This article introduces a human-in-the-loop (HITL) algorithm designed to address a real-world individual animal identification problem: identifying distinct animals from a set of unlabeled camera-trap images when the total number of individuals is unknown. A key contribution of this article is its explicit balance of two critical factors in evaluating identification algorithms: identification accuracy and the cost of expert human annotation, i.e., the number of manual decisions required to determine whether an image pair belongs to the same individual. Our HITL strategy uses human involvement only in the most difficult cases, employing autonomous identification for easily distinguishable cases, to obtain high overall identification accuracy with minimal human effort. The proposed framework consists of three components: HITL-based core clustering generation, HITL-based clustering verification, and HITL-based clustering growth. Experimental validation was performed on an African leopard dataset provided by Panthera, with the algorithm achieving identification accuracy comparable to a human baseline method, where each image is manually compared to its top-2 most similar images. The proposed approach reduces human involvement by 77.3%, requiring only 0.05% of all pairs to be manually labeled as belonging to the same or different individuals. Note to Practitioners-This paper addresses the real-world challenge of individual animal identification in an unlabeled image dataset. We propose a human-in-the-loop algorithm that achieves high identification accuracy with minimal human involvement and without prior knowledge. The method is particularly effective for unlabeled camera-trap image datasets of species with distinctive identifiable markings and a high individual-to-image ratio, especially when many animals appear in only a single image. By incorporating human confirmation, the algorithm can differentiate image pairs of the same individual that appear visually similar from those of different animals, due to significant variations in animal poses and other factors. Importantly, the algorithm significantly reduces the human effort required to determine whether image pairs belong to the same identity. This approach is particularly valuable for researchers studying new habitats or working with species for which deep learning techniques are not viable due to the lack of labeled datasets.

This article describes an algorithm to solve the real-world animal identification problem, i.e., determine the unknown number of $K$ individual animals in a dataset of $N$ unlabeled camera-trap images of African leopards, provided by Panthera. To determine the leopards’ IDs, we propose an effective automated algorithm, that consists of segmenting leopard bodies from images, scoring similarity between image pairs, and clustering followed by verification. To perform clustering, we employ a modified ternary search that uses a novel adaptive $k$ -medoids $++$ clustering algorithm. The best clustering is determined using an expanded definition of the silhouette score. A new post-clustering verification procedure is used to further improve the quality of a clustering. The algorithm was evaluated using the Panthera dataset that consists of 677 individual leopards taken from 1555 images, and resulted in a clustering with an adjusted mutual information score of 0.958 as compared to 0.864 using a baseline $k$ -medoids $++$ clustering algorithm. Note to Practitioners —We proposed an effective automated algorithm to solve the real-world animal identification problem: identifying $K$ unknown individual animals in $N$ images of a given species, with most animals only represented by a single image. This algorithm is different from other methods that assume all images in a dataset are from known individuals and thus regard the animal ID problem as a retrieval identification task. Our approach consists of a new adaptive $k$ -medoids $++$ clustering algorithm and a novel post-clustering verification procedure. The clustering is performed based on the degree of similarity between all image pairs in the dataset with the result validated using an expanded definition of the silhouette score. The accuracy of our algorithm was demonstrated on a real-world image dataset of African leopards, a small dataset with a relatively large ratio of $K/N$ , provided by Panthera. Code has been made available at: https://github.com/obaiga/Automatic-individual-animal-identification.

Artificial swarms have the potential to provide robust, efficient solutions for a broad range of applications from assisting search and rescue operations to exploring remote planets. However, many fundamental obstacles still need to be overcome to bridge the gap between theory and application. In this characterization work, we demonstrate how a human rescuer can leverage minimal local observations of emergent swarm behavior to locate a lone survivor in maze-like environments. The simulated robots and rescuer have limited sensing and no communication capabilities to model a worst-case scenario. We then explore the impact of fundamental properties at the individual robot level on the utility of the emergent behavior to direct swarm design choices. We further demonstrate the relative robustness of the simulated robotic swarm by quantifying how reasonable probabilistic failure affects the rescue time in a complex environment. These results are compared to the theoretical performance of a single wall-following robot to further demonstrate the potential benefits of utilizing robotic swarms for rescue operations.

Fabiola P. Ehlers-Zavala was named INTO Colorado State University (CSU)’s Center Director in November 2014, having previously fulfilled the role of INTO CSU Academic Director (March 2013-November 2014). In her CD capacity, she works with Colleges across campus, and has a particular interest in the preparation of international students pursuing engineering degrees at the undergraduate and graduate levels. She earned her B.A. in English Language and Literature together with her teaching certificate from the Pontificia Universidad Catolica de Valparaiso in 1992. She then pursued graduate education in the U.S., and she earned both her M.A. (1994) in English and Ph.D. in English Studies (1999) from Illinois State University. Upon her graduation, she worked for Illinois State University (ISU) as Assistant Professor in Bilingual/Bicultural Education until she received her tenure and promotion to Associate Professor. While at ISU, in her last year, she directed the Bilingual/Bicultural Education Program in the College of Education. In 2006, she moved to Colorado to teach in the M.A. in English at Colorado State University where she received her second tenure and promotion to Associate Professor in 2009. Between 2009 and 2013, she directed the M.A. in English (TESL/TEFL). Her areas of expertise include: second language/bilingual reading, second language assessment, and ESL/bilingual teacher preparation. She is the coauthor of Reading Strategies for Spanish Speakers. Her publications in books and journals include ”Meeting the reading comprehension challenges of diverse English language learners in K-12: Key contributions from reading research” (2016), ”Advocacy in Language Teaching” (2013), ”History of Bilingual Special Education” (2011), ”Bilingualism and Education: Educating At-Risk Learners” (2010), ”How Can Teachers Help Adolescent English Language Learners Attain Academic Literacy?” (2009), ”Teaching Adolescent English Language Learners” (2008), ”Assessing English Language Learners (ELLs) in Mainstream Classrooms” in The Reading Teacher (2006 & reprinted in 2010); ”Bilingual Reading from a Dual Coding Perspective” (2005) in Proceedings of the 4th International Symposium on Bilingualism; ”Preparing Quality Bilingual/Bicultural Teachers in the 21st Century: A PDS Model for Educational Change and Success” (2004), ”Use of Lexical Borrowings in Sonoran Border Spanish” (2003). She serves on the editorial board of The International Multilingual Research Journal (IMRJ) and TESOL Journal. She is Past President of Illinois TESOL/Bilingual Education, Past Chair TESOL International Bilingual Education Interest Section and Past Chair of TESOL International’s Nominating Committee. Most recently, Dr. Ehlers-Zavala served as a member of the International TESOL Diversity and Inclusion Committee (2014-2015), and has been invited to serve in the Editorial Review Board for The Reading Teacher (RT) Volume 70 review year (2015-2016).

In this study, we design, evaluate, and compare multiple heuristic techniques for mission scheduling of distributed systems comprising unmanned aerial vehicles (UAVs) in energy-constrained dynamic environments. These techniques find effective mission schedules in real-time to determine which UAVs and sensors are used to surveil which targets. We develop a surveillance value metric to quantify the effectiveness of mission schedules, incorporating the amount and usefulness of information obtained from surveilling targets. We use the surveillance value metric in simulation studies to evaluate the heuristic techniques with a reality-based randomized model. We consider two comparison heuristics, three value-based heuristics, and a metaheuristic that intelligently switches between the best value-based heuristics. Additionally, preemption and filtering techniques are applied to further improve the metaheuristic. We show that, for all scenarios that we consider, the novel modified metaheuristics find solutions that are the best on average compared to all other techniques that we evaluate.

The aim of this paper is the development of a redundancy resolution scheme for manipulators able to cope with kinematic constraints. In detail, the structure of the controller is of weighted least norm (WLN) type. The constraints are modeled as unilateral inequalities and can be general scalar functions (linear or nonlinear) of both the joint position and the joint velocity variables. In this work, a general procedure is proposed in order to include constraints of different types, namely functions of joint position or velocity only, functions of both joint position and velocity with a time dependent or time independent threshold. Simulations are performed in Matlab-Simulink environment and two tests are performed: the first employs a single 7-DOF arm, while in the second a dual-arm system composed of two 7-DOF manipulators is used. Results show that the proposed redundancy resolution scheme is capable of satisfying complex inequality constraints where other known methods fail.

It has been shown that one can guarantee a reachable workspace for a kinematically redundant robot after an arbitrary locked-joint failure if one artificially restricts the range of its joints prior to the failure. This work presents an algorithm for computing the optimal kinematic parameters and artificial joint limits for a robot to maximize this so-called "failure-tolerant workspace". The proposed technique employs a genetic algorithm that incorporates a novel method for selecting an initial population that results in fast convergence to high-quality solutions. The algorithm is illustrated on multiple examples of kinematically redundant robots and is shown to be computationally tractable even for robots that perform tasks in 6D workspaces.

This article considers the problem of planning a trajectory that maximizes the probability that a robot will be able to complete a set of point-to-point tasks, after experiencing locked joint failures. The proposed approach first develops a method to calculate the probability of task failure for an arbitrary trajectory based on its failure scenarios, which are efficiently computed by identifying the ranges of task point self-motion manifolds. Then, a novel trajectory planning algorithm is proposed to find the optimal trajectory with maximum probability of task completion. The planning algorithm exploits the overlap of self-motion manifold bounding boxes, as opposed to always using the shortest distance, to determine an optimal trajectory. The proposed trajectory planning algorithm is demonstrated on planar positioning 3R, spatial positioning 4R, and spatial positioning/orienting 7R redundant robots, resulting in average improvement of 17%, 22%, and 30%, respectively, compared to the best shortest distance trajectory.

At Colorado State University, we are actively reinvigorating our Electrical and Computer Engineering (ECE) curriculum to increase diversity and retention. In Fall 2019, we implemented a novel, one-credit hour, career emulation course for first-year students considering a degree in ECE. The course was deliberately designed to help students imagine working as a professional engineer so they could make more informed decisions about their academic endeavors. Throughout the course, students were engaged in realistic engineering tasks and interacted with a diverse range of professional engineers. These experiences were created to ensure all students had the opportunity to visualize themselves in a professional engineering environment. Eighteen students were initially enrolled in the Fall 2019 implementation and a new cohort of 16 students enrolled in the Fall 2020 course offering. Course surveys, instructor observations, and discussions with students regarding their future career expectations were used to assess the effectiveness of the course. Based on these metrics, we achieved our primary goal of helping students make an informed decision about pursuing a degree in ECE by emulating informative workplace learning opportunities.

This paper discusses an integrated approach to electrical-engineering education that incorporates computer-assisted MATLAB-based instruction and learning into the junior-level electromagnetics course and newly created learning studio modules. In this model, creativity class sessions are followed by two comprehensive and rather challenging multi-week homework assignments of MATLAB problems and projects in electromagnetic fields. This is enabled by a unique and extremely comprehensive collection of MATLAB computer exercises and projects, reinforcing all important theoretical concepts, methodologies, and problem-solving techniques in electromagnetic fields and waves, developed by one of the faculty team members. These tutorials, exercises, and codes constitute a modern tool for learning electromagnetics via computer-mediated exploration and inquiry, exploiting the technological and pedagogical power of MATLAB software as a general learning technology. The novel approach introduces students to MATLAB programming of electromagnetic fields, as opposed to just passive demonstrations of MATLAB's tools and capabilities for computation and visualization of fields. MATLAB programming tutorials and assignments are designed to deepen student engagement and accommodate different learning styles so students can learn more effectively. In addition to improving students' understanding and command of MATLAB use and programming within the electromagnetics context and beyond, these exercises increase their motivation to learn and appreciation of the practical relevance of the material, and equip them with the tools and skills to excel in other courses and projects. The results of this project were qualitatively analyzed through feedback surveys given to the students at the end of each MATLAB assignment. The Electromagnetics Concept Inventory was also used.

Survey-based data of three home appliances are included in a residential demand response (DR) aggregation algorithm that performs resource re-allocation for peak demand reduction in a notional electric distribution system. In addition, new constraints are integrated into the resource allocation approach to alleviate the inconvenience of the participating customers due to rescheduling their home appliances. Our effort replaces some assumptions from prior work on the mathematical model of customer preferences with actual data from a survey to validate the prior work. The results confirm the feasibility of the DR aggregation approach in achieving profits for the aggregator while considering the comfort of the participating customers.

Previous work has shown that it is possible to guarantee a reachable workspace for a kinematically redundant robot after an arbitrary locked-joint failure if one artificially restricts the range of its joints prior to the failure. Identifying the optimal articial joint limits has been the subject of previous work to maximize this so-called “failure-tolerant workspace.” Unfortunately, these techniques are not feasible for a highly redundant robot operating in a spatial workspace. This work presents a novel hybrid technique for estimating the failure-tolerant workspace size for robots of arbitrary kinematic structure and any number of degrees of freedom performing tasks in a 6D workspace. The method presented combines an algorithm for computing self-motion manifold ranges to estimate workspace envelopes and Monte-Carlo integration to estimate orientation volumes to create a computationally efficient algorithm. This algorithm is then combined with the coordinate ascent optimization technique to determine optimal artificial joint limits that maximize the size of the failure-tolerant workspace of a given robot. This approach is illustrated on multiple examples of robots that perform tasks in 3D planar and 6D spatial workspaces.

Abstract The success of an efficient and effective aggregator‐based residential demand response system in the smart grid relies on the day‐ahead customer incentive pricing (CIP) and the load shifting protocols. An artificial neural network model is designed to generate the day‐ahead CIP for the aggregator based on historical data. Load scheduling is proposed as a day‐ahead optimization problem that is solved using a blocked sliding window technique using parallel computing. With the assumptions made, the proposed algorithm improved the aggregator performance by reducing the overall simulation time from 275 to 45 min and increasing the aggregator forecast profits and customer savings by 11.85% and 35.99% compared to the previous genetic algorithm‐based approach.

Kinematically redundant robots have extra degrees of freedom so that they can tolerate a joint failure and still complete an assigned task. Previous work has defined the "failure-tolerant workspace" as the workspace that is guaranteed to be reachable both before and after an arbitrary locked-joint failure. One mechanism for maximizing this workspace is to employ optimal artificial joint limits prior to a failure. This current work presents a technique for determining these optimal artificial joint limits that is based on the gradient ascent method. The proposed technique is able to deal with the discontinuities of the gradient that are due to changes in the boundaries of the failure tolerant workspace. The technique is illustrated using two examples of three degree-of-freedom planar serial robots. The first example is an equal link length robot where the optimal artificial joint limits are computed exactly. In the second example, both the link lengths and artificial joint limits are determined, resulting in a robot design that has more than twice the failure-tolerant area of previously published locally optimal designs. (C) 2019 Elsevier Ltd. All rights reserved.

One measure of the global fault tolerance of a redundant robot is the size of its self-motion manifold. If this size is defined as the range of its joint angles, then the optimal self-motion manifold size for an n-degree-of-freedom (DoF) robot is n × 2π, which is not typical for existing robot designs. This letter presents a novel two-step algorithm to optimize the kinematic structure of a redundant manipulator to have an optimal self-motion manifold size. The algorithm exploits the fact that singularities occur on large self-motion manifolds by optimizing the robots kinematic parameters around a singularity. Because a gradient for the self-motion manifold size does not exist, the kinematic parameter optimization uses a coordinate descent procedure. The algorithm was used to design 4-DoF, 7-DoF, and 8-DoF manipulators to illustrate its efficacy at generating optimally fault-tolerant robots of any kinematic structure.

Instruction in ethical considerations is an important part of every engineering discipline. In many cases, a student’s exposure to ethical issues is delayed until the capstone senior design experience. For example, we have included lectures devoted to ethics in our Electrical and Computer Engineering senior design program that start with an introduction to the National Society of Professional Engineers (NSPE) and Institute of Electrical and Electronics Engineers (IEEE) codes of ethics, and is then followed by a discussion of various ethical case studies. While this is common in many programs, surveys of our students have revealed that they do not value this instruction to the same level as the technical content that they acquire. To address this issue, our department is exploring ways of integrating ethics education throughout the curriculum as part of our NSF-sponsored RED (Revolutionizing Engineering and Computer Science Departments) project. The core goal of our RED framework is to provide a holistic education, where we view our program as an integrated system that is a collaboration among faculty and students. Our new organizational model emphasizes knowledge integration at many levels and includes three key threads that extend throughout the curriculum, namely, foundations, creativity, and professionalism. The professional formation thread is designed to convey the importance of professional skills in the development of engineers, so that they are prepared to enter the workplace. One critical component of this thread is exposing students to ethical considerations that they may encounter in their professional careers and preparing them to deal with them. This paper discusses the process by which we have identified how to deconstruct the components of a traditional delivery of ethics education and integrate them throughout the instruction of technical content. By crafting case studies to the technical material that the students are currently studying, we hope to have students make the explicit connection that ethical considerations are part of the engineering design process and not a component that is tacked on at the end. In addition, because the same faculty who are presenting the technical material are also involved in the discussion of the ethical issues that arise, we believe students will make the implicit correlation that these issues should be valued as much as the technical material. Finally, by reinforcing the ethical content at multiple touch points throughout the curriculum, we hope to see an increased sophistication of ethical analysis as the students move through our program.

Contribution: This article presents quantitative support that the changes implemented as part of Colorado State University's (CSU's) Revolutionizing Engineering Departments (REDs) grant produce statistically significant positive change through a series of nonparametric analysis techniques. Additionally, the set of nonparametric analysis techniques provides a novel approach to quantitatively analyzing student data after significant pedagogical changes are made to the undergraduate curriculum. Background: As part of the grant, a series of significant pedagogical changes were made to the electrical and computer engineering (ECE) undergraduate curriculum. A large portion of these changes relates to knowledge integration techniques, which are used to highlighting the intricate relationships between the three topics of electronics, signals and systems, and electromagnetics. This article presents an analysis of the outcomes that are in part due to these changes. Intended Outcomes: As a result of the grant and the associated curriculum changes, it was anticipated that the cumulative in-major grade point average for third-year students would increase. It was also anticipated that the in-major intercourse grades would be more positively correlated. The analysis techniques that were used provide novel examples of applications to student data. Application Design: The implemented changes described in this article directly follow from the goals of the National Science Foundation's RED program. Findings: Three nonparametric analysis techniques are performed on a collection of data from ECE undergraduates that was collected over 20 years. It is shown that the intertopical correlations between courses increase immediately following the implementation of the intervention discussed in this article, and statistically, significant evidence is presented supporting that the distribution of grades has positively changed following the intervention.

Communication skills are one of several professional skills that are required for engineering graduates that pose difficulties for engineering educators. The issues around these skills include what to teach, how to teach them, and how to assess students’ abilities. As part of a curriculum reform project that is a component of a larger department change effort, three required classes of the third year curriculum of an Electrical and Computer Engineering program have added a knowledge integration component that occurs approximately every five weeks. During these integration efforts, students are required to integrate knowledge from the three courses to evaluate the design of a cell phone. The goal of these efforts is to have the students connect the knowledge across the three courses using a practical real-world device. Additionally, students are required to produce short video presentations to demonstrate their abilities in integrating the knowledge and the ability to communicate this via a video presentation. The video presentation assignment includes several components. To provide scaffolding for the student efforts, a couple of high-quality example video presentations are made available. These videos were developed by graduate students involved with the knowledge integration project. After the graduate students developed initial versions of their videos, they were critiqued by several faculty and then finalized and made available to the students. A time limit of 7.5 minutes was given for each video. After producing the videos, students were required to perform an anonymous peer review of three classmates’ presentations. Additionally, one graduate student, not responsible for an example video, then also performed an assessment of the videos. The students were also required to provide guided self-reflections on their communication skills after they had completed their videos and performed and received peer reviews. Herein we present results of the assessment data collected for this project. There are a couple of goals related to the assessment of the videos. First, a comparison is made between the assessments of the GTA and peer assessments with the students’ self-reflections –looking for areas of consistency. Then a second evaluation was performed where a random selection of videos were evaluated by members of an industry advisory board to look for similarities and differences between their evaluation and the in-house, or academic, evaluations. The methodology for this work includes collecting the text-based evaluations from each constituents. These texts were then coded for emergent themes that are compared across the various constituents. The results of this work demonstrate the efficacy of combining peer reviews with self-reflections in the development of students’ communication skills.

An emission rate-based carbon tax is applied to fossil-fueled generators with a demand response approach called Smart Grid resource allocation (SGRA). The former reduces the capacity factors (CFs) of base load serving fossil-fueled units, while the latter reduces the CFs of peak load serving units. The objective is to quantify the integration of the carbon tax and the SGRA approach on CO2 emissions and electricity prices in a multi-area power grid. We illustrate this using the Roy Billinton test system and the results show potential for significant reductions in fossil fuel-based generation and CO2 emissions.

Robots in a swarm are programmed with individual behaviors but then interactions with the environment and other robots produce more complex, emergent swarm behaviors. One discriminating feature of the emergent behavior is the local distribution of robots in any given region. In this work, we show how local observations of the robot distribution can be correlated to the environment being explored and hence the location of openings or obstructions can be inferred. The correlation is achieved here with a simple, single-layer neural network that generates physically intuitive weights and provides a degree of robustness by allowing for variation in the environment and number of robots in the swarm. The robots are simulated assuming random motion with no communication, a minimalist model in robot sophistication, to explore the viability of cooperative sensing. We culminate our work with a demonstration of how the local distribution of robots in an unknown, office-like environment can be used to locate unobstructed exits.