Francesco Peri

Machine reasoning about anomalous sensor data

Abstract We describe a semantic data validation tool that is capable of observing incoming real-time sensor data and performing reasoning against a set of rules specific to the scientific domain to which the data belongs. Our software solution can produce a variety of different outcomes when a data anomaly or unexpected event is detected, ranging from simple flagging of data points, to data augmentation, to validation of proposed hypotheses that could explain the phenomenon. Hosted on the Jena Semantic Web Framework, the tool is completely domain-agnostic and is made domain-aware by reference to an ontology and Knowledge Base (KB) that together describe the key resources of the system being observed. The KB comprises ontologies for the sensor packages and for the domain; historical data from the network; concepts designed to guide discovery of internet resources unavailable in the local KB but relevant to reasoning about the anomaly; and a set of rules that represent domain expert knowledge of constraints on data from different kinds of instruments as well as rules that relate types of ecosystem events to properties of the ecosystem. We describe an instance of such a system that includes a sensor ontology, some rules describing coastal storm events and their consequences, and how we relate local data to external resources. We describe in some detail how a specific actual event—an unusually high chlorophyll reading—can be deduced by machine reasoning to be consistent with being caused by benthic diatom resuspension, consistent with being caused by an algal bloom, or both.

Adaptive Decentralized Control of Mobile Underwater Sensor Networks and Robots for Modeling Underwater Phenomena

Understanding the dynamics of bodies of water and their impact on the global environment requires sensing information over the full volume of water. In this article, we develop a gradient-based decentralized controller that dynamically adjusts the depth of a network of underwater sensors to optimize sensing for computing maximally detailed volumetric models. We prove that the controller converges to a local minimum and show how the controller can be extended to work with hybrid robot and sensor network systems. We implement the controller on an underwater sensor network with depth adjustment capabilities. Through simulations and in-situ experiments, we verify the functionality and performance of the system and algorithm.

Design and implementation of a wireless video camera network for coastal erosion monitoring

Abstract The short-term rate of coastal erosion and recession has been observed at island shoreline bluffs near waterways among Boston Harbor, Massachusetts, USA. This erosion has been hypothesized partially related to waves from high-speed wakes. Recording the physical erosion events during extreme high waves is significant to evaluate the dynamics of bluff erosion and to document these short-term processes. Still and motion imagery are important media to observe rare and extreme events in ecology, geology, and environmental condition. The study of coastal erosion requires recording devices for these modalities capable of long-term, low-cost, low-power operation with low maintenance, and with the ability to support a large dynamic range in both time and space. We describe recent work in the development of a wireless video camera network for an ecosystem observation platform. These cameras are enclosed in weatherproof housings and supported by solar energy harvesting. The cameras are Internet-enabled and thus live video can be accessed remotely. Video streams are transmitted via wireless network, and delivered to and stored at a remote server. This system has been functional as designed since installation in October 2012 on Thompson Island, Massachusetts, and is expected to operate indefinitely. To date, a number of erosion-related events have been successfully captured. This platform has shown the potential to be used in a large scale for a variety of environmental monitoring studies.

Classifying ecosystems with metaproperties from terrestrial laser scanner data

Abstract In this study, we introduce metaproperty analysis of terrestrial laser scanner (TLS) data, and demonstrate its application through several ecological classification problems. Metaproperty analysis considers pulse level and spatial metrics derived from the hundreds of thousands to millions of lidar pulses present in a single scan from a typical contemporary instrument. In such large aggregations, properties of the populations of lidar data reflect attributes of the underlying ecological conditions of the ecosystems. In this study, we provide the Metaproperty Classification Model to employ TLS metaproperty analysis for classification problems in ecology. We applied this to a proof‐of‐concept study, which classified 88 scans from rooms and forests with 100% accuracy, to serve as a template. We then applied the Metaproperty Classification Model in earnest, to separate scans from temperate and tropical forests with 97.09% accuracy (N = 224), and to classify scans from inland and coastal tropical rainforests with 84.07% accuracy (N = 270). The results demonstrate the potential for metaproperty analysis to identify subtle and important ecosystem conditions, including diseases and anthropogenic disturbances. Metaproperty analysis serves as an augmentation to contemporary object reconstruction applications of TLS in ecology, and can characterize regional heterogeneity.

Seasonal Export of Dissolved Organic Matter from a New England Salt Marsh

ABSTRACT Schiebel, H.N.; Gardner, G.B.; Wang, X.; Peri, F., and Chen, R.F., 2018. Seasonal export of dissolved organic matter from a New England salt marsh. The seasonality of dissolved organic matter (DOM) export from a New England salt marsh was investigated using multiple estuary transects along with laboratory plant leaching experiments. From 2001 through 2013, 35 transects of the Neponset Estuary in Boston, Massachusetts, were conducted to measure high-resolution in situ chromophoric dissolved organic matter (CDOM) fluorescence, along with other standard oceanographic parameters, via sensors incorporated into an undulating depressor wing known as the mini-shuttle. Water samples were collected during each cruise for discrete dissolved organic carbon (DOC) and CDOM fluorescence analyses. Seasonal salt marsh DOC export rates ranged from 1.1 × 107 to 15 × 107 mol C season−1 (10–140 mol C m−2 season−1). The total export was found to be 23 ± 17 × 107 mol C y−1 (210 mol C m−2 y−1) with river discharge, temperature, and day length significantly affecting this rate. Results from seasonal salt marsh vegetation (Spartina alterniflora, Spartina patens, and Phragmites australis) incubation experiments showed that the leaching of DOC and CDOM from plant biomass was a rapid process occurring over several days. Sunlight exposure, biomass type (above- versus belowground biomass), and microbial activity had significant effects on DOC leaching rates from plant matter. These incubations, in conjunction with the cruise data, support the fall dump hypothesis, in which a pulse of DOC leaches from plants during the fall as marsh plants begin to senesce for winter. The long-term trend in salt marsh DOM contributions to the estuary suggested a decrease in the in situ inputs from the marsh to the estuary during the length of this study.

Bounding uncertainty in volumetric geometric models for terrestrial lidar observations of ecosystems

Volumetric models with known biases are shown to provide bounds for the uncertainty in estimations of volume for ecologically interesting objects, observed with a terrestrial laser scanner (TLS) instrument. Bounding cuboids, three-dimensional convex hull polygons, voxels, the Outer Hull Model and Square Based Columns (SBCs) are considered for their ability to estimate the volume of temperate and tropical trees, as well as geomorphological features such as bluffs and saltmarsh creeks. For temperate trees, supplementary geometric models are evaluated for their ability to bound the uncertainty in cylinder-based reconstructions, finding that coarser volumetric methods do not currently constrain volume meaningfully, but may be helpful with further refinement, or in hybridized models. Three-dimensional convex hull polygons consistently overestimate object volume, and SBCs consistently underestimate volume. Voxel estimations vary in their bias, due to the point density of the TLS data, and occlusion, particularly in trees. The response of the models to parametrization is analysed, observing unexpected trends in the SBC estimates for the drumlin dataset. Establishing that this result is due to the resolution of the TLS observations being insufficient to support the resolution of the geometric model, it is suggested that geometric models with predictable outcomes can also highlight data quality issues when they produce illogical results.