively. Moreover, we show how such negative impact can be mitigated by using multi-threaded executors. Finally, we illustrate the observed behavior on a real-world centralized multi-agent robot system. Our work aims to raise awareness within the ROS 2 developer community, regarding possible problems of timer blocking, and propose a mitigation solution of the latter.

Multi-agent robot systems, specifically mobile robots in dynamic environments interacting with humans, e.g., assisting in production environments, have seen an increased interest over the past years. To better understand the ROS2 communication in a network with a high load of nodes, this paper investigates the communication handling of multiple robots to a single tracking node for centralized multi-agent robot systems using ROS2. Thereore, a quantitative analysis of two publisher-subscriber communication architectures and a comparative study between DDS vendors (CycloneDDS, FastDDS and GurumDDS) using ROS2 Galactic is performed. The architectures of consideration are a many-to-one approach, where multiple robots communicate to a central node over one topic, and the one-to-one communication approach, where multiple robots communicate over particular topics to a central node. Throughout this work, the increase in the number of robots at different publishing rates is simulated on a single computer for the different DDS vendors. A further simulation is done using a distributed setup with CycloneDDS. The simulations show that with an increase in the number of nodes, the average data age and the data miss ratio in the one-to-one approach were significantly lower than in the many-to-one approach. CycloneDDS was shown as the most robust regarding crashes and response time under system launch, while FastDDS showed better results regarding the data ageing.

Federated learning (FL) is an emerging and privacy-preserving machine learning technique that is shown to be increasingly important in the digital age. The two challenging issues for FL are: (1) communication overhead between clients and the server, and (2) volatile distribution of training data such as class imbalance. The paper aims to tackle these two challenges with the proposal of a federated fuzzy learning algorithm (FFLA) that can be used for data-based construction of fuzzy classification models in a distributed setting. The proposed learning algorithm is fast and highly cheap in communication by requiring only two rounds of interplay between the server and clients. Moreover, FFLA is empowered with an an imbalance adaptation mechanism so that it remains robust against heterogeneous distributions of data and class imbalance. The efficacy of the proposed learning method has been verified by the simulation tests made on a set of balanced and imbalanced benchmark data sets.