Aquatic environment has been facing an increasing number of threats from various harmful aquatic processes such as oil spills, harmful algal blooms (HABs), and aquatic debris. These processes greatly endanger the aquatic ecosystems, marine life, human health, and water transport. Hence, it is of great interest to detect these processes and monitor their evolution such that proper actions can be taken to prevent the potential risks. This dissertation explores four representative problems in aquatic environment monitoring, which include diffusion processing profiling, spatiotemporal field reconstruction, aquatic debris surveillance, and water surface monitoring.First, we propose an accuracy-aware approach to profiling an aquatic diffusion process such as oil spill. In our approach, the robotic sensors collaboratively profile the characteristics of a diffusion process including its source location, discharged substance amount, and evolution over time. In particular, the robotic sensors reposition themselves to progressively improve the profiling accuracy. We formulate a novel movement scheduling problem that aims to maximize the profiling accuracy subject to limited sensor mobility and energy budget. To solve this problem, we develop an efficient gradient-ascent-based algorithm and a near-optimal dynamic-programming-based algorithm.Second, we present a novel approach to reconstructing a spatiotemporal aquatic field such as HABs. This approach features a rendezvous-based mobility control scheme where robotic sensors collaborate in the form of a swarm to sense the aquatic environment in a series of carefully chosen rendezvous regions. We design a novel feedback control algorithm that maintains the desirable level of wireless connectivity for a sensor swarm in the presence of significant environment and system dynamics. Moreover, information-theoretic analysis is used to guide the selection of rendezvous regions so that the reconstruction accuracy is maximized subject to the limited sensor mobility.Third, we develop a vision-based, cloud-enabled, low-cost, yet intelligent solution to aquatic debris surveillance. Our approach features real-time debris detection and coverage-based rotation scheduling algorithms. Specifically, the image processing algorithms for debris detection are specifically designed to address the unique challenges in aquatic environment, e.g.,, constant camera shaking due to waves. The rotation scheduling algorithm provides effective coverage of sporadic debris arrivals despite camera's limited angular view. Moreover, we design a dynamic task offloading scheme to offload the computation-intensive processing tasks to the cloud for battery power conservation.Finally, we design Samba -- an aquatic surveillance robot that integrates an off-the-shelf Android smartphone and a robotic fish for general water surface monitoring. Using the built-in camera of on-board smartphone, Samba can detect spatially dispersed aquatic processes. To reduce the excessive false alarms caused by the non-water area, Samba segments the captured images and performs target detection in the identified water area only. We propose a novel approach that leverages the power-efficient inertial sensors on smartphone to assist the image processing. Samba also features a set of lightweight and robust computer vision algorithms, which detect harmful aquatic processes based on their distinctive color features.
Read
