Developing surface classification models manually requires significant time and detracts from the goal of automating systems. We create a system that automatically collects the data using an Unmanned Aerial Vehicle (UAV), extracts features, trains a large number of classifiers, selects the best classifier, and programs the UAV with that classifier. Motivating our work is a prior project [1] that manually developed a surface classifier using an accelerometer; to verify our system functionality, we replicate those results with our new automated system and improve on those results, providing a four-surface classifier with a 75% classification rate and a hard/soft classifier with a 100% classification rate. We further verify our system through a field experiment that collects and classifies new data, proving its end-to-end functionality. Overall, our system reduces the time and machine learning expertise needed by the user to develop new time-series classifiers usable by the UAV. The general form of our system provides a valuable tool for automation of classifier creation and is released as an open-source tool [2].

Using Unmanned Aerial Vehicles (UAVs) to deploy sensor networks promises an autonomous and useful method of installation in remote or hard to access locations. Some sensors, such as soil moisture sensors, must be physically installed in soft soil, yet UAVs cannot easily determine soil softness with remote sensors. In this paper, we use data from an onboard accelerometer measured during UAV landings to determine the softness of the ground. We collect and analyze over 200 data sets gathered from 8 different materials: foam, carpet, wood, tile, grass, dirt, concrete, and woodchips. Based on this analysis, we examine a number of features from the accelerometer and four classification algorithms: LDA, QDA, SVM, and binary decision trees. The decision tree performs well and is simple to implement onboard the UAV. We implement this in our UAV control system and perform experiments to verify that the UAV can accurately classify the softness of the surface with 90% accuracy. This lays the groundwork for our future work on developing a UAV capable of installing sensors in soft soil.

Ocean monitoring using underwater sensor networks faces communication challenges in retrieving data, communicating large amounts of data between nodes, and covering increasing spatial regions while remaining connected. With underwater sensor networks that are capable of surfacing, unmanned aerial vehicles (UAVs) provide a solution to this by providing radio-based data muling services, but, as this area is still unexplored, the utility of this solution is unclear. In this paper, we examine the theoretical expectations, perform several field experiments, and analyze the communication success rates of 802.15.4 radios near the water surface both communicating between surface nodes as well as between a node and the UAV. These indicate that on the water surface internode radio communication is poor, but node to UAV communication can provide both reasonable ranges and success rates. We additionally measure and analyze the energy aspects of the systems, determining the impacts of parameters such as network size and distance between nodes on the UAV energy. Finally, we consolidate the information into an algorithm outlining how to configure and design hybrid UAV and underwater sensor network systems.

Long-term sensor network deployments demand careful power management. While managing power requires understanding the amount of energy harvestable from the local environment, current solar prediction methods rely only on recent local history, which makes them susceptible to high variability. In this article, we present a model and algorithms for distributed solar current prediction based on multiple linear regression to predict future solar current based on local, in situ climatic and solar measurements. These algorithms leverage spatial information from neighbors and adapt to the changing local conditions not captured by global climatic information. We implement these algorithms on our Fleck platform and run a 7-week-long experiment validating our work. In analyzing our results from this experiment, we determined that computing our model requires an increased energy expenditure of 4.5mJ over simpler models (on the order of 10 -7 % of the harvested energy) to gain a prediction improvement of 39.7%.

Robots improve wireless sensor network (WSN) deployments by reducing deployment times, deploying nodes to improve coverage, and ferrying data. Utilizing Unmanned Aerial Vehicles (UAVs) to install sensor networks in environmentally sensitive areas is especially valuable, as the UAVs are able to quickly traverse rough and environmentally sensitive terrain. UAV based deployments are challenging, as the UAVs may need to install nodes in a specific orientation or location type, which is difficult to sense from a UAV. We present our work towards resolving these difficulties by first classifying the surface a UAV has landed on, and then conducting a post-deployment analysis of the installation.

Predictive environmental sensor networks provide complex engineering and systems challenges. These systems must withstand the event of interest, remain functional over long time periods when no events occur, cover large geographical regions of interest to the event, and support the variety of sensor types needed to detect the phenomenon. Prediction of the phenomenon on the network complicates the system further, requiring additional computation on themicrocontrollers and utilizing prediction models that are not typically designed for sensor networks. This paper describes a system architecture and deployment to meet the design requirements and to allow model-driven control, thereby optimizing the prediction capability of the system. We explore the application of river flood prediction using this architecture, describing our work on a centralized form of the prediction model, network implementation, component testing and infrastructure development in Honduras, deployment on a river in Massachusetts, and results of the field experiments. Our system uses only a small number of nodes to cover basins of 1000-10000 square km2 using an unique heterogeneous communication structure to provide real-time sensed data, incorporating self-monitoring for failure, and adapting measurement schedules to capture events of interest.

In developing countries, flooding due to natural disasters such as hurricanes and earthquakes results in massive loss of life and property. Warning communities of the incoming flood provides an effective solution to this by giving people sufficient time to evacuate and protect their property. However, the range of early warning system solutions introduces a tangle of conflicting requirements including cost and reliability, and creates several interesting problems from factors as diverse as technological, social, and political. The complexity of these systems and need for autonomy within the context of a developing country while remaining maintainable and accessible by non-technical personnel provides a challenge not often solved within developed countries, much less the developing. After describing this problem, the paper discusses a proposed solution for the problem, initial experiments in implementing the solution, and lessons learned through that work.