Publications

People’s perceptions of Autonomous Vehicles (AVs) are critical to understanding the role of AVs in future transportation systems. Most previous work on AVs perceptions is based on large cities or metropolitan areas. This study provides a unique perspective regarding perceptions of impacts of AVs in small and rural communities through an online survey in Central Texas (n = 1153). Our questionnaires gathered basic socio-demographic characteristics and AV impacts variables identified from the literature. We used summary statistics and ordered logistic regression models to reveal the perceived impacts of AVs. Residents of small and rural communities, particularly older adults (65 + years), were more enthusiastic about the development of AVs than the national average. Our findings reveal that being an employed, married male with a higher income increases the likelihood of accepting the impacts of AVs, suggesting further research to explore a feasible approach to utilizing AVs in small, rural communities.

IEEE Access 2024

Voronoi diagrams are widely used for area partitioning and coverage control. Nevertheless, their utilization in non-convex domains often necessitates additional computational procedures, such as diffeomorphism application, geodesic distance calculations, or the integration of local markers. Extending these techniques across diverse non-convex domains proves challenging. This paper introduces the adaptive centroidal Voronoi tessellation (aCVT) algorithm, which combines iterative centroidal Voronoi tessellation (iCVT) with an innovative agent dropout and reinsertion strategy. This integration aims to enhance area coverage control in non-convex domains while maintaining adaptability across varied environments without the need for complex computational processes. The efficacy of this approach is validated through simulations involving non-convex domains with disjoint target areas, obstacles, and shape constraints for both homogeneous and heterogeneous agents. Additionally, the aCVT algorithm is extended for real-time coverage control scenarios. Performance metrics are employed to assess the distribution of partitioned Voronoi regions and the overall coverage of the target areas. Results demonstrate improved performance compared to methods that do not incorporate the agent dropout and reinsertion strategy.

System architecture for virtual-physical hybrid swarm simulation

This paper introduces a hybrid robotic swarm system architecture that combines virtual and physical components and enables human–swarm interaction through mixed reality (MR) devices. The system comprises three main modules: (1) the virtual module, which simulates robotic agents, (2) the physical module, consisting of real robotic agents, and (3) the user interface (UI) module. To facilitate communication between the modules, the UI module connects with the virtual module using Photon Network and with the physical module through the Robot Operating System (ROS) bridge. Additionally, the virtual and physical modules communicate via the ROS bridge. The virtual and physical agents form a hybrid swarm by integrating these three modules. The human–swarm interface based on MR technology enables one or multiple human users to interact with the swarm in various ways. Users can create and assign tasks, monitor real-time swarm status and activities, or control and interact with specific robotic agents. To validate the system-level integration and embedded swarm functions, two experimental demonstrations were conducted: (a) two users playing planner and observer roles, assigning five tasks for the swarm to allocate the tasks autonomously and execute them, and (b) a single user interacting with the hybrid swarm consisting of two physical agents and 170 virtual agents by creating and assigning a task list and then controlling one of the physical robots to complete a target identification mission.

This paper presents a mobile robot for amphibious surface locomotion called ARMoR. The locomotion system of ARMoR consists of two wheel-and-leg transformable mechanisms and a customizable balancing tail. A sphere body chassis containing electronic components assembles the wheels and the tail. A combination of chassis design and transformable wheels allows ARMoR to safely navigate various environments, including diverse terrains and water surfaces. The robot is controlled and operated using an embedded microprocessor interfacing with sensing, communicating, and powering modules, including the Global Positioning System (GPS), camera, Inertial Measurement Unit (IMU), wireless communication module, and batteries. ARMoR was tested for its locomotion capabilities on concrete, dirt, grass, rocky surface, low brush, stairs, and water. On concrete, dirt, and grass, ARMoR operated in the wheeled mode; on other surfaces, the wheels transformed into the legged configuration enabling the robot to traverse challenging surface conditions effectively. ARMoR successfully traversed all terrains, and the traversal speeds were measured.

Background: Cognitive assessment using tangible objects can measure fine motor and hand-eye coordination skills along with other cognitive domains. Administering such tests is often expensive, labor-intensive, and error prone owing to manual recording and potential subjectivity. Automating the administration and scoring processes can address these difficulties while reducing time and cost. e-Cube is a new vision-based, computerized cognitive assessment tool that integrates c
omputational measures of play complexity and item generators to enable automated and adaptive testing. The e-Cube games use a set of cubes, and the system tracks the movements and locations of these cubes as manipulated by the player.

Objective: The primary objectives of the study were to validate the play complexity measures that form the basis of developing the adaptive assessment system and evaluate the preliminary utility and usability of the e-Cube system as an automated cognitive assessment tool.

Methods: This study used 6 e-Cube games, namely, Assembly, Shape-Matching, Sequence-Memory, Spatial-Memory, Path-Tracking, and Maze, each targeting different cognitive domains. In total, 2 versions of the games, the fixed version with predetermined sets of items and the adaptive version using the autonomous item generators, were prepared for comparative evaluation. Enrolled participants (N=80; aged 18-60 years) were divided into 2 groups: 48% (38/80) of the participants in the fixed group and 52% (42/80) in the adaptive group. Each was administered the 6 e-Cube games; 3 subtests of the Wechsler Adult Intelligence Scale, Fourth Edition (WAIS-IV; Block Design, Digit Span, and Matrix Reasoning); and the System Usability Scale (SUS). Statistical analyses at the 95% significance level were applied.

Results: The play complexity values were correlated with the performance indicators (ie, correctness and completion time). The adaptive e-Cube games were correlated with the WAIS-IV subtests (r=0.49, 95% CI 0.21-0.70; P<.001 for Assembly and Block Design; r=0.34, 95% CI 0.03-0.59; P=.03 for Shape-Matching and Matrix Reasoning; r=0.51, 95% CI 0.24-0.72; P<.001 for Spatial-Memory and Digit Span; r=0.45, 95% CI 0.16-0.67; P=.003 for Path-Tracking and Block Design; and r=0.45, 95% CI 0.16-0.67; P=.003 for Path-Tracking and Matrix Reasoning). The fixed version showed weaker correlations with the WAIS-IV subtests. The e-Cube system showed a low false detection rate (6/5990, 0.1%) and was determined to be usable, with an average SUS score of 86.01 (SD 8.75).

Conclusions: The correlations between the play complexity values and performance indicators supported the validity of the play complexity measures. Correlations between the adaptive e-Cube games and the WAIS-IV subtests demonstrated the potential utility of the e-Cube games for cognitive assessment, but a further validation study is needed to confirm this. The low false detection rate and high SUS scores indicated that e-Cube is technically reliable and usable.

This paper presents a facial emotion recognition technique using two newly defined geometric features, landmark curvature and vectorized landmark. These features are extracted from facial landmarks associated with individual components of facial muscle movements. The presented method combines support vector machine (SVM) based classification with a genetic algorithm (GA) for a multi-attribute optimization problem of feature and parameter selection. Experimental evaluations were conducted on the extended Cohn-Kanade dataset (CK+) and the Multimedia Understanding Group (MUG) dataset. For 8-class CK+, 7-class CK+, and 7-class MUG, the validation accuracy was 93.57, 95.58, and 96.29%; and the test accuracy resulted in 95.85, 97.59, and 96.56%, respectively. Overall precision, recall, and F1-score were about 0.97, 0.95, and 0.96. For further evaluation, the presented technique was compared with a convolutional neural network (CNN), one of the widely adopted methods for facial emotion recognition. The presented method showed slightly higher test accuracy than CNN for 8-class CK+ (95.85% (SVM) vs. 95.43% (CNN)) and 7-class CK+ (97.59 vs. 97.34), while the CNN slightly outperformed on the 7-class MUG dataset (96.56 vs. 99.62). Compared to CNN-based approaches, this method employs less complicated models and thus shows potential for real-time machine vision applications in automated systems.

Pre-print author copy (peer-reviewed and accepted): XLiuIEEESensors2020