The light detection and ranging (LiDAR) technology allows to sense surrounding objects with fine-grained resolution in a large areas. Their data (aka point clouds), generated continuously at very high rates, can provide information to support automated functionality in cyberphysical systems. Clustering of point clouds is a key problem to extract this type of information. Methods for solving the problem in a continuous fashion can facilitate improved processing in e.g. fog architectures, allowing continuous, streaming processing of data close to the sources. We propose Lisco, a singlepass continuous Euclidean-distance-based clustering of LiDAR point clouds, that maximizes the granularity of the data processing pipeline. Besides its algorithmic analysis, we provide a thorough experimental evaluation and highlight its up to 3x improvements and its scalability benefits compared to the baseline, using both real-world datasets as well as synthetic ones to fully explore the worst-cases.
Considering the needs for continuous availability of information out of data generated in Cyber-Physical production systems, we investigate the use of continuous stream processing as a paradigm for generating useful information out of the data, to support efficient and safe operation, as well as planning activities.Our contributions and expected benefits: (i) we show possibilities to automate and pipeline the validation and analysis of the data, hence providing an automated way to improve the quality of the latter and parallelizing the two phases; (ii) we show how to induce lower latency in generating the desired information, enabling it to be continuously made available, before whole batches of data are gathered, in cost-efficient ways; (iii) besides the automation of the above procedures that are commonly done in a batch fashion and with significant manual effort by the production system analysts, we show additional options for configuring ways in which to automate deeper analysis of the data; in particular, we provide evidences about how the rich semantics of stream processing frameworks can ease the development and deployment of data analysis applications in production systems.Moreover, using the problem of bottleneck detection as a sample scenario, we illustrate the above in a concrete fashion, on cost-efficient systems, that are plausible to have in existing deployments. The experimental study is on a 2-year data-set with more than 8.5 million entries, from a system including more than 30 interconnected machines and it demonstrates the benefits of the proposed methods, in providing timely and multidimensional information from the data, enabling possibilities for deeper analyses.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.