PhD, University of Cincinnati, 2024, Engineering and Applied Science: Electrical Engineering

Echolocating bats demonstrate sonar's remarkable ability to support various complex tasks, from navigating dense vegetation to foraging in flight. However, achieving these tasks using echolocation alone presents challenges due to the inherent limitations of sonar. Drawing inspiration from biological systems, we can narrow the performance gap between artificial and biological sonar. In this context, we develop behavior-based systems using sensorimotor loops and their interactions to accomplish the foraging task. We showcase a solution to the foraging task by training machine learning models to arbitrate between the approach and avoid sensorimotor loops using solely echoes as perception inputs. These models demonstrate a high success rate in foraging for insect-like targets and flower targets. In addition, they are robust against different arena geometries, traps, and noisy motor control. While arbitrating between the two sensorimotor loops proves effective, the system does not demonstrate attention to a single target, failing to investigate a flower target thoroughly. To address this, we develop a novel sensorimotor loop, termed the flower home-in loop, designed to guide nectarivorous bats toward a flower opening while maintaining attention on the target. The flower home-in loop perfectly performs when approaching the flower from the front. However, the performance gets worse when approaching the flower from other sides. Continuously re-estimating the target flower pose and updating the bat's path is a more robust strategy for the flower home-in loop in acoustically ambiguous scenarios. The flower home-in loop can also guide the bat toward the echoes from a flower's opening while being interfered with by echoes from nearby objects.

Committee: Dieter Vanderelst Ph.D. (Committee Chair); Ali Minai Ph.D. (Committee Member); Herbert Peremans Ph.D M.A B.A. (Committee Member); John Gallagher Ph.D. (Committee Member); Zachariah Fuchs Ph.D. (Committee Member) Subjects: Robotics

MS, University of Cincinnati, 2023, Arts and Sciences: Biological Sciences

Aquatic habitat is one of the most important drivers of sportfish population dynamics and behavior in freshwater ecosystems, and fisheries management efforts often concentrate on the inventorying, monitoring, and enhancement of aquatic habitat features. However, the collection of aquatic habitat data has been historically difficult due to the limitations and deficiencies of traditional, quadrat-based sampling methods. Specifically, traditional sampling methods are often arduous, time-intensive, and limited by environmental factors such as depth and turbidity. Side-scan sonar has been identified as a tool that can improve aquatic habitat sampling. Side-scan sonar systems are able to efficiently collect benthic imagery encompassing a wide range of aquatic habitat features and are not subject to the same limitations as traditional sampling methods. The primary purpose of this project was to develop a standardized aquatic habitat assessment methodology using side-scan sonar that can be implemented by the Ohio Department of Natural Resources – Division of Wildlife (ODNR-DOW) in reservoirs throughout the state. This methodology will allow for the creation of large-scale habitat inventories and will provide insight into potential management actions by ODNR-DOW that can improve sport fisheries throughout Ohio. As part of developing this methodology, we first compared the relative utility and performance of a recreational Lowrance and survey-grade EdgeTech side-scan sonar system to determine which was the superior tool for collecting littoral habitat data in Ohio reservoirs. We manually quantified submerged woody debris, standing timber, aquatic vegetation, and benthic substrate using imagery from each sonar system and compared habitat classification accuracy, habitat values, GIS processing times, and the level of variation in habitat values generated by separate GIS users/operators. Based on our results, we ultimately concluded that recreational side-scan sonar systems s (open full item for complete abstract)

Committee: Michael Booth Ph.D. (Committee Chair); Jeremy Pritt Ph.D. (Committee Member); Stephen Matter Ph.D. (Committee Member) Subjects: Natural Resource Management

Master of Science (MS), Bowling Green State University, 2022, Biological Sciences

Sandusky Bay is both a valuable nursery for important sport and commercial fishes like Walleye and White Bass, and a degraded system with excessive sediment and nutrient inputs from the agricultural watershed. Proposed restorations could alter the bay from a uniformly turbid system to one capable of supporting submerged aquatic vegetation and accompanying changes in the fish community. The current bay conditions support rapid growth of larval and juvenile Walleye and Yellow Perch, but it also harbors Channel Catfish and Walleye fisheries. With the potential changes in habitat through restoration, I aimed to quantify the abundance and composition of large fish predators (piscivores) before proposed restoration begins, so the possible impacts to the bay as a nursery habitat can be assessed. Fish abundance was determined in early June 2021 when larval and juvenile fishes are in the bay. Using Dual Frequency Identification Sonar (DIDSON) and accompanying software, I quantified the abundance of fishes in Sandusky Bay (i.e., dividing the bay into four sectors, with multiple recordings in each sector). Due to species-specific size distributions, I divided the observations into size classes (100-250mm, 251-450mm, >451-mm total length (TL)) and then coupled these with proportions of species by size class from experimental gillnet sets in each sector to obtain abundance by species. By using the areal estimator in the ODNR Coastal Viewer and correcting for actual depth of water surveyed, I extrapolated these densities to determine that there were 29.6 ± 9.3 million (mean ± 1SE) fish >100mm TL in Sandusky Bay at the time we sampled. Gizzard Shad dominated the system (46% of all fish), but notably ~4.1 million Channel Catfish and ~2.9 million Walleye inhabit the bay. With absolute numbers, bioenergetic models may be built to predict community-wide impacts to larval and juvenile fishes and predict how restoration may affect this fish community.

Committee: Jeffrey Miner PhD (Committee Chair); Shannon Pelini PhD (Committee Member); Joseph Schmitt PhD (Committee Member); Daniel Weigmann PhD (Committee Member) Subjects: Aquatic Sciences

MS, University of Cincinnati, 2020, Engineering and Applied Science: Electrical Engineering

Recent evidence suggests that sonar provides bats only with limited information about the environment. Nevertheless, they can fly swiftly through dense environments while avoiding obstacles. Previously, we proposed a model of sonar-based obstacle avoidance that only relied on the interaural level difference of the onset of the echoes. In this paper, we extend this previous model. In particular, we present a model that (1) is equipped with a short term memory of recent echo trains, and (2) uses the full echo train. Because handcrafting a controller to use more sonar data is challenging, we resort to machine learning to train a robotic model. We find that both extensions increase performance and conclude that these could be used to enhance our models of bat sonar behavior. We discuss the implications or our method and findings for both biology and bio-inspired engineering.

Committee: Dieter Vanderelst Ph.D. (Committee Chair); Zachariah Fuchs Ph.D. (Committee Member); Ali Minai Ph.D. (Committee Member) Subjects: Artificial Intelligence

MS, University of Cincinnati, 2020, Engineering and Applied Science: Mechanical Engineering

Echolocating bats use sonar for navigating to salient locations. Navigation requires the recognition of locations. Previously, we proposed template-based place recognition to explain how bats recognize locations using sonar. We hypothesized that bats recognize places based on their echo signature and do not compute the 3D layout of a scene. We collected a large corpus of echoes from different habitats and processed them using a model of the bat's auditory system. We showed that biologically plausible templates contain sufficient information to recognize previously visited locations. In this previous work, we assumed that the bat has perfect memory. We assumed that the bat stores all previously encountered templates and recognizes locations by comparing a new template with all stored templates. As the bat would need to store many templates, this kind of perfect memory is computationally expensive and biologically implausible. In this work, we explore how bats could store and use templates more efficiently using neural networks, a biologically plausible substrate for the memory of bats. We trained a feed-forward neural architecture to return the location and viewing direction associated with a given template as a model of a bat recognizing its current location and ight direction from the echoes. The performance of this imperfect memory approach was comparable to the previously implemented perfect memory, where performance was fundamentally limited by the noise inherent in sonar data and the signal to noise ratio of the templates. Hence, our results indicate that a large corpus of sonar data can be stored in a neural network (which is small compared to the neural capacity available to bats). This work indicates that our hypothesized mechanism for scene recognition is biologically plausible. In addition, our work opens up avenues for more efficient implementations of sonar-based navigation for robots using neuromorphic hardware.

Committee: Dieter Vanderelst Ph.D. (Committee Chair); Manish Kumar Ph.D. (Committee Member); Ali Minai Ph.D. (Committee Member) Subjects: Mechanical Engineering

MS, University of Cincinnati, 2019, Engineering and Applied Science: Electrical Engineering

Sonar-based navigation systems have been immensely used for over a century. Underwater navigation is primarily achieved using Sonar, as Laser-based systems are largely ineffective due to the darkness of the oceans. However, the use of Sonar waves for navigation other than underwater has been very limited. The rise in demands of autonomous systems like self-driving cars and unmanned aerial vehicles have posed new challenges in designing sensor systems. Sonar-based sensors, with the unique properties of sound waves, present themselves as a possible solution, especially in scenarios where visibility is heavily compromised by smoke, dust or other suspended particles. The motivation for using Sonar arises from the flight pattern of bats. Bats are able to effectively navigate in dense vegetation using echolocation and are able to map surroundings for which other animals rely on their vision. In this thesis, we present the design of a data acquisition system with the ability to simulate bat echolocation by generating sonar-waves and recording the resulting echoes. The first part of the thesis focuses on the hardware circuit design for the ultrasonic peripherals. A software interface for the hardware circuit is developed which allows control and flexibility over the hardware not possible on existing sonar-based solutions in the market. To evaluate the practicality of the sonar-based sensor, we design an experimental setup to quantify its performance using resultant echoes, in conditions where visibility is compromised. An artificially generated fog and a LIDAR-based sensor are used to simulate the environment. Experiments are carried out to determine the effect of distance and fog density on the sensor's performance. The simulated environment and the sonar-board are also used to determine the response of different frequencies in the sonar spectrum. We conclude that the sonar-based board can be effectively used in environments with very high levels of fog to detect obj (open full item for complete abstract)

Committee: Ranganadha Vemuri Ph.D. (Committee Chair); Wen-Ben Jone Ph.D. (Committee Member); Dieter Vanderelst Ph.D. (Committee Member) Subjects: Electrical Engineering

Master of Science (MS), Bowling Green State University, 2015, Biological Sciences

Many of the lake-dwelling fish species of Lake Erie rely on shallow, heavily vegetated bays as spawning grounds to increase offspring probability of survival during early life stages. Multiple complex abiotic and biotic factors can affect mortality especially during early life stages; the loss or absence of suitable habitat is one of these key factors leading to poor recruitment of fish species. Submerged aquatic vegetation (SAV)/macrophyte beds in clearwater systems act as refuges for juvenile fish decreasing mortality from predation while foraging on prey resources. However, it also has been shown that river discharge “plumes” (areas of high turbidity) may act as habitat/refuge for young-of-the-year fishes. The Maumee River and Maumee Bay, once with abundant macrophyte beds, have experienced substantive increases in suspended solids over the last century. Historical introduction of benthivorous feeding carp (especially Cyprinus carpio), sediment pollution from surface runoff in the surrounding watershed and relatively high wave energy further increases the levels of turbidity in bays and decrease the amount of SAV habitat. The potential colonization of western Lake Erie by Grass Carp (Ctenopharyngodon idella) could further reduce the distribution of the SAV that may serve as a crucial habitat for economically and ecologically important Lake Erie fish species. I mapped the distribution of macrophytes in the northern section of Maumee Bay to quantify the utilization of SAV by juvenile fishes and the current distribution of SAV. In summer 2014 I used side scan sonar images processed in Quester Tangent™ computer programs in order to provide this baseline distribution. The 300-hectare mapped area was primarily inhabited by two SAV species, eel grass (Vallisneria americana) and variable pondweed iv (Potamogeton gramineus), and this SAV was distributed over 43.7% of the area (131.2 hectares). The distribution of SAV seemed to be more related to the influx of sediments (open full item for complete abstract)

Committee: Jeff Miner Dr. (Advisor); Patrick Kocovsky Dr. (Committee Member); Dan Wiegman Dr. (Committee Member) Subjects: Aquatic Sciences; Biology; Environmental Management; Wildlife Conservation; Wildlife Management

MS, University of Cincinnati, 2001, Engineering : Industrial Engineering

Autonomous robots with mobile capability are finding lot of applications in manufacturing, medicine, space and defense. Design of such a robot is truly a daunting task. The issue is complex because the robot has to interact with its environment when performing the task One of the possible application for a robot might entail moving the robot through a dynamically decided safe path. Such navigation could be seen as a guiding of a series of mobile robots to a desired destination along a just decided safe path. Numerous research works has been done in the area of path planning and obstacle avoidance algorithms for navigating a robot intelligently through a unknown, unexplored Environment. This research work was done towards fulfilling the requirement of designing a mobile robot to follow a moving leader. The Center for Robotics Research at the University of Cincinnati has built a mobile robot named Bearcat II for the International Ground Robotics Competition being conducted by the Association for Unmanned Ground Systems (AUVS) every year. The objective is to make the Bearcat II follow a lawn mower driven by one of the judges while maintaining a safe distance of about 3 meters. A Polaroid ultrasonic transducer mounted on a micro-motor with an encoder feedback was used to track the co-ordinates and motion of the leader and the steering system is suitably adjusted for re-orienting the robot and to maintain the fixed distance between the robot and the leader. The readings of the sonar at the known adjustable angles are translated to the co-ordinate and relative motion of the leader. The Galil DMC controller suitably drives the left and right motor to steer the robot in the proper direction and at proper speed. This design yields a portable independent system, which could be suitably integrated or replaced with any different kind of sensor like a laser sensor, which could ascertain the position and motion of the leader.

Committee: Dr. Ernie Hall (Advisor) Subjects: Engineering, Industrial

MS, University of Cincinnati, 2002, Engineering : Industrial Engineering

Autonomous robots with mobile capability are finding lot of applications in manufacturing, medicine, space and defense. Technology for autonomous vehicles is a very active field right now. The University of Cincinnati's Bearcat robot is a test bed for research in various technologies related to autonomous vehicle navigation. One of the challenges is to steer the robot autonomously to follow a moving vehicle over an unmapped area at a specified distance. A single rotating sonar sensor is used with a restricted angle of sweep to obtain readings to develop a range map to plan an unobstructed path for an autonomous guided vehicle. The new rotating sonar system is certainly an improvement over the old stationary system in terms of better obstacle avoidance as well as reduced use of resources. Tuning of the motor and the installation of a new gearbox has helped in achieving better results. The laser further improves the object detection by providing us with more accurate data, larger range and the exact profile of the object. The new algorithm for follow the leader uses a laser scanner, extending and improving the rotating sonar logic. The data from the laser scanner/sensor and the vehicle position are used as input variables. The outputs are the new curvature and velocity of the robot in order to follow the leader. Autonomous transportation is one of the ways in which vehicle technology is developing. Intelligent Transportation Systems (ITS) which involve communication between vehicle and the road are the next step in automobile future. The development of automatic collision avoidance systems is improving the safety levels on highway by reducing number of accidents and at the same time making driving easier. These systems draw ideas from the follow the leader concept in their development.

Committee: Dr. Ernest Hall (Advisor) Subjects: Engineering, Industrial

MS, University of Cincinnati, 2002, Engineering : Industrial Engineering

Obstacle avoidance is one of the most critical factors in the design of autonomous vehicles such as mobile robots. One of the major challenges in designing intelligent vehicles capable of autonomous travel on highways is reliable obstacle avoidance. Obstacle avoidance may be divided into two parts, obstacle detection and avoidance control. Numerous methods for obstacle avoidance have been suggested and research in this area of robotics is done extensively. Three different methods for obstacle detection and avoidance are available on the BEARCAT III. These include fixed mounting of sonar sensors, a rotating sonar sensor and a laser scanner. The fixed mounting system uses two sonar sensors which are mounted at the outer front edges of the vehicle. The rotating sonar system consists of a Polaroid ultrasound transducer element mounted on a micro motor with an encoder feedback. The motion of this motor is controlled using a Galil DMC 1000 motion control board. It is possible to obtain range readings at known angles with respect to the center of the robot. The laser range scanner system consists of a SICK Optics laser scanner which returns a two dimensional profile of the horizontal region in front of the vehicle. The data from these systems can be used to detect and avoid obstacles. The systems were tested in July 2002 at the International Ground Robotics Competition. The BEARCAT III placed third in the autonomous challenge contest. This test bed system provides experimental evaluation of the tradeoffs among the systems in terms of resolution, range and computation speed as well as mounting arrangements. The significance of this work is in the increased understanding of obstacle avoidance for robot control and the applications of autonomous guided vehicle technology for industry, defense and medicine.

Committee: Dr. Ernest L. Hall (Advisor) Subjects:

Doctor of Philosophy, The Ohio State University, 2012, Geodetic Science and Surveying

The ability of living creatures to navigate their environment is one of the great mysteries of life. Humans, even from an early age, can acquire data about their surroundings, determine whether objects are movable or fixed, and identify open space, separate static and non-static objects, and move towards another location with minimal effort, in infinitesimal time spans. Over extended time periods humans can recall the location of objects and duplicate navigation tasks based purely on relative positioning of landmarks. Our ability to emulate this complex process in autonomous vehicles remains incomplete, despite significant research efforts over the past half century. Autonomous vehicles rely on a variety of electronic sensors to acquire data about their environment; the challenge is to transform that data into information supporting the objective of navigation. Historically, much of the sensor data was limited to the two dimensional (2D) instance; recent technological developments such as Laser Ranging and 3D Sonar are extending data collection to full three dimensional (3D) acquisition. The objective of this dissertation is the development of an algorithm to support the transformation of 3D ranging data into a navigation solution within unknown environments, and in the presence of dynamically moving objects. The algorithm reflects one of the very first attempts to leverage the 3D ranging technology for the purpose of autonomous navigation, and provides a system which enables the ability to complete the following objectives: • Separation of static and non-static elements in the environment • Navigation based upon the range measurements of static elements This research extends the body of knowledge in three primary topics. 1) The first is the development of a general method to identify n features in an initial data set from m features in a subsequent data set, given that both data sets are acquired via 3D ranging sensors. Accomplishing this objective, particularly (open full item for complete abstract)

Committee: Dorota Grejner-Brzezinska PhD (Advisor); Alper Yilmaz PhD (Committee Member); Ralph von-Frese PhD (Committee Member); Charles Toth PhD (Committee Member) Subjects: Geotechnology

Master of Science (MS), Ohio University, 2009, Electrical Engineering (Engineering and Technology)

As the desire and need for unmanned aerial vehicles (UAV) increases, so to do the navigation and system demands for the vehicles. While there are a multitude of systems currently offering solutions, each also possess inherent problems. The Global Positioning System (GPS), for instance, is potentially unable to meet the demands of vehicles that operate indoors, in heavy foliage, or in urban canyons, due to the lack of signal strength. Laser-based systems have proven to be rather costly, and can potentially fail in areas in which the surface absorbs light, and in urban environments with glass or mirrored surfaces. SONAR based systems, however, do not fall victim to any of the outlined problems above, and for this reason, is an area of navigation which requires further study and development. This thesis will cover techniques for localizing and classifying targets in a vehicles environment using a SONAR-based system and an inertial measurement unit (IMU). The thesis will also cover the extension of these localization and classification techniques for the use of 3D motion detection and tracking of objects.

Committee: Maarten Uijt de Haag (Advisor); Michael Braasch (Committee Member); Frank Van Graas (Committee Member); William Kaufman (Committee Member) Subjects: Engineering

Doctor of Philosophy (PhD), Ohio University, 1990, Electrical Engineering & Computer Science (Engineering and Technology)

This dissertation studies the theoretical underpinnings of active sonar classifiers. We present a systematic approach for designing optimal Bayesian classifiers and analyzing their performance. We emphasize the ternary case where three hypotheses are considered: H 0(noise only), H 1(reverberation plus noise) and H 2(target plus noise). We start by deriving a sufficient statistic to decide between H 1and H 2, assuming H 0has already been eliminated. Then, closed-form solutions for classification and false alarm probabilities are obtained and several receiver operating characteristics curves illustrating meaningful physical scenarios are presented. Two classes of illuminating signals are considered: high resolution and linear FM signals. Many design parameters affecting classifier performance are studied. Perhaps the most important issue is classifier performance when incorrect a priori knowledge of the target's spatial properties is processed. Other parameters such as target resolution, signal-to-noise ratio, transmitter constant in linear FM signals, etc. are investigated as well. The final issue presented is acoustic target imaging. A minimum variance linear unbiased estimator of the scattering coefficients of the test volume encompassing the target is derived. Furthermore, we investigate error minimization of the MVLU estimator in terms of system characteristics such as array and/or signal design. We also discuss the relation between classification and imaging. In summary, ideas from decision theory, detection and estimation theory are combined in order to implement optimal Bayesian classifiers and acoustic imagers.

Committee: John Tague (Advisor) Subjects: